^y4>.^^^ ;^- ^^^^
n;
EN MEM01E2AM
Homird V/. Estill
Gift of
Digitized by the Internet Archive
. in 2007 with funding from
IVIicrosoft Corporation
http://www.archive.org/details/comprehen.sivetreOOmellrich
A COMPREHENSIVE TREATISE ON INORGANIC AND
THEORETICAL CHEMISTRY
VOLUME I
H, O
BY THE SAME AUTHOR
INTRODUCTION TO MODERN INORGANIC
CHEMISTRY.
With 232 Illustrations. Crown 8vo, gs.
MODERN INORGANIC CHEMISTRY.
With 334 Illustrations. Crown 8vo, 12s. 6^.
HIGHER MATHEMATICS FOR STUDENTS OF
CHEMISTRY AND PHYSICS. With special
reference to Practical Work.
With Diagrams. 8vo, 21s. net.
THE CRYSTALLISATION OF IRON AND
STEEL : an Introduction to the Study of Metallo-
graphy.
With 65 Illustrations. Crown 8vo, 8s. 6d. net.
LONGMANS, GREEN AND CO.
LONDON, NEW YORK, BOMBAY, CALCUTTA, AND MADRAS.
A COMPREHENSIVE TREATISE
ON
INORGANIC AND THEORETICAL
CHEMISTRY
yCJ BY
J. W. MELLOR, D.Sc
VOLUME I
W/TB 274 DIAGRAMS
LONGMANS, GREEN AND CO
39, PATERNOSTER ROW, LONDON
55, FIFTH AVENUE, NEW YORK
BOMBAY, CALCUTTA, AND MADRAS
1922
All rights ^reserved
D]5
■)1fi-
2Deb(cate&
TO THE
PRIVATES IN THE GREAT ARMY
OF WORKERS IN CHEMISTRY
THEIR NAMES HAVE BEEN FORGOTTEN
THEIR WORK REMAINS
PREFACE
With due regard to the law hwnanum errare est, this work aims at giving a
complete description of all the compounds known in Inorganic Chemistry, and,
where possible, these are discussed in the light of the so-called Physical Chemistry.
The separation of Organic from Inorganic Chemistry is nothing more than a
conventional convenience ; it is probable that the sharper the line of demarcation,
the greater the loss which each of these divisions of chemistry will suffer. In the
analysis of inorganic compounds, for example, some extraordinarily sensitive tests
are available, and some extraordinarily clean separations can be effected by utilizing
the properties of certain organic compounds of the metals.
In the past, several complete records have been made. Starting from
W. Nicholson's A Dictionary of Chemistry (London, 1795-1808), there have
appeared in England : A. Ure's A Dictionary of Chemistry (London, 1821-35), and
H. Watts' A Dictionary of Chemistry (London, 1866-68), which was later edited by
H. F. Morley and M. M. P. Muir. There is also Sir Edward Thorpe's A Dictionary of
Applied Chemistry (London, 1890-92), a new and revised edition of which is now
in the press (1921). The English translation of L. Gmelin's Handbook of Chemistry
(London) appeared in nineteen volumes between 1848 and 1872. This work
covered both organic and inorganic chemistry. The sixth German edition appeared
as the Eandbmh der anorganischen Chemie (Heidelberg) in 1871-86, while the
seventh edition, commenced in 1905, is not yet complete. A number of other
related books have appeared in Germany. The more important of these are
A. Ladenburg's Eandworterbuch der Chemie (Breslau, 1882-89) ; H. von Fehling's
Neues Handworterhuch der Chemie (Braunschweig), which commenced in 1874 and
is not yet completed. It was founded on J. von Liebig, J. C. Poggendorff, and
F. Wohler's Handivorferhitch der reinen und angeivandten Chemie (Braunschweig,
1837-64). There is 0. Dammer's Handhuch der anorganischen Chemie (Stuttgart,
1892-1903), and his Handhuch der chemischen Technologie (Stuttgart, 1895-98).
R. Abegg's Handhuch der anorganischen Chemie (Leipzig), commenced in 1905, is
not yet completed. In France there are E. Fremy's Encyclopedie chimique (Paris,
1882-1905); C. A. Wurtz's Dictionnaire de chimie (Paris, 1868-1908); and
H. Moissan's Traite de chimie minerale (Paris, 1904-6). In Italy, L. Guareschi's
JVuovo enciclopedia di chimica (Torino), commenced in 1900, is still in progress.
I have been more or less indebted for hints and ideas to all the above-named
works, as well as to H. Kopp's Geschichte der Chemie (Braunschweig, 1843-47).
Much of the material of this work was compiled in card-index form long before
my Modern Inorganic Chemistry appeared ; and that work was really an abridge-
ment of this one. The references which were not included in the scheme of that
work will be found here. It was not originally intended to make the larger work
assume the exhaustive character which this book has now acquired. Rightly or
wrongly, I came to the conclusion that it is a mistake to load up a student with
vii
viii PREFACE
facts as if he were going to be a specialist in all branches of inorganic chemistry.
In addition to the general principles, the salient features of certain type-compounds
should be taught, and anything further should be left for works of reference, where
full information may be obtained — to be absorbed or forgotten as may be expedient.
Consequently, in the ideal case, a work of reference should not only give the
authorities for statements of fact, but it should also indicate what knowledge has
been gleaned on the particular subject in question. To do this in a practicable
manner, attention must be directed to the original publications on the subject.
This natui-ally makes the work of compilation extremely laborious ; in some cases,
indeed, it happens that scores of independent references are involved in the state-
ment of one particular fact. Fortunately I have rather a unique collection of
dissertations and theses ; in a few cases, these have not appeared in the regular
channels of publications. Where the original references are not in the libraries of
this country, I have had to depend on an assistant in Berlin, who has generally
been successful in tracking them where our libraries have failed. This has made
some references very costly. A large proportion of the references will be found in
the Abstracts of the London and American Chemical Societies, and of the Society
of Chemical Industry.
In the references, the usual abbreviation for the title of a periodical is given,
then follow in clarendon type the volume number, the page or pages, and, last of
all, the year of publication. In cases where a volume is made up from a number of
bulletins with independent pagination, the number of the bulletin is employed
instead of the page. In the cross-references, the first number in clarendon type
refers to the volume, the second to the chapter, and the third to the § (section).
It will be observed that the diagrams of the chapters have an independent
numeration.
In former times little more than a mere qualitative knowledge of the so-called
physical and mechanical properties of elements and compounds was considered
ample, but with the tremendous ramifications of the various industries increasing
demands for precise data have been made from the workers in pure science. In
reviewing the data I have been impressed with the prevailing lack of perspective in
the measurements of physical properties, for, in some cases, these have been carefully
measured with elaborate apparatus involving an experimental error hundreds of
times smaller than the magnitude of the disturbing effects produced by impurities.
It is not always enough to say that the materials were Herren X.Y.Z.'schen
" chemically pure " preparations. For instance, in pre-war days I have had to
make very serious complaints about the quantities of glass contained in their
highest grade " chemically pure " potassium pyrosulphate. This would not have
been suspected had its use not been attended by an epidemic of bad analyses. Of
course, the best representative values of the physical constants of pure elements
and compounds are very important, but in commercial work, materials of an
extremely high degree of purity are regarded more as chemical curiosities, and
larger errors may be introduced by using data — atomic weights, etc. — derived
from pure materials, than by using data obtained with material of '* commercial "
purity.
I think it was P. J. Macquer who apologized for the alphabetic form of the
subject-matter of his Dictionnaire de chymie (Paris, 1766), by stating that chemistry
was little more than a collection of facts scarcely entitled to the name of science,
or capable either of synthetic or analytic explanation ; and hence he concluded that
the dictionary form was the best mode of arranging the facts. The dictionary thus
PREFACE ix
belongs to a primitive stage in the development of a science in that it is but a
collection of facts to be employed in building up the science.
We now flatter ourselves that the periodic law has given inorganic chemistry a
scheme of classification which enables the facts to be arranged and grouped in a
scientific manner. The appearance of order imparted by that guide is superficial
and illusory. Allowing for certain lacunae in the knowledge of the scarcer elements
prior to the appearance of that law, the arrangements employed by the earlier
chemists were just as satisfactory, and in some cases, indeed, more satisfactory than
those based on the periodic law.
The arrangement of the subject-matter of inorganic chemistry according to the
periodic scheme is justified solely by expediency and convention. It has a tendency
to make teachers over-emphasize unimportant and remote analogies, and to under-
estimate important and crucial differences. I imagine that when we have found
a truer basis of classification, such differences as are displayed between, say, /err oswm
a.nd.ffirricum compounds will be exhibited as if two different elements are involved,
and that iron alone appears as the stable form when separated from these com-
pounds. Similar remarks apply to other multi-valent elements. The difference
between the higher and lower valent forms of an element with a given acid are
often greater than between the compounds of two totally different elements with
the same acid.
The first volume of this work is mainly introductory, and in it the atom is
considered to be the chemist's unit, or the unit of chemical exchange. The newer
work on the structure of atoms, and the so-called elements with variable atomic
weights will be introduced in the third volume, as a sequel to the radio-active
elements. The collection in the first volume of most of the generalizations required
for application to special cases in subsequent volumes has simplified many explana-
tions. This applies, for example, to thermal diagrams, equilibrium diagrams for
ternary systems, etc. The general historical sketches in this volume facilitate the
reviews of the histories of the elements and their compounds which appear in
subsequent volumes.
Hydrogen and oxygen, and the compounds of these two elements, have been
worked in with the introductory volume. The second volume includes the halogens
and the alkali metals. The ammonium compounds are included with the com-
pounds of the alkalies. The other elements will appear mainly in the order of the
periodic law. The metal hydrides, oxides, halides, sulphides, sulphates, carbonates,
nitrates, and phosphates are included with the metals ; the other compounds are
described with the acids, or the acidic elements. With the complex salts and inter-
metallic compounds of an element are included analogous compounds of ammonium,
hydrazine, and hydroxylamine, as well as of all those elements which have been
previously discussed. It should therefore be possible to locate a desired compound
from an inspection of the backs of the volumes, which are lettered to show what
elements are discussed inside. The indexes and cross-references are also available.
In the 1778 edition of his Dictionnaire, P. J. Macquer referred to la nomencla-
ture tres complete which was available. We are not so well provided to-day. Our
nomenclature is inadequate and insufficient ; nor has it sufficient elasticity to adapt
itself to increasing knowledge. Unfortunately, we have grown so accustomed to
the system inaugurated near the beginning of the last century that we are afraid
to make a drastic change.
The systematic names of many compounds naturally depend on what view is
taken of their constitution. Many names are thus determined by the prevailing
xii CONTENTS
by Weight (132); § 5. The Decomposition of Water by Metals (134); § 6. The
Decomposition of Water by Electricity (136) ; § 7. Cavendish's Experiments on
the Synthesis of Water by Volume (138).
CHAPTER IV
THE PHYSICAL PROPERTIES OF GASES
1. The Atmosphere (147); § 2. The Influence of Pressure on the Volume of Gases
— Boyle's Law (150) ; § 3. Deviations from Boyle's Law (152) ; § 4. Dalton's Law
of Partial Pressures (155); § 5. The Laws of Nature (157); § 6. The Influence
of Temperature on the Volume of Gases— Charles' Law (158); § 7. Deviations
from Charles' Law (162); § 8. The Critical State of Gases (164).
CHAPTER V
COMBINATION BY VOLUME
1. Gay Lussac's Law of Combining Volumes (171) ; § 2. Amadeo Avogadro's Postulate
(172); § 3. The Eelative Weights of the Molecules (174); § 4. The Formula of
Compounds (178) ; § 5. The Relative Weights of the Atoms (179) ; § 6. Methods
for Measuring the Vapour Densities of Gases, and of Volatile Liquids and Solids
(181) ; § 7. The Struggle of Avogadro's Hypothesis for Recognition (186) ; § 8.
Deviations from Avogadro's Law (192) ; § 9. Radicals or Radicles (197) ; § 10. The
Atomic Weights of the Elements (198) ; § 11. The Relation between the Molecular
Weights and the Volumes of Gases (201) ; § 12. Chemical Equations and Chemical
Arithmetic (202) ; § 13. The Relation between Atomic and Combining Weights —
Valency (204) ; § 14. The Polarity of Valency (211) ; § 15. The Association of Atoms
in Three Dimensions (213) ; § 16. The Evolution of the Valency Concept (216) ;
§ 17. Attempts to Explain Valency (225) ; § 18. Atomic, Molecular, and Specific
Volxmies (228).
CHAPTER VI
THE CLASSIFICATION OF THE ELEMENTS
1. The Classification of the Elements (248) ; § 2. Triads, and the Law of Octaves
(252) ; § 3. The Periodic Law— D. I. Mendeleeff and L. Meyer (255) ; § 4. The
Gaps in Mendeleeffs Tables of the Elements (260) ; § 5. The Application of the
Periodic Law (262) ; § 6. Some Defects in the Periodic Law (263).
CHAPTER VII
HYDROGEN
1. The Occurrence of Hydrogen in particular and of the Elements in general (270) ;
§ 2. The Preparation and Purification of Hydrogen (275) ; § 3. Chemical Affinity
(291); § 4. The Measurement of the Affinity between the Acids and the Metals
(294) ; § 5. Opposing Reactions. Guldberg and Waage's Law (297) ; § 6. The
Solubility of Hydrogen (301); § 7. The Physical Properties of Hydrogen (313);
§ 8. The Chemical Properties of Hydrogen (325) ; § 9. The Diffusion of Gases (338).
CONTENTS 3dii
CHAPTEB VIII
OXYGEN
§ 1. History of the Discovery of Oxygen (344) ; § 2. The Action of Heat on Mercuric
Oxide (347) ; § 3. The Action of Heat on Potassium Chlorate (349) ; § 4. The
Occurrence and Preparation of Oxygen (351) ; § 5. Catalysis (357) ; § 6. Consecutive
Reactions (359); § 7. Concurrent or Side Reactions (360); § 8. The Physical
Properties of Oxygen (363) ; § 9. The Chemical Properties of Oxygen (378) ; § 10.
The Origin of the Terms: Acid, Alkali, Base, Salt (382); § 11. Acids (385);
§ 12. Salts (387); §13. Neutralization (389),- § 14. Bases (393); § 15. Hydroxides
and Anhydrides (395); § 16. The Polar Theory of Chemical Combination (397);
§ 17. Binary and Unitary Theories of the Constitution of Acids and Salts (402).
CHAPTEE IX
WATER
1. The Cycle of Water in Nature (405); § 2. The Purification and Distillation of
Water (409) ; § 3. The Effect of Temperature and Pressure on the Volume of Water
(410) ; § 4. The Vapour Pressure of Water— Fusion and Boiling (423) ; § 5. Gibbs'
Phase Rule (444) ; § 6. Undercooling, Supersaturation, and MetastabiHty (450) ;
§ 7. The Allotropic Forms of Water (457) ; § 8. The Physical Properties of Water
(463) ; § 9. The Chemical Properties of Water (483) ; § 10. Hydrates and Hydrated
Salts (498) ; § 11. The Vapour Pressure of Hydrated Salts (501).
OHAPTEE X
SOLUTIONS
1. The Solubility of Solids in Water (506); § 2. The Freezing of Solutions (516);
§ 3. The Solubility of Liquids in Liquids (522) ; § 4. The Solubility of Gases in
Liquids— Henry's Law (527) ; § 5. The Solubility of Mixed Gases in Liquids —
Dalton's Law (533) ; § 6. Diffusion in Gases and in Liquids (536) ; § 7. Solution
Pressure — Osmotic Pressure (538) ; § 8. The Osmotic Pressure of Dilute Solutions
and the Gas Laws (543) ; § 9. The Relation between the Vapour Pressure of a
Solution and the Molecular Weight of the Solute (548) ; § 10. Distillation (553) ;
§ 11. Other Hypotheses explaining Osmosis (557) ; § 12. The Relation between the
Boiling Point pf a Solution and the Molecular Weight of the Solute (561) ; § 13.
The Relation between the Freezing Point of a Solution and the Molecular Weight
of the Solute (565) ; § 14. The Relation between the Solvent Power of a Solvent
and the Molecular Weight of the Solute (568); § 15. Anomalous or Abnormal
Results for the Molecular Weights of Substances in Solution (569) ; § 16. The Cause
of Solution (574) ; § 17. The Physical Properties of Solutions (578).
CHAPTER XI
CRYSTALS AND CRYSTALLIZATION
1. The Crystallization of Salts from Solutions (589); §2. Fractional CrystaUization
(590) ; § 3. Crystals (593) ; § 4. The Crystallization of Solids en masse (602) ;
§ 5. The Internal Structure of Crystals (607); § 6. The Seven Styles of Crystal
xiv CONTENTS
Architecture (613); § 7. The Growth of Crystals (623); § 8. Analysis of the
Structure of Crystals by X-rays (633) ; § 9. Liquid Crystals ; Crystalline Liquids ;
or Anisotropic Liquids (645) ; § 10. Isomorphism— Mitscherlich's Isomorphic Law
(651) ; § 11. The Eectifi cation of Atomic Weights by Isomorphism (668) ; § 12. The
Formulae of Minerals, and of Isomorphous Mixed Salts (668) ; § 13. Index of
Refraction and Dispersion (670).
CHAPTER XII
THERMODYNAMICS AND THERMOCHEMISTRY
1. Matter and Energy (688) ; § 2. Thermochemistry (697) ; § 3. The Principle of
Maximum Work (703) ; § 4. The Principle of Reversibility (706) ; § 5. Hess' Law
(708) ; § 6. The Degradation or Dissipation of Energy (711) ; § 7. Bound and Free
Available Energy (716) ; § 8. The Amount of Heat which can be Utilized for doing
Work (719) ; § 9. Non-productive Energy. Entropy (721) ; § 10. The Work done
by Afi&nity during a Chemical Reaction (730) ; § 11. The Effect of Temperature on
Chemical Equilibria (732).
CHAPTER XIII
THE KINETIC THEORY OF ATOMS AND MOLECULES
§ 1. The Molecular Theory of Matter (740) ; § 2. The Kinetic Theory of Gases— Boyle's
Law (742)^ § 3. The Kinetic Theory of Gases— Charles' Law and Avogadro's
Hypothesis (747); § 4. Attempts to Obtain a More Exact Gas Equation (754);
§ 5. J. D. van der Waals' Theory of Corresponding States (759) ; § 6. Summary of
the Kinetic Theory of Molecules (765) ; § 7. Ultramicroscopic Particles — Ultra-
microscopy (768) ; § 8. The Kinetic Theory of Atoms (782) ; § 9. The Two Specific
Heats of Gases (786); § 10. The Relation between the Two Specific Heats of a
Gas and the Degree of Freedom of its Molecules (790) ; § 11. The Molecular Heats
of Gases (795) ; § 12. The Specific Heats of Elementary Solids — Dulong and Petit's
Rule (798) ; § 13. Molecular Heats— Neumann's and Joule's Rules (805) ; § 14. The
Meaning of Dulong and Petit's Rule (808) ; § 15. The Quantum Theory of Energy
and Dulong and Petit's Rule (811) ; § 16. Debye's Theory of Atomic or Specific
Heats (815); §17. The Kinetic Theory of Solids (818); § 18. Reactions between
Solids — Spring's Experiments (824) ; § 19. The Vibration Frequency of Atoms and
Molecules (828) ; § 20. Empirical Relations between the Properties of Solids (834) ;
§ 21. The Kinetic Theory of Liquids (840) ; § 22. The Surface Tension and Surface
Energy of Liquids and Solids (846) ; § 23. The Association or Polymerization of
Liquids (860) ; § 24. Thermal Effects attending the Expansion and Compression
of Gases (862) ; § 25. The Liquefaction of Gases (868) ; § 26. The Manufacture of
Oxygen and Nitrogen from Liquid Air (874).
CHAPTER XIV
OZONE AND HYDROGEN PEROXIDE
1. The Discovery of Ozone and of Hydrogen Peroxide (877); § 2. The Modes of
Formation and Preparation of Ozone (878); § 3. The Occurrence of Ozone and
Hydrogen Peroxide (891); § 4. The Physical Properties of Ozone (893); § 5.
Oxozone, Ozonides, and Oxozonides (899) ; § 6. The Chemical Properties of Ozone
(901) ; § 7. The Constitution of Ozone (914) ; § 8. The Modes of Formation and
CONTENTS
XV
Preparation of Hydrogen Peroxide (922) ; § 9. The Physical Properties of Hydrogen
Peroxide (929) ; § 10. Quantitative Application of the Law of Mass Action (933) ;
§11. The Chemical Properties of Hydrogen Peroxide (936) ; § 12. The Qualitative
and Quantitative Determination of Ozone and Hydrogen Peroxide (949) ; § 13. The
Composition and Constitution of Hydrogen Peroxide (952) ; § 14. Peroxides and
Peracids (956).
OHAPTEE XV
ELECTEOLYSIS AND THE IONIC HYPOTHESIS
§ 1. The Products of Electrolysis (962); § 2. Faraday's Laws of Definite Electrolytic
Action (963) ; § 3. The Velocity of Electrolytic Conduction (967) ; § 4. The Effect
of the Solvent (968) ; § 5. The Ionic Hypothesis (969) ; § 6. The Electrolytic Con-
ductivity of Solutions (977) ; § 7. The Number of Ions in a Solution (978) ; § 8.
The Migration of Ions (983) ; § 9. The Speeds of Moving Ions— Kohlrausch's Laws
(986); § 10. "Abnormal" Osmotic Pressures and Ionization (990); § 11. Equili-
brium between Ionized and Non-ionized Solute (992) ; § 12. The Solubility Law
(995) ; § 13. Acids and Bases according to the Ionic Hypothesis (1000) ; § 14. The
Strengths of Acids and of Bases (1003) ; § 15. The Neutralization of Acids and
Bases (1006).
CHAPTEE XVI
ELECTRICAL ENERGY
§ 1. The Factors of Energy (1011) ; § 2. Electrochemical Series of the Elements (1013) ;
§ 3. Solution Pressure -Contact Differences of Potential (1015); § 4. The Ionic
Hypothesis and Chemical Eeactions (1026) ; § 5. Polarization— Back Electromotive
Force (1027) ; § 6. Decomposition Voltages (1030) ; § 7. Gas Cells (1033) ; § 8. The
Relation between Electrical and Thermal Energy (1036) ; § 9. Fractional Electro-
lysis—G. Magnus' Rule (1039).
INDEX 1041
2 INORGANIC AND THEORETICAL CHEMISTRY
there may be a retrograde movement by the advent of an age of intellectual dark-
ness ; yet, in the main, these three periods characterize the growth of science as surely
as the child, the boy, and the man characterize the development of an individual's
mind. Chemistr}'' is a particularly happy illustration of Comte's idea.
L The first, the msrthological, anthropomorphical, or superstitious stage.—
This represents the childhood of chemistry, for, as man emerged from the mists of
prehistoric antiquity, everything must have appeared to be full of wonder and
mystery. He was overawed by the wind and the rain ; by the lightning and the
thunder ; by the eclipse and the comet ; and by the rainbow and the clouds. The
student of nature lived in a bewildering dreamland of mixed magic and myth which
led him to ascribe supernatural explanations to inaccurately known facts, and
consequently, he seemed to be surrounded on all sides by un monde invisible des
esprits et des demons. Just as man's own actions seemed to be the result of his own
efforts and volitions, so did natural phenomena appear to be the work of benignant
or mahgnant spirits in air, earth, or sea ; and man accordingly made oblations to
their residing deities to secure their kindly offices. Chemical phenomena were
produced by spirits — the salamander or the sylph, the naiad or the nymph, the
undine or the gnome — indwelling in different bodies, whose aid was invoked by
incantation or charm to produce successful experiments.
Accordingly, men who studied nature in those days were often suspected of
tampering with the spirits of evil, and chemistrj'^ came to be known as one of the seven
devilish arts. So too arose a childish fear and hatred of science, and the belief—
widespread in the Middle Ages — that science is dangerous, and its votaries ought
to be suppressed. In illustration, in 1287, the Order of Dominicans proposed to
suppress chemical studies as had been attempted with physics in 1243 ; again, the
Accademia dei Segreti — Academy of Nature's Secrets — founded by ,1. B. Porta in
1602 for the discussion of scientific subjects, was dissolved by Pius III, after une
existence courte mais glorieuse, apparently because it was believed that magic and the
black arts were practised at its meetings. In the thirteenth century, Roger Bacon
was arraigned at Oxford on an indictment for practising sorcery and magic ; and
in order to disprove these accusations, he wrote his celebrated Epistola de secretis
operihus artis et naturce et de nullitate magice to show that phenomena and appear-
ances, then attributed to supernatural agencies, were simply due to the operation
of natural laws. Again, in his Magice naturalis (Naples, 1558), J. B. Porta tried to
show that the magic of nature is quite as wonderful as that of wizards and witches.
T. Thomson opens his work, The History of Chemistry (London, 1830), by
pointing out that chemistry sprang originally from delusion and superstition, and
was at its commencement exactly on a level with magic and astrology. Superstition
can flourish only where knowledge is imperfect and fragmentary. Day, adds C. J.
Keyser (1914), is just as mysterious as night, and the mystery of knowledge and
understanding is more wonderful and awesome than the darkness of the unknown.
Mysterious phenomena, explained in one generation as the vagarious work of
invisible demons or deities., appear to succeeding generations as the ordered workings
of natural laws. The mists of superstition are always dissipated as positive know-
ledge extends into wider and wider fields.
The cuneiform inscriptions and the records of antiquity which have been
transmitted to us, show that the early chemists were dominated by the gratuitous
assumption that *' the interior agencies which keep the world in motion were personal
forces essentially out of and above nature." The magician and the sorcerer, the
necromancer and the wizard were the founders and keepers of the first rudimentary
knowledge of nature. Accordingly, knowledge and superstition were interwoven
with wondrous ingenuity and subtilty. The alchemists, following the mysticism
introfcluced by the Alexandrian and Arabian schools, had virtually reverted to this
stage of development when they spoke of red bridegrooms (gold) and lily brides
(silver) ; of green dragons (mercury) and red lions (gold) ; of black crows (lead),
and yellow scorpions (sulphur) ; and of flying eagles, fugitive stags, and inflated
p
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 3
toads. One of the older chemists described the result of triturating mercuric
chloride with mercury, resulting in the formation of mercurous chloride, in these
pompous words :
The fierce serpent is tamed and the dragon so reduced to subjection as to oblige him to
devour his own tail.
The anonymous work, Artis aurifer(B quam chemiam vacant (Basil, 1572), represents
the dissolution of gold in aqua regia by a lion devouring the sun, as depicted in Fig. 1.
This language persisted even as late as the
eighteenth century. In W. Clarke's The
Natural History of Nitre (London, 1670), for
example, the red vapours formed when nitre
is heated in a retort are called " the flying
dragon."
The seven metals — gold, silver, an alloy,
copper, tin, iron and lead — known to the
early Chaldeans, were also designated by the
names and symbols of the seven greater
heavenly bodies — the Sun, Moon, Mercury,
Venus, Jupiter, Mars, and Saturn. A close
relation was supposed to subsist between
the metals and their respective planets so
that nothing could happen to the one which
was not shared by the other ; and it was
further supposed that experiments with any
particular metal were more likely to succeed Fia. 1.— Copied from an old Symbol repre-
when the governing planet was in the ascend- gf j^^ *^^ Dissolution of Gold in Aqua
ant, and near its zenith. Thus, in Para-
celsus' directions for preparing an amalgam of lead and mercury, the two
fluid metals are to be mixed " at the very moment of the conjunction of
Saturn and Mercury." In some cases it is possible to see a fanciful reason why
a particular metal was assigned to ^ particular heavenly body, but in other
cases the connection is too remote to hazard even a guess! En passant^ it
may be pointed out that an ingenious hypothesis to explain how the metals are
affected by the planets was in circulation long after the original fancies had been
forgotten. As N. Lemery expressed it in his Cours de chimie (Paris, 1675) :
An infinite number of minute corpuscles pass to and from the metals and the planets,
these corpuscles can easily pass through the pores of the metals and the planets they repre-
sent, but they cannot pass into other bodies whose pores are not figured properly to receive
them, or if they do get into other bodies, they cannot stay there to contribute any nourish,
ment. The metals are thus perfected and nourished by the influence which comes from the
planets and conversely.
n. The second or philosophical stage. — At last man roused himself from his
stupor of helpless wonder and childish guessing. He dimly realized some rnethod
in nature's inscrutable complexity. Unfortunately, his vision was soon bedimnied
and his mind intoxicated. Accordingly, we now find him arrogantly proclaiming
the supremacy and omnipotence of the human reason. The majority of educated
people of that age believed it to be undignified for a self-respecting man to make
experiments, and they did not consider knowledge obtained by observing nature
to be a serious subject worthy of mental occupation. Indeed, men were so proud
of their intellectual supremacy that they persuaded themselves that their fancies
about nature were finer, nobler, and more worthy of belief than nature herself ;
and Plato apparently considered that the secret laws of nature could be invented
by abstract thinking ; for, in his Republic, he said that " real knowledge is obtained
by a simple process of reasoning independently of all information furnished by the
senses." In his Phcedo, Plato expresses his delight with Anaxagoras' saying that
4 INORGANIC AND THEORETICAL CHEMISTRY
** the mind is the cause and orderer of all things." The numerous absurdities
obtained by the application of this principle are well exemplified in the pages of
Plato's Timceus, where there are many illustrations of the vanity of the attempt
to explain incomprehensible facts by nebulous words ; for example, Plato there
states :
The universe is a unique, perfect, and spherical production, because the sphere is the
most perfect of figures ; and it is animated and endowed with reason, because that which
is animated and endowed with reason is better than that which is not.
Even Aristotle, the father of logic, reasoned that a vessel containing ashes would
hold as much water as when the vessel contained no ashes. The conclusion is not
true, showing that Aristotle did not always recognize the need for the discipline of
the imagination by relentlessly checking reason against inexorable fact.
Thus, man did not always see with Cicero that nature is a better teacher than
the most ingenious philosopher. Prompted by a sublime imagination, R. Descartes,
in his Principia philosophice (Amsterdam, 1644), built a hypothetical universe
which had no substance, and is now regarded as little more than an idle dream.
Well might T. Bergmann's essay De indagando vero (1779) claim :
The philosophical method, by pretending to unlock the secrets of nature with ease and
expedition, soothes a natural impulse to explain all things ; and by assimiing everything
to be accessible to the human intelligence, administers pleasing flattery to vanity and
arrogance.
The methods of thinking, the much vaunted philosophy of Plato and Socrates,
in its attempt to proclaim the laws of nature from the throne of human reason,
actually obscured the path of progress for many centuries, for it became the fashion
to look with lofty scorn on knowledge gleaned by observing nature. Accordingly,
the leading philosophers worshipped what Erancis Bacon might have called idola
cogitationis— idols of the imagination ; they devoted themselves to fantastic and
chimerical hypotheses about material things ; and made no earnest attempt to
discriminate between the unreal and the real. As a result, their minds became so
prejudiced that the facts were either denied, or else explained by extravagant
ideas and fancies uncontrolled by truth and reality as we understand these terms
to-day.
in. The third, the scientific, or the positive era. — The marvellous Greeks gave
promise of inaugurating this era before the advent of Christianity, but the feeble
light kindled by Aristotle flickered and almost expired in the atmosphere of
mysticism which prevailed in the Middle Ages. During this period, man almost
reverted to the pandemonium of miracle and magic of his childhood days. The
light re-appeared about the thirteenth century, and gained brilliancy during the
succeeding centuries ; man then learned to see that nature is as she is, and is not
subjected to the capricious will of deity or demon ; man recognized that nature
is always conformable with herself without contradictions and without incon-
sistencies.
The growth of chemistry as a science was nourished in the seventeenth century
by the establishment of academies and societies for the cultivation of science.
The famous Society of Rosicrucians,'^ which flourished mightily towards the end of
the sixteenth or beginning of the seventeenth century, was perhaps an exception,
for, judging from the many books which were poured from its presses between 1600
and 1630, it rather fostered mysticism and obscurity, and was not favourable to
the true scientific spirit. Long before the advent of the scientific societies,
there were associations which fostered human knowledge, for example, the priests
of Egypt had their temple laboratories ; and the same spirit led to the formation of
the various schools of philosophy in Greece ; but the special feature of the later
associations was their energetic protest against the worship of antiquity, where the
authority of an ancient master was placed above experience.
The Accndemia del Cimento, founded at Florence, in 1657, under the presidency
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 5
of Prince Leopold de Medici, was the first scientific society of any importance ; its
main object was " the repudiation of any favourite system or sect of philosophy,
and the obligation to investigate nature by the pure light of experiment." Although
it lived but ten years, it enriched the world by leaving a volume of important
records of experiments, chiefly in pneumatics — Saggi di naturali esperienze fatti
nelV Accademia del Cimento (Firenz, 1666). This work has been reprinted a number
of times in several languages. The Royal Society of London was founded in 1660 ;
VAcademie des Sciences of Paris in 1666 ; the Academia naturce curiosorum of
Germany in 1652 ; and many others were founded in the eighteenth century.
In some exceptional cases, these associations degenerated into " fastnesses from
which prejudice and error were latest in being expelled ; and they joined in perse-
cuting the reformers of science." The attitude of the University of Paris towards
Galilei, and of the University of Oxford towards Roger Bacon have been cited as
examples. In general, however, the policy of these associations was to encourage
the investigation of nature by observation and experiment ; Arrierc les theories,
vivent les f aits ! was their watchword ; and, instead of clothing their results in the
enigmatical and allegorical language of the Rosicrucians, they sought to give a
candid and straightforward account of their investigations and thoughts. In this
way, the obscure mysticism of the Middle Ages was gradually dispelled. Man
thus rediscovered that he does not bring any knowledge into the world with him ;
that "the subtilties of nature far transcend the subtilties of the human reason"
(F. Bacon) ; and that " knowledge cannot be invented, it must be discovered."
Progress was then assured, and the manifold achievements of the observational
and positive sciences during the past century are in striking contrast with the
paucity of the results of philosophical thinking applied in vain for thousands of
years.
References.
1 H. Martineau, The Positive Philosophy of Auguste Comte, London, 1875 ; J. S. Mill, Auguste
Comte and Positivism, London, 1865 ; L. L. Bruhl, The Philosophy of Auguste Comte, London, 1903 ;
S. Brown, Essays Scientific and Literary, Edinburgh, 1858 ; C. J. Keyser, Science and Rdigion,
New Haven, 1914.
2 A. E. Waite, The Real History of the Rosicrucians, London, 1 887 ; H. Jennings, The Rosi-
crucians, their Rites and Mysteries, London, 1887.
§ 2. The Observation and Record of Facts. Collecting Data
The mind is like a blank tablet upon which experience writes that which is perceived
by the senses.- — Aristotle (b!c. 320).
To what can we refer for knowledge ? What can be a more certain criterion than the
senses themselves ? If we cannot trust the senses, how is it possible to distinguish what
is true from what is false ?• — Lucretius (b.c. 60).
I know only that truth is in the things and not in my mind which judges them, and that
the less I put my mind in my judgments about them, the more sure am I to come near to
the truth. — J. J. Rousseau (1770).
H. Poincare, in his La science et Vhyjpothese (Paris, 1904), emphasized in a very
telling manner that true knowledge about material things can be acquired only
through the senses — experientia docet ; there is no other way. Experience is the
well-spring of true knowledge ; experience alone can teach something new ; it alone
is irrefutable ; it alone can give certainty. The same idea was suggested by
Aristotle and the peripatetical philosophers : nihil est in intellectu, quod non prius
in senst^— nothing is in the intellect which was not first in the senses ; and by-
Roger Bacon in his Opus majus about 1266, when he said : Sitie experientia nihil
sufficienter sciri potest. Experience comprises all the impressions we observe and
perceive through the various organs of sense. These impressions are recorded in
our notebooks, dictionaries of chemistry, etc., as empirical realities or facts.
Although knowledge cannot transcend the human faculties, much of the data of
science is not directly furnished by the senses, for the senses are quite unable to
6 INORGANIC AND THEORETICAL CHEMISTRY
discriminate the subtilties of nature. For instance, the speed of light and the
size of atoms are magnitudes either too great or too small to be accessible to sense
perceptions. Yet much data derived indirectly from the insensible physical world
are assumed to be realities or facts, when actually they are known only by inference
from data furnished by the senses. Without facts, science can do nothing ; they
are the foundation and building stones of the whole superstructure. The edifice
can be stable only in so far as it is founded upon the immutability of facts. The
facts must be accurate, or the edifice will be unstable.
Not very many years ago, an apt quotation from one of the classical writers —
say Aristotle — was considered ample proof of the truth of any statement, and this
in spite of repeated warnings ; even in the thirteenth century, Albertus Magnus
could say :
I pored over the books of all the sages from Morienus, Aristotle, and Plato downward,
but yet I went wrong, until, by trial and mistakes, I at length discovered the truth.
Science does not accept P. Bonus' dictum, in his Margarita novella (Basil, 1572) :
** The mere fact that a great body of learned men believe a statement supersedes
the necessitv for proof." To-day, science looks askance on records of mere opinions,
and focuses its attention on records of facts. It is not always easy to record facts
faithfully without unconscious distortion or bias. What we wish, said Demosthenes,
that we believe ; what we expect, said Aristotle, that we find.i Things are not
always what they seem. Seeing is not always believing. It is often difficult to
distinguish appearances from realities for we are easily deceived by the mockery
of sensations. The senses cannot be divorced from the mind ; neither is always
to be trusted alone. The sun appears to rise and set ; in reality it does neither.
So, although experience is the source of truth, it may also be a source of error.
Superficial appearances may obscure hidden realities. Plato of old was unduly
oppressed with the illusions and deceptions of sensory impressions, and he was
accordingly led to deny the validity of knowledge derived from the sensations ;
but Aristotle rightly showed that difficulties arise only when the mind wrongly
interprets the testimony of the senses.
In 1689, John Locke 2 emphasized the view that the senses are the tentacula of
the mind because the mind primarily derives its knowledge of the external world
through the senses. All our knowledge, said he, consists of a stock of ideas which
were primarily produced in the mind by sensation, and which have remained after
the sensation had ended. Our knowledge of chemistry, physics, etc., depends on
the ability of the senses (i) to receive accurate impressions of the external world ;
and (ii) to convey these impressions to the mind or brain. When the mind receives
a sensation, it immediately begins to interpret the meaning, and it usually infers the
existence of something outside itself which gave rise to the sensation. It may
seem as if the mind directly perceives the external object which gives rise to the
sensations ; but this is an illusion. The mind apprehends the sensation alone ;
and it assumes that there exists a cause of the sensation external to itself. There
is no doubt about the sensation, but there is less certainty about the inference ;
the sensation must be accepted as a fact, but the inferential knowledge will be true
or false according as the interpretation of the external cause of the sensation was
correct or otherwise. The sensation does not err, it is the mind which fails when
it misinterprets the material furnished by the senses. Hence, Plato could say that
we do not see with the eyes but with our reason ; J. W. Goethe, that we see only
what we know ; and E. Mach (1883), that the adaptation of thoughts to facts is
the aim of all scientific research.
It is therefore sometimes necessary to receive with caution the testimony of
evidence derived from sensations. The mind interprets a sensation by comparing
it with some former sensation, the source of which has been previously determined.
Consequently, the faithfulness of the interpretation is dependent upon the memory
of past sensations, or upon the sensitiveness of the mind to detect resemblances and
THE EVOLUTION ANB METHODOLOGY OF CHEMISTRY 7
differences. Otherwise expressed, the accuracy of an inference as to the nature of
the objective source of a subjective sensation varies from a mere guess to virtual
certainty. 3 The idea has been aptly illustrated this wise : just as a nimiber of bits
of glass irregularly arranged always form symmetrical patterns when viewed
through the kaleidoscope, so does the understanding of each man impose a pattern
of its own upon the various sensations which it perceives. Consequently, as Robert
Hooke ■* once said : It is necessary to be on guard against deep-rooted errors which
may have been grafted upon science by the slipperiness of the memory, the narrow-
ness of the senses, and the rashness of the understanding. The greatest caution
must be exercised in accepting, on secondhand evidence, facts which cannot be
verified. No reliance can be placed on vague impressions. Evidence must be
clear and precise.
Few persons can estimate and register facts impartially and fairly. As W. S.
Jevons 5 puts it: " Among uncultured observers, the tendency to remark favourable,
and forget unfavourable events is so great that no reliance can be placed on their
supposed observations." T. Bergmann long ago drew attention to this very trait.
He said :
One observer will relate an event with the most extravagant encomiums ; another will
detract from its real merit ; a third, by some oblique insinuation, will cast suspicion on the
motive ; and a fourth will represent it as a crime of the blackest dye. These different
descriptions represent the character of the respective observers.
Untutored minds are very prone to mistake inferences for observations, and pre-
possessions for facts ; their observations and their judgments are alike vitiated by
dogma and prejudice ; they do not seek to investigate, they seek to prove. The
old proverb is inverted, believing is seeing. The student of science must pledge
himself to do his best to eliminate prepossession and dogma from his judgments,
and he must spare no pains to acquire the habit of recording phenomena as they
are observed ; and to distinguish sharply between what is or has been actually
seen, and what is mentally supplied. It requires a mind disciplined like a soldier
to avoid the natural inclination to look away from unwelcome facts.
The purity of truth is almost certain to be corrupted when the observer is ruled
by preconceived opinions, for, as 0. W. Holmes puts it : When we have found one
fact, we are very apt to supply the next out of the imagination ; or as T. Bergmann
said in his essay De indagando vero (1779) :
An observer swayed by preconceived opinions, may be considered as one who views
objects through coloured glasses, so that each object assumes a tinge similar to that of the
glasses employed. He who seeks the truth must learn to observe with equal candour
those facts which controvert his opinions, and those which favour them.
It is only in a pseudo-science, said 0. W. Holmes, that positive evidence, or
such as tells in favour of its doctrines, is admitted ; and all negative evidence, or
such as tells against it, is excluded. C. Darwin, in his Autobiography (London, 1887),
states that one of his golden rules was to make a memorandum of any fact or
thought which he found to oppose his general results, because he noticed by ex-
perience that such facts or thoughts were far more apt to escape the memory than
favourable ones. Above all, said Robert Hooke (1665), a good observer needs a
sincere hand and a faithful eye, to examine and record things themselves as they
really appear. " The mind and the reason of the trustworthy observer must be
trained to rebel against all desire, and to disobey all inclinations."
The belief that bodies contained a definite quantity of heat substance or caloric
prevented Black's successors from regarding the fact, known to every savage, that
heat is produced by friction ; the theory of phlogiston prevented some of the early
chemists from recognizing the increase in weight which occurs when metals are
calcined — oculos habent et non videbunt (Psalm 116. 5) ; the assumption that air is
absorbed when lead is roasted prevented Stephen Hales recognizing oxygen as the
gas evolved when red lead is heated ; and, as E. Mach (1892) has pdinted out in
8 INORGANIC AND THEORETICAL CHEMISTRY
his Populdre Vorlesungen (Leipzig, 1903), the undulatory theory of light prevented
C. Huygens marking the fact of polarization which Isaac Newton, undisturbed by
theories, perceived at once.
Refbbences.
* W Hamilton, Lectures on Metaphysics, Edinburgh, 1. 74, 1859.
" J. Locke, An essay concerning human understanding, London, 1689; E. Mach, Populdre
Vorlesungen, Leipzig, 1903.
* E. Mach, Beitrage zur Analyse der Empfindungen, Leipzig, 1885 ; Chicago, 1897 ; A, Philips,
Essays toirards a Theory of Knowledge, London, 1915 ; A. Rau, Empfindung und Denken, Giessen,
1896 ; P. Carus, The Primer of Philosophy, Chicago, 1904.
* R. Hooke, Micrographia, London, 1665.
^ W. S. Jevons, The Principles of Science, London, 1874.
§ 3. The Collating, Sifting, and Clarifying of Observations. Classifying Data
History teaches that the commencement of every branch of science is nothing more
than a series of observations and experiments which had no obvious connection with one
another.- — J. von Liebig (1846).
In order that the facts obtained by observation and experiment may be capable of
being used in furtherance of our exact and solid knowledge, they must be apprehended and
analysed according to some conceptions which, applied for this purpose, give distinct and
definite results, such as can be steadily taken hold of, and reasoned from.* — W. Whewell.
The record of facts obtained by observation and experiment, jper se, is empirical
knowledge. Empirical is derived from the Greek word ifxireLpLKo^, meaning ex-
perienced. It has just been emphasized that all knowledge is derived from
experience, and hence empiricism would appear to be the right method of
acquiring knowledge. The term, however, has slightly changed in meaning, for
it is now usually applied to chance experiences which occur irregularly without any
orderly plan of investigation.
All true science, said T. Huxley, must begin with empirical knowledge. Nature,
however, presents to our senses a panorama of phenomena co-mingled in endless
variety so that we are sometimes overwhelmed and dazed by the apparent com-
plexity of empirical knowledge. It is work for the intellect to educe the elements
of sameness amidst apparent diversity, and to see differences amidst apparent
identity. It is work for the judgment to reject accidental and transient attributes,
and to consolidate essential and abiding qualities. Consequently, while the primary
aim of science is to collect facts, the higher purpose of science is to show that,
amidst wild and terrible disorder, order and law reign supreme. The man of
science seeks a refuge from this bewildering complexity in unifying principles by
which the facts can be grouped and classified into systems. As he gazes into
nature, the man of science must be quick to discern hidden resemblances amidst
a thousand differences ; he must be quick to disentangle natural relations
from a medley of detail ; and quick to detect dissemblances amidst alluring
similarities.
Empirical knowledge describes facts ; science begins by comparing facts.
Empirical facts, in consequence, can form a science only when they have been
arranged, rearranged, grouped, or classified so as to emphasize the elements of
similarity and identity in different phenomena. Accordingly Thomas Hobbes
expressed the opinion that the main purpose of science is the tying of facts into
bundles. This bundle-tying, indeed, forms no small or insignificant part in the
development of science ; otherwise expressed, a significant advance has been made
in the development of a science when the observed facts have been codified into a
system so that a medley of empirical facts is systematically summarized under a
small number of heads. This means that the facts must be arranged in a methodical
and systematic manner until finally all the relevant facts taken together may form
one system. The process of classification and correlation is one of the methods
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 9
of scientific investigation. Knowledge so systematized is scientific knowledge.
T. Bergmann (1779) illustrated the idea in his essay previously cited :
A vast number of observations without order or regularity is not luilike a confused
heap of stones, lime, beams, and rafters requisite for constructing an edifice, but which
being combined with no skill fail in producing the proposed effect.
The material framework of the world appears in a myriad different guises and
combinations, but the chemist can resolve each combination into a few definite
elementary forms of matter ; similarly, a multitude of forces can be resolved into
comparatively a few primitive forms of energy. About 150 a.d., the Egyptian
astronomer Claudius Ptolemy measured the angles of incidence and refraction of a
beam of light passing from air into water, but more than fourteen hundred years
elapsed before W. Snell (1621) detected the law of refraction hidden in Ptolemy's
data. By tabulating his measurements of the volumes of air confined under different
pressures, Robert Boyle discovered the law known by his name. Each of these
laws summarizes in one simple rule myriads of possible measurements.
Scientific knowledge is not necessarily more accurate than empirical knowledge.
Empirical uncoordinated facts are no less true, definite, and real than scientific
facts, for all facts are equally true fer se. A collection of empirical facts always
requires some theory to serve as framework in order that the facts may be arranged,
grouped, and pigeon-holed. According to F. Hoefer (1843) :
II n'y a rien de plus stupide qu'un fait, quand il ne se rattache a aucune cause connue,
a aucune loi dominante. II faut done concilier I'individualisation des faits avec leur
generalisation. C'est la que reside le vrai crit^rium, I'avenir de la science.
If a group of facts — scientific facts — has been organized on an erroneous system,
the facts are no less true though the system be false. Chemistry presents a
curious mixture of empirical facts with isolated fragments of scientific knowledge.
§ 4. The Generalization of Observations
Facts are the body of science, and the idea of those facts is its spirit. — S. Brown.
It is the intuition of imity amid diversity which impels the mind to form science. —
F. S. Hoffman.
The correlation of empirical facts requires qualities of the mind different from
those employed in observation and experiment. Both qualities are not always
located in the same individual. Some excel in the one, not in the other.
J. Priestley, C. W. Scheele, and H. Davy, for instance, were admirable observers, but
they were not brilliant in the work of correlation ; J. Dalton and A. L. Lavoisier
were not particularly distinguished as experimenters, but they excelled in correlating
observed data. W. Hamilton ^ did not rate the fact-collecting faculty very highly.
He said :
In physical science the discovery of new facts is open to every blockhead with patience,
manual dexterity, and acute senses ; it is less effectively promoted by genius than by
co-operation, and more frequently the result of accident than of design.
J. Priestley (1783) recognized his own limitations when he said : " I have a tolerably
good habit of circumspection with respect to facts, but as to conclusions from them,
I am not apt to be very confident." Skill in the critical analysis of observational
data, and in collating, sifting, and clarifying records, is not a sufficient recommenda-
tion to the adytum — the sanctorum sanctissimum — of science. There is still a
higher type of work for but a few seekers after knowledge. It is
To search thro' all
And reach the law within the law. — Tennyson.
It is the sprite imagination which usually reveals the deeper meaning of facts which
have been diligently garnered, and laboriously sifted . ^
10 INORGANIC AND THEORETICAL CHEMISTRY
It cannot be doubted that science in its higher work, requires a supple and
well-developed imagination 2 which T. Gomperz says is the instrument of genius,
no less for scientific discovery than for artistic creation. The secret charm of
scientific discovery is not in the facts per se, but rather in the extrication of natural
relations among the facts one with another. Particular groups of facts must be
unified or generalized into a system — the so-called law. Science begins with facts
and ends with laws. Law is the essence of facts. As pointed out elsewhere, Newton's
celebrated law epitomizes in one simple statement how bodies have always been
observed to fall in the past. Immortal Newton did not discover the cause or the why
of the falling of the apple, but he did show that it was due to the operation of the
same forces which hold the earth, the planets, and their satellites in their appro-
priate orbits. Newton's simple and comprehensive law epitomizes in one single
principle the many and varied phenomena associated with falling bodies, planetary
motions, etc., and generally, the scientific generalization explains the operations
of nature by showing the elements of sameness in what at first sight appears to be
a confused jumble of phenomena. Generalization is the golden thread which
binds many facts into one simple description. That peculiar type of genius, that
rare quality of mind required for the, work of generalization, is found only in a Newton
or a Darwin. Plato said that if ever he found a man who could detect the one in
inany he would follow him as a god.
Unification is the supreme goal of modern science, or, as Heracleitus (c. 450 B.C.)
proclaimed, the highest goal of knowledge is the one law regulating all events.
However, with A. Comte,^ the majority will have la profonde conviction personelle,
that the attempt to explain all phenomena by une hi unique is chimerical. Several
natural phenomena belong to different categories, and are irreducible one to another.
At best, man has to apply a very weak intellect to a very complicated world ; and
the resources of the human intellect are too narrow, and the universe is too complex
to leave any hope that it will ever be within man's power to carry scientific perfection
to Tennyson's last degree of simplicity :
. . . one law, one element.
References.
1 W. Hamilton, Discussions on Philosophy and Literature, London, 239, 1852.
2 T. Gomperz, Greek Thinkers, London, 4. 125. 1912.
' A. Comte, Cours de philosophie positive, Paris, 1. 44, 1864 ; H. Martineau, The Positive
Philosophy of Av^uste Comte, London, 1. 13, 1875.
§ 5. The Aim of Science in General, and of Chemistry in Particular
Let us remember, please, that the search for the constitution of the world is one of the
greatest and noblest problems presented by nature. — G. Galilei.
The ordered beauty of the world of nature suggests an infinite inteUigence with powers
of action such as no man possesses.' — Benjamin Moore.
Science embraces the sum-total of human knowledge, and it ranges over the
whole realm of nature. Science is not a mass of empirical knowledge gained by
observation and experiment, but it is an organized body of facts which have been
co-ordinated and generalized into a system. Science tacitly assumes that nature
is a harmonious unity, and that rational order pervades the universe. Science seeks
a complete knowledge of the multitude of inter-related parts of the universe which
act and react on one another producing endless variety. In fine, science aims at
omniscience. The target, however, appears to recede with increasing knowledge.
As man grows in wisdom and knowledge, he begins dimly to realize that the unknown
multiplies into boundless proportions.
The sciences are too complex and too vast to be comprehended by one man's
mind.
One science only will one genius fit,
^ So vast is art, so narrow human wit. — Pope.
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY U
Our feeble wit has rendered it necessary to rear a tree of scientific knowledge with
many branches : astronomy, physics, chemistry, mineralogy, geology, biology,
sociology, etc. " The divisions of the sciences," said Francis Bacon, " are like the
branches of a tree that join in one trunk," and they are therefore more or less closely
related with one another. The astronomer, the physicist, the chemist, each usually
keeps to his own particular branch. This separation of the sciences is mere con-
vention. Even in the middle of the thirteenth century Roger Bacon saw that there
are no real lines of demarcation between the different sciences, for he pointed out
in his Opus tertium (1267) :
All the sciences are connected ; they lend each other material aid as parts of one great
whole. Each does its own work, not for itself alone, but for the other parts. ... No
part can attain its proper result separately ; since all are parts of one and the same com-
plete wisdom.
The science of chemistry is man's attempt to classify his knowledge of all the
different kinds of matter in the universe ; of the ultimate constitution of matter ;
and of the phenomena which occur when the different kinds of matter react one
with another. The science of chemistry is itself so vast, that many branchlets are
necessary for useful work, and thus we have : inorganic chemistry, organic chemistry,
physical chemistry, mineralogical chemistry, bio- chemistry, agricultural chemistry,
pharmaceutical chemistry, etc. The chemist also frequently aims at applying his
knowledge to useful purposes in the arts and industries ; and thus arises appHed,
industrial, or technical chemistry.
Applied chemistry. — About the middle of the thirteenth century, Roger Bacon
distinguished between knowledge sought for the sake of truth, and knowledge
utilized in the practice of the various arts ; or, as I. R. Averroes expressed it a
century earlier : In pure science, scimus ut sciamus ; and in applied science, scimus
ut operemur. The distinction, however, was recognized in the fourth century B.C.,
for it was explicitly expounded in Aristotle's Metaphysics, and it was also intimated
still earlier in Plato's Republic.'^ The purpose of pure science is to observe pheno-
mena and to trace their laws ; the purpose of art is to produce, modify, or destroy.
Strictly speaking there is no such thing as applied science, for, the moment the
attempt is made to apply, science passes into the realm of art. It has been well
said that " science is indebted to art for the means of experimenting, but she
instructs art concerning the properties and laws of the materials upon which the
latter operates." In an essay on The usefulness of experimental philosophy, Robert
Boyle (1663) emphasized the mutual benefits which would obtain when science,
or, as he called it, when natural philosophy is applied to the various arts and crafts ;
and he claimed that it is prejudice, no less pernicious than general, which has kept
science so long a stranger in the industries. Boyle's ideas have been still further
emphasized by Lord Kelvin (W. Thomson), who said in 1883 :
There cannot be a greater mistake than looking superciliously upon practical applica-
tions of science. The life and soul of science is its practical application, and just as the
great advances in mathematics have been made through the desire of discovering the
solutions of problems which were of a highly practical kind in mathematical science, so in
physical science many of the greatest advances that have been made from the beginning
of the world to the present time have been in the earnest desire to turn the knowledge of
the properties of matter to some purpose useful to mankind.
The so-called applications of science to the industrial arts — say, applied chemistry
— may be (i) An attempt to extend the methods of scientific investigation to the
industrial arts ; or (ii) To adapt known operations and laws to useful purposes.
When the chemist is occupied in the systematic observation of phenomena, and in
tracing their laws, he is engaged in scientific investigation, no matter if the work be
conducted in academy, in counting house, or in factory.
References.
^ W. Hamilton, Lectures on Metaphysics, Edinburgh, 1859; Anon., Chem. News, 18. 215,239,
263, 1868; 19. 1, 61. 109, 1869.
12 INORGANIC AND THEORETICAL CHEMISTRY
§ 6. Experiment
Experiment is the interpreter of nature. Experiments never deceive. It is our judg-
ment which sometimes deceives itself because it expects results which experiment refuses.
We must consult experiment, varying the circumstances, iintil we have deduced general
rules, for experiment alone can furnish reliable rules.- — Leonardo da Vinci.
Nature speaks to us in a peculiar language, the language of phenomena. She answers
all the questions we ask her, and these questions are our experiments. — J. von Liebig.
Chemistry is largely an experimental science. Experiment is really a method
of observation, which is employed when the facts are so masked by other conditions
that they cannot be accurately observed unless the obscuring conditions are sup-
pressed. The chemist would not make much progress if it were only possible to
observe phenomena just as they occur in nature, and not possible to make observa-
tions under determinate conditions. By experiment, it is possible to make combi-
nations of different forces, and different forms of matter which are not known to
occur in nature ; to eliminate complex disturbing conditions ; and to observe
phenomena under simplified conditions. An experiment has been well defined as
une observation provoquee. Experiment, said G. A. Reid, is useful only when there
are conditions which obscure direct observations. The most successful experiment
does no more than make a fact which was previously obscure as patent as one that
was open to direct observation from the first. Chemical phenomena, per se, are
usually too complex for our minds to grapple, and they must be simplified by
simple experiments. Consequently, chemistry is an experimental science because
its facts can rarely be observed in any other way. If data could be obtained by
direct observation, there would be no need for experiment.
It requires much acumen to determine the precise conditions under which an
experiment shall give a successful result. Every experiment has the character of
a specific question. The skilled questioner — the experimenter — knows what he is
asking, and he tries his best to interpret nature's reply, be it affirmative, negative,
or evasive. If the answer be negative or evasive, the question has not been properly
asked, and it must be plied again and again until
A sharphooked question baited with such skill
It needs must catch the answer.
Paradoxically enough, the investigator can usually say with " Dr. Moreau " : " ]
asked a question, devised some method of getting an answer, and got — a fresh
question." Some such ideas were in Robert Hooke's mind when he said :
The footsteps of nature are to be traced, not only in her ordinary course, but when she
seems to be put to her shifts, to make doublings, and turnings, and to use some kind of art
in endeavouring to avoid our discovery.
The more intricate the experiment, the greater the probability of an obscure
and ambiguous result. As A. L. Lavoisier has pointed out, " it is a necessary
principle in experimental work to eliminate every complication, and to make experi-
ments as simple as possible." The quality of an experiment, not the quantity, is
best adapted to throw light upon a phenomenon. Experiments carelessly performed
may be sources of error and obscurity. Many of the results obtained by the alchemists
in the Middle Ages show how ineffective or abortive are the results of experiments
in incompetent hands — here, the experiments wandered into eccentric by-paths,
and furnished preposterous conclusions. Experiment is an art, said G. A. Lewes
(1864) and demands an artist.
Joseph Priestley believed in making a large number of haphazard experiments,
and said that he discovered oxygen by trying the effect of heat on many substances,
apparently selected at random by John Warltire of Birmingham. Thomas A.
Edison, also, appears to have discovered the phosphorescence of calcium tungstate
when exposed to Rontgen's rays by deliberately trying the effects of these
rays on a large collection of different substances. This old prosaic method of
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 13
experimenting by trying everything is necessary in some cases, and, though usually
dubbed empirical or rule-of-thumb, the process is fundamentally scientific, but it
is not generally economical in time and labour. Discoveries are then due, as
J. Priestley once argued, more to " chance than to any proper design or preconceived
theory." More frequently, the track of the experimenter is blazed by means of
working hypotheses.
§ 7. Hypothesis, Theory, and Law
We are gifted with the power of imagination, and by this power we can enlighten the
darkness which surrounds the world of senses. Bounded and conditioned by co-operant
reason, imagination becomes the mightiest instrument of the physical discoverer. —
J, Tyndall.
The nearer to the practical men keep, the mightier their power. The theorist who
dreams a rainbow dream, and calls his hypothesis true science, at best is but a paper financier
who palms his specious promises for gold.- — T. L. Harris.
Hypotheses are cradle songs by which the teacher lulls his pupils to sleep. — L. W.
Goethe.
It is a popular belief that the aim of science is to explain things ; as a matter of
fact, the so-called explanations of science do not usually get much beyond describing
the observed facts in the simplest possible terms so as to make their relations with
one another clear and intelligible. i The description may emphasize the history of
a phenomenon, or the conditions under which the phenomenon occurs : In other
words, science may explain a phenomenon by describing how one event is determined
by an antecedent action — sometimes called a cause; and how one particular set of
conditions — the cause — can give rise to another set of conditions — the effect.
Science explains a phenomenon (the effect) by showing that it is a necessary or rather
a probable consequence of another phenomenon (the cause).
Classical scholars tell us that Aristotle has lorty-eight, and Plato sixty-four
meanings for the word cause. The later metaphysicians have also played a game
of shuttle-cock with the term. The word cause is usually appHed to an event,
action, or process which " produces " an effect ; or, with R. Shute, cause may be
regarded as that which the mind selects as a sign of the coming of that other phe-
nomenon which it calls the effect ; or conversely, an effect is regarded as something
which the mind selects as a sign of the past existence of a cause. There can there-
fore be no cause without an effect, and no effect without a cause. The one pre-
supposes and completes the other. Hence, as P. Carus has observed, the law of
causation describes a transformation in which form alone is changed ; and conse-
quently, the law of causation is nothing more nor less than another aspect of the
famous law of the conservation of matter and energy. The search for the cause of
an event is a search for the determining factors which would produce that event.
When the cause of an event has been discovered, the event is said to be explained
by the cause.
There are certain circumstances or conditions which may exercise, directly or
indirectly, a determinative influence on the effect produced by the activity of a
cause ; and very often certain conditions must obtain before an event can occur,
thus the temperature of hydrogen must be raised above its ignition point before
combustion can ensue. The effect obtained by burning hydrogen is more vigorous
if the flame be in oxygen gas than if it be in air. Hence, an atmosphere of oxygen
gas is a favourable condition for the combustion of hydrogen ; a reduced pressure
is a retarding condition because it hinders the speed of combustion and reduces the
vigour of the flame. The term cause is frequently employed when reason is mtended.
The difference is marked in different countries by the use of different terms— Greek :
ah-id (cause), alxv (principle, reason) ; Latin : causa, ratio ; French : cause,
raison d'etre ; German : Ursache, Grund ; Italian : causa, ragione ; etc. Gravita-
tion is said to be the cause of the falling of a vase from the mantelpiece, whereas
the cause of the fall may have really been a push from the elbow. In the former
U INORGANIC AND THEORETICAL CHEMISTRY
case, the reason why the vase fell downwards is the very same reason why all masses
gravitate, and a push was the real cause of the catastrophe. Here the reason of
the fall is referred to an inherent quality of bodies, just as the reason why bodies
react chemically is explained by investing matter with an inherent quality or vii<
occulta — chemical affinity. If these distinctions be borne in mind, there is no need
for confusing cause, reason, and condition, even if one term be used for all three
concepts.
The law of continuity — emphasized by G. W. von Leibniz (1687) — assumes that
no interruption between cause and event is possible, and that there is a connected
chain in the order of natural phenomena so that when several of the links are
known, the intermediate links can be inferred. Consequently, men of science
assume that each phenomenon is an efiect of a previous event, and is itself the cause
of a succeeding effect, and that under like conditions, the same causes produce
the same effects. Apart altogether from the question whether or not nature can
do precisely the same thing again under precisely similar circumstances as she has
done before, the principle of continuity or uniformity assumes that any phenomenon
will be repeated if all the preceding phenomena be precisely repeated ; otherwise
expressed : the same antecedents are invariably accompanied by the same conse-
quents. Hence, it has been said that science does not now seek for the reason or
the why of events, but rather for invariable relations between phenomena. The
law of causation is taken to describe a sequence of changes starting with the cause
and ending with the effect. G. Kirchhoff introduced the term description as a
synonym for cause at the very beginning of his Vorlesungen uber mathematische
PhysiJc (BerHn, 1876), where he said : " The object of mechanics is to give a complete
description in the simplest possible manner of such motions as occur in nature."
Although every effect may be traced to a previous event as its cause, in the
physical world, phenomena follow one another as links in an unbroken chain of cause
and effect. It is soon recognized that the cause of a phenomenon is an effect which
itself needs explaining by some ulterior cause, so that causes can be traced back-
wards in a never-ending chain of events. Owing to the limited range of man's
understanding in a world of infinite complexity, we are far, very far, from compre-
hending the true conditions, the true causes, or the true reasons for natural pheno-
mena.
The mind cannot receive a long series of details without encircling and con-
necting them by a common bond which is a kind of mental nexus ; similarly,
in the attempt to find the causes of many phenomena, man is compelled to build
an imaginary model showing how a given set of conditions — the hypothesis or theory
— is always followed by particular effects. A phenomenon is then explained by
showing that it is bound to occur by the operation of the set of conditions postu-
lated by the hypothesis. Consequently, hypotheses are essentially guesses at truth.
The rational observer does not trust to random guesses, but he is guided by a more
or less vague intuitive conjecture (hypothesis) as to the meaning of the phenomena
under investigation, and experiments are devised accordingly, for
Man's work must ever end in failure,
Unless it bear the stamp of mind.
The head must plan with care and thought,
Before the hand can execute.- — Schiller.
The Spanish philosopher J. L. Balmes emphasized this same idea m his Filosofia
/ondamental (Barcelona, 1846), when he said :
Although one accepts as a real truth the most uncontested and the most certain fact, it
remains sterile if ideal truths do not f ecimdate it. . . . To acquire scientific value, the facts
must become objective, or, being submitted to reflection, must be impregnated by the
mind with the light it lends to necessary truths.
Hypotheses precede observation and prompt experiments, for they indicate the
conditions under which the search for new facts is likely to be successful. Hence,
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 15
when Leonardo da Vinci (c. 1500) 2 said that " hypothesis is the general, and experi-
ments are the soldiers," he probably meant that hypotheses direct or indicate what
experiments should be made. Accordingly, hypotheses are indispensable aids in
the systematic quest after the secret meaning in nature's deeds. Those who refuse
to go beyond fact, said T. H. Huxley (1887), rarely get as far as fact. It is difficult
to believe that so astute an investigator as Joseph Priestley really overlooked this
niode of investigation, as might be supposed from some preceding remarks — nor
did he. On the contrary, he said :
It is by no means necessary to have just views, and a true hypothesis, a priori, in order
to make real discoveries. Very lame and imperfect theories are sufficient to suggest
useful experiments which serve to connect those theories, and give birth to others more
perfect. These then occasion further experiments, which bring us still nearer to the truth,
and in this method of approximation, we must be content to proceed, and we ought to think
ourselves happy if, in this slow method, we make any real progress.
The many gaps in our knowledge are temporarily bridged by the assumptions
called hypotheses. Hypotheses thus help to render intelligible the interrelations
between different facts, and they are employed by men of science to extend and
deepen their experience by predicting and disclosing new facts ; to correct and
purify their knowledge of natural phenomena by eUminating errors and contra-
dictions ; and to systematize their description of facts so as to obtain the greatest
control over them with the least possible effort.
An hypothesis contains a speculative term, an assumption which goes beyond the
observed facts ; while a law is a generalization which does not extend beyond the observed
facts. A law is thus limited by the facts it describes. When an hypothesis has been so
extended that it has a wide and comprehensive scope, the hypothesis becomes a theory.
Like the hypothesis, a theory usually contains an unproved assumption — e.g. the kinetic
theory, the electron theory, etc. Some writers — e.g.W. Ostwald — apply the term theory to
a generalization which does not extend beyond the observed facts, and in that case, theory
becomes law when the generalization has a wide and comprehensive scope. There are
several other uses of the term theory. For historical reasons the term may appear to be
confused because the passage from hypothesis to theory, or from theory to law, has not
always been attended by a change in the corresponding terms — e.g. Avogadro's hypothesis,
by the definitions here given, might be called a theory.
The verification of hypotheses. An hypothesis may seem to be the logical
consequence of known facts, or it may be a random flash of the imagination.
However probable an hypothesis might appear, both the hypothesis and the
logical consequences of the hypothesis must be tested by comparison with facts.
Aristotle (c. 320 B.C.) certainly recognized the need for basing reasoning on observed
facts, but, as G. H. Lewes (1864) has emphasized, Aristotle did not reahze the very
vital importance of verifying his logic by comparing its conclusions with facts, nor
did he recognize that the true purpose of experiment is to verify the accuracy of
data and of theoretical conclusions. We are indebted to Roger Bacon (c. 1280),
perhaps more than to any other, for first insisting on verification as the essential
pre-requisite for every trustworthy conclusion. He said :
Experunental science is the mistress of speculative science. She tests and verifies the
conclusions of other sciences. ... In reasoning we commonly distinguish a sophism from a
demonstration by verifying the conclusion through experiment.
Experiments have a way of giving results which differ from those which rigorous
logic concluded must occur ; and when the prediction fails, it is necessary to fmd
what has been overlooked. This does not mean that constant verification is needed
to establish the validity of the process of reasoning, for that may be irreproachable
and yet the conclusion may be false because the facts or premises upon which the
reasoning was founded may have been interpreted to mean something very different
from what actually obtains in nature, or because some unrecognized or undiscovered
factor was involved. It is not wise to dogmatize when direct trial is possible : " Do
not think," said J. Hunter, " try."
16 INOKGANIC AND THEORETICAL CHEMISTRY
It has been aptly said that the remarkable discoveries of modern science have
been made by invariably sifting the truth through a fine mesh of logical experiment.
One of C. Darwin's favourite methods of applying this method was to reason : "If
my hypothesis be true, then certain consequences must also be true. Now let us
find if they are true ; " and H. St. C. Deville used to say that there is no need to
argue if an experiment can be made. In fine, it is necessary to submit all con-
jectures to the incorruptible test of fact in order to avoid being seduced by im-
material creations of the imagination. Faith without facts availeth nothing. The
ad experiinentum test must be made with unremitting diligence, rigorously and
impartially, without conscious bias. Trial by a combat of wits in disputations has
no attraction for the seeker after truth ; to him, the appeal to experiment is the last
and only test of the merit of an opinion, conjecture, or hypothesis.
If one hypothesis does not fit the facts, it is discarded, and a modification of the
old, or totally new hypothesis is tried. Thus, J. Kepler, in his De inotibus stellce
martis (1608), is said to have made nineteen hypotheses respecting the form of
planetary orbits, and to have rejected them one by one until he arrived at that which
assumed their orbits to be elliptical. " To try wrong guesses," said W. Whewell,
" is apparently the only way to hit the right ones." This method of trial and
failure is continued until the golden guess crowns the investigation ; but one single
real conflict between fact and hypothesis will destroy the most plausible hypothesis.
Of fifty hypotheses, only one may prove fruitful ; the unsatisfactory ones are weeded
out, until that particular one remains which has established its right to live by
proving itself useful or by satisfying some need. Quoting M. Faraday :
The world little knows how many of the thoughts and theories which have passed through
the mind of a scientific investigator have been crushed in silence and secrecy by his own severe
criticism and adverse examination ; that in the most successful instances not a tenth of the
suggestions, the hopes, the wishes, and the preliminary conclusions have been realized.
This quotation may give a wrong impression,^ for Michael Faraday displayed
consummate skill, not only in framing hypotheses per se, but in deducing hypotheses
that were worth testing. Without hypotheses, the experimental method may
degenerate into empiricism ; without experiments, hypotheses may degenerate
into speculation.
The promulgation of immature or premature hypotheses without a substantial
basis of fact is discouraged by most scientific societies. The celebrated nebular
hypothesis was ushered in by P. S. de Laplace (1796) with those misgivings and
doubts which must of necessity becripple all hypotheses which are not based upon
observation or calculation. An hypothesis may be invaluable when it can be
verified or refuted by a definite appeal to observation. If this check be not possible,
the imagination riots in the wildest speculations. If the evidence of an alleged
phenomenon cannot be tested by verification, it is outside the range of science.
A. W. Hofmann is reported to have said that he would readily listen to any suggested
hypothesis, but on one condition — that he be also shown a method by which it might
be tested. Accordingly, scientific inquiry is limited to such objects and phenomena
as admit of direct or indirect observational or experimental verification. On the
other hand, science cannot enter into the dark territory beyond the scope of man's
faculties, and where verification, direct or indirect, is not possible. A vivid imagina-
tion can people this region with phantasms and be deluded with the hallucination
that these creatures of the imagination are real, substantial, objective facts. It
is now generally recognized that imagination, uncontrolled by facts, has produced
all the palsying superstitions which have blinded and cursed the human race — past
and present.
Rival hsrpotheses. — Two or more contradictory hypotheses may be consistent
with the facts ; both cannot be right. There is then need for an experimentum
crucis, an experiment which will decide in favour of the one and exclude the other.
An hypothesis is supposed to be established when it, and it alone, is in harmony
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 17
with known facts. The hypothesis then ranks as a theory or law. In the majority
of cases, the so-called laws of nature can be regarded as prophecies which becaui
they have always been fulfilled in the past, are expected to be also fulfilled in innu-
merable cases in the future. Laws, theories, and hypotheses are all on probation.
However successful a theory or law may have been in the past, directly it fails to
interpret new discoveries its work is finished, and it must be discarded or modified.
However plausible the hypothesis, it must be ever ready for sacrifice on the altar
of observation. On account of the unproved assumption embodied in all hypotheses,
they are of necessity transient, fleeting, and less stable than theories ; and theories]
in turn, are less stable than laws. A theory believed to-day may be abandoned
to-morrow. New facts need new laws. An hypothesis is invalid when it fails to
unite and coordinate facts. All our hypotheses and theories are to be superscribed
" subject to revision," for they are continually changing. " Science in making is
a battlefield of competing theories," the path of progress is strewn with dying and
dead hypotheses. For example, W. Ostwald (1893) claims that the theory of chemical
combination is a strange and contradictory conglomerate of the fossil constituents
of earlier hypotheses. Science is not a state, but is rather a stage of progress. Even
Isaac Newton's law of gravitation is included in this category ; and the astronomer
R. Ball 4 could say :
When the law of gravitation is spoken of as being universal, we are using language
infinitely more general than the facts warrant. At the present moment we know only that
gravitation exists to a very small extent in a certain indefinitely small portion of space.
Ever since T. Bergmann's time (1779), science has been compared with a building
in the course of erection, and scientific hypotheses have been compared with the
scafEolds and ladders required by the builder in order to place the stones of ex-
perience where they belong. The scaffolding must be rejected when it hinders
further developments, and when the purpose for which it was erected has been
fulfilled. Accordingly, an hypothesis is not the end, but rather the means of
attaining that end. To think otherwise would be to suppose that the builder
erects a mansion for the sake of showing off the ladders and scaffolds used in its
construction. The imperfect notions and hypotheses of men of science must not
be mistaken for descriptions of observed facts. In the chemica docens of our
schools, the term science usually includes both the growing building and the auxiliary
scaffolding ; otherwise expressed, the term includes the immutable facts, the
ephemeral hypotheses, the transient theories, and the more or less incomplete
generalizations from observations. The facts alone are certain to endure throughout
all time. When S. Brown (1849) inquired : Is it necessary to the nature of a science
that it be all true, and that it contains no admixture of error ? and answered : By
no means ! Otherwise chemistry was no science during the reign of phlogiston, and
the Lavoisierian chemistry no science so long as oxygen was taken for the principle
of acidity — he included in the term science those transient theories which are
necessarily employed in the erection of the temple of truth.
Deductive and inductive induction.— The term induction is applied by the
logician to the quest of science for generahzations, that is, for the camnes or uni-
versales regulce of Roger Bacon. In deduction, the attempt is made to widen the
bounds of knowledge without stepping outside known facts— the Euchdean method
is a good illustration ; in induction, a leap is taken from the known into the ilhmit-
able beyond. Two important methods of induction will be recognized— one may
be called the deductive method, the other the inductive method. The former was
favoured by Francis Bacon, the latter by Isaac Newton.
1. Bacon's deductive method, by what he called the interpretaho naturo'.
Here the facts are exhaustively classified until the generalization becomes clear,
a is either M or N, or 0, or P, or . . . ; but a is not N, nor 0, nor P, nor . . . ;
and consequently, a is 31. Thus, in the 105th aphorism of his Novum Organum
(London, 1620), F. Bacon said :
VOL. I. ^
18 INORGANIC AND THEORETICAL CHEMISTRY
The induction which is to be available for discovery and demonstration . . . must analyse
nature by proper rejections and exclusions ; and then, after a sufficient number of negatives,
come to a conclusion on the affirmative instances.
The method appears to proceed from known facts to general conclusions, a parti-
culari ad universale. It is based on facts already known, and has therefore been
called a priori reasoning. The method by which Boyle's and Charles' laws were
discovered might be cited in illustration of one form of the method of deductive
induction.
2. Newton's inductive inethod, by what F. Bacon called the anticipatio
naturoB. Here the attempt is made to infer the hidden generalization from the
consequences of the assumption (hypothesis) what that generalization is. The
process is sometimes called a posteriori reasoning. This method of investigation
was extensively employed with glorious results by Isaac Newton, although it had
been advocated by Aristotle tv/o thousand years earlier. Francis Bacon, indeed,
before Newton's time, protested against anticipating nature by hypotheses, but
the greatest triumphs of modern science have been won by the application of the
Newtonian method while the Baconian method has been singularly unfruitful.
Francis Bacon's failure in the practice of his own method was complete.
The particular form which the Newtonian method takes in science is to devise
provisional generalizations called hypotheses or working hypotheses to explain facts
and phenomena. The appeal is then made to observation and experiment in order
to test the validity of the proposed generalization. Examples : The cause of the
increase in the weight of metals calcined in air ; A. L. Lavoisier's theory of com-
bustion, and his experiments on the transformation of water into earth ; J. Mayow's
work on combustion ; etc. The application of this method of inquiry involves
(a) The accumulation of facts by observation and experiment ; (6) The employment
of the imagination in framing hypotheses to explain the facts ; and (c) The appeal
to facts to prove or disprove the hypotheses. By this procedure, said W. Whewell,
the hypothesis becomes the guide of its former teacher — observation. There is a
kind of cycle from facts to hypothesis, and from hypothesis to facts.
Induction, said Aristotle, does not prove. I. Newton's phrase : Hypotheses non
jingo — I do not frame hypotheses — is often quoted to show that he discountenanced
the inductive method of scientific investigation. This is based upon a misunder-
standing, for Newton here referred to hypotheses not suggested by observation.
On the contrary, Newton's own procedure was to use hypotheses deduced from
phenomena similar to the way science uses them to-day. Accordingly he
asserted that " no great discovery was ever made without a bold guess," and his
immortal Philosophice naturalis principia ^nathematica (London, 1687) is a wonderful
record of discoveries made possible only by the exercise of the greatest freedom in
the elaboration of hypotheses. Indeed, from the first of his communications on
light to the Royal Society to the last revision of his Principia, Isaac Newton seems
to have been steadily and persistently guessing.
The method of investigation employed in scientific, positive, or modern chemistry
thus involves four operations : (i) observation and experiment ; (ii) classification
and comparison ; (iii) deduction, or speculation and hypothesis ; (iv) testing and
verification. Francis Bacon did not grasp the prime importance of testing his
induction by comparison with facts. A. de Morgan (1872) ^ puts this rather
cleverly : According to Francis Bacon, facts are used to make theories from, and
according to Isaac Newton, to try ready-made theories hy. Chemistry could
progress as a science only when this method of investigation was discovered, so that,
as S. Brown stated in 1843, before discovering chemistry it was necessary to discover
the art of discovering chemistry.
References.
1 P. Carus, The Primer of Philosophy, Chicago, 137, 1904 ; Truth on Trial, Chicago, 1911 ;
R. Shute, A Discourse on Truth, London, 103, 1877 ; K. Pearson, The Grammar of Science,
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 19
London, li3, 1900; C. A. Mercier, On Cavsation, London, 1916; B. Russell, Myaticiem and
Logic, London, 1919.
2 H. Grote, Leonardi da Vinci als Ingenieur und Philosophy Berlin, 1874 ; P. Duhem, ttudea
sur Leonard de Vinci, Paris, 1906-1913 ; W. R. Thayer, Moniat, 4. 507, 1894.
» M. Faraday, Lectures on Education, London, 1855 ; Experimental Researches in Chemistry
and Physics, London, 486, 1859 ; G. J. Stoney, B. A. Rep., 243, 1879.
4 R'. Ball, Pop. Science Monthly, 23. 94, 1883.
^ A. de Morgan, A Budget of Paradoxes, Chicago, 1. 88, 1915.
§ 8. The History of Chemistry in China, India, and Chaldea
It is vain and ridiculous to attempt to trace the origin of chemistry to the first men who
worked in the metals, cut and polished stones, fluxed sand, or dissolved and crystallized
the salts. This would be analogous to an attempt to trace the elements of geometry in
the efforts of the savage to trim irregular fragments of rock to a more regular form in order
to adapt them to his first needs.- — ^A. F. de Fourcroy (1782).
There can be no doubt that the chemical arts had their origin in the darkness
before the dawn of history ; the very etymology of the word chemistry is lost in
obscurity. Many have been the attempts to fix a date at which chemistry began,
and as often have these attempts proved abortive. The names of mythological,
classical, and scriptural writers have been enrolled among the adepts, and as often
have these names been expunged from the list. What L. Blanc (1847) said of the
beginning of the Frencl^ Ke volution applies also to chemistry. Its history begins
and ends nowhere. The origins are so confused and the many facts known to the
ancients are so obscurely connected that there is no event which can be regarded
with certainty as a first cause.
The historians and antiquarians in chemistry now recognize how futile must be
the attempt to fix time or place for the birth of chemistry. They see that inquiries
can be profitably directed only in the attempt to find what particular form chemistry
took, or what particular ideas concerning chemical phenomena prevailed during any
given epoch. Thus, in his work Les origines de Valchimie (Paris, 1885), M. Berthelot ^
says :
Chemistry is not a primitive science like geometry or astronomy, because it is constructed
from the debris of a previous scientific formation which, half chimerical and half positive,
is itself founded on the treasure slowly accumulated by practical discoveries in metaUurgy,
medicine, industry, and domestic economy.
Evidence of an old prehistoric civilization, long prior to that indicated at the
beginning of the biblical record, has been laid bare during excavations in Egypt
and elsewhere. The antiquities which have been unearthed are arranged by
archaeologists in three successive periods — the stone age, the bronze age, and the
iron age. It is assumed that stone would be used by a rude savage people before
metal, and that copper, being oftenest found native, and readily hammered into
shape, would come into use before iron. This view was taken by Lucretius in his
De rerum natura (5. 1282) written about 60 B.C. He said :
The first weapons used by man were the hands, the nails, and teeth, also stones and the
branches of trees ; and then was discovered the power of iron and copper. The use of copper
was known earlier than that of iron, since copper is more abundant and easier to work
than iron.
Long before Lucretius, Hesiod (c. 700 B.C.) stated that the earth was first tiUed
with copper instruments because iron had not been discovered.
The three periods do not altogether represent divisions of tmie, but rather
stages of human culture, and they were not uniform in all parts of the world ;
rather did they merge more or less one into the other so that stone weapons were
used throughout the age of bronze, while bronze and iron were known m the stone
age ; and similarly, stone and bronze were used in the iron age. Hence this
classification is not altogether reliable historically, but it is so convenient that it
20 INORGANIC AND THEORETICAL CHEMISTRY
has been adopted by the leading museums in the world for the classification of
antiquities or ancient relics.
The Aryans. — Comparative philologists 2 who have studied the languages of the
different countries of Europe and Asia, have brought forward evidence in favour of
the theory that most of the European languages were derived from a family of
people speaking one language — now called the Aryan language—and that this
primitive language is also the source of much of the Indian, Iranian, and Armenian
languages. The common parentage is suggested by striking similarities in the roots
of many words in the languages of these different peoples. The evidence further
indicates that the primitive Aryan tongue was spoken by nomad herdsmen
wandering over the plains of Europe during the neolithic age, that is, when man had
learned to polish his flint weapons — very roughly about 6000 B.C. There is no
satisfactory evidence to prove that the Aryans were a civilized people which invaded
Europe from the East — as was once supposed. In time, the geographical continuity
of the primitive Aryans was disturbed and local variations in speech — dialects — ■
began to arise which ultimately were fractionally crystallized, producing the different
languages which now separate the different families derived from the original Aryans.
Owing to the absence of any common root for words connected with the smith's
craft, we are told that the arts of extracting and working the metals were developed
after the linguistic separation ; and for similar reasons, the philologists suppose
that the Aryans were not acquainted generally with iron, tin, or gold. Their
common knowledge of copper is supposed to be shown by i^e relation of the different
words — Sanscrit, ayas ; Gothic, aiz ; Latin, ces ; German, erz ; English, ore — for
the metal or its ore. The probability is increased by the fact that copper occurs
native in the metallic state. Some of the oldest metal implements, imitating the
older stone implements, found in old tombs and in the remains of pile-dwellings in
various parts of Europe, are of copper, not bronze. The knowledge of the metals
seems to have spread in Europe from the Mediterranean northwards, and is supposed
to have been introduced by Phoenician traders. The different stages of development
of the people, after the differentiation of the language, were not synchronous, since,
when one nation was in the stone age, another was in the bronze age, and a third in
the iron age.
Chaldea. — In Chaldea the remains of ancient cities and temples have been
ransacked, and the existence of another civilization before that of Egypt has been
revealed. The early Chaldeans must have been skilful workers in the metals over
sixty centuries ago — 4000 B.C. — and, since there were no mines and very little fuel
in the country, it is thought that the Chaldeans must have got some of their know-
ledge from another people more favoured in this respect. There is no written record
of early Chaldean chemistry, nor of any historical names in connection with their
chemical arts. Zoroaster (c. 1500 B.C.) is reputed to have been the founder of the
philosophy of the early Chaldeans and Persians. This subject, however, is very
obscure. Zoroaster is said to have made a very special study of the movements of
the planets. The cuneiform inscriptions show that the Chaldean wise men or
priests were practised in the arts of astrology, incantation, divination, and conjur-
ing. The number 7 appears to have been considered very important in their
philosophy and religion ; and the Chaldeans recognized this number of gods, devils,
planets, colours, metals, etc. The Babylonians established the divisions of time
which are employed to-day ; the seven days in a week thus originated from religious
and astrological considerations before 2300 B.C. The same number is sacred in the
Zarathustrian faith, the Mithras religion, and among the Buddhists, Jews, and early
Christians. A. Origen in his Contra Celsum (c. 22), says that the Persians repre-
sented the revolutions of the heavenly bodies by seven stairs which led to the same
number of gates each of a different metal — lead, tin, copper, iron, a mixed metal,
silver, gold. He added :
The leaden gate had the slow tedious motion of Saturn ; the tin gate the lustre and
gentleness of Venus ; the third gate of copper was dedicated to Jupiter ; the fourth, iron,
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 21
was dedicated to Mercury on account of its strength and fitness for trade ; the fifth, mixed
metal, to Mars ; the sixth, silver, to the moon ; and the last, gold, to the sun.
The astrological nomination of the metals has thus been traced to the Chaldeans,
and it appears to have been used by the Hindus, for F. Philostratus said in his
Vita Apollonii (3. 41), that the Brahmin larchas gave Apollonius seven rings named
after the seven planets ; one ring to be worn daily — each one on the day of the
week which bore its name.
The characters employed by the early writers to represent the planets were also used
for the corresponding metal, » but they were not agreed in the dedication of particular
metals to particular planets, and the characters themselves were subjected to certain
changes of form. Thus, G. F. Rodwell says that in a manuscript written by Antonio Neri
before 1613, mercury is designated by no less than thirty -five different names and twenty-
two symbols ; lead by sixteen names and fourteen symbols ; and sulphur by two names,
and sixteen symbols. The mythological symbols used largely by the alchemists of the
Middle Ages were :
0
9
6
11
h
v(
^
Gold
Copper
Iron
Tin
Lead
SHver
Mercury
Sun
Venus
Mars
Jupiter
Saturn
Moon
Mercury
J. Beckmann has suggested that these symbols are the remains of Egyptian hieroglyphics,
or else corrupted forms of the initial letters of the names of the deities which were supposed
to reside in particular planets ; and he claims that the symbol for copper 9 , said to
symbolize the looking-glass of Venu3, may really be a distorted form of the initial letter
of the Greek term ^aa-cpSpos for that goddess ; the so-called scythe of Saturn, a corruption
of the first two letters of his Greek name Kp6vos ; the imaginary caduceus of Merciuy, a
modified form of the initial letter of his Greek name ^Tifiwi which in the oldest manuscript
was written C or o with the next letter added below ; the lance and shield of Mars, €ui
abbreviation of the Greek name of the deity @ovpos, obtained by placing the last letter
above the first ; and the symbol for the thunderbolts of Jupiter was similarly derived from
the initial letter of the Greek equivalent Zeds for Jupiter with the last letter added below,
as is actually done in some of the older writings. The circle, the symbol for the sun, was
also the symbols of divinity and perfection. The semicircle for the moon is appropriate
since it is the only one of the heavenly bodies which appears in that form to the naked eye.
The following excerpt from K. Digby's Chemical Secrets (London, 1683) illustrates the way
the alchemists employed the symbols :
Take good mineral 9 , mortifie it with radicated vinegar ; then separate its
quintessence with pure S.V. ; with that quintessence, dissolve ^ duplicatiun of 9 >
that both become an oyl, which unite with a subtle calx of 0, and bring them to an
incombustible oyl, which will transmute ^ into 0.
Hence, astrology, and the emphasis which the alchemists later sometimes laid on
the number 7, are relics of Chaldean thought. The Chaldeans supposed that the
planets influenced the properties of the metals, the organs of the body, and the
destiny of man.
The Chaldeans seem to have had some knowledge of metallurgy, dyeing, weaving,
the manufacture of colours, glass, and the imitation of gem stones. The chemical
arts practised by the early Chaldeans were probably adaptations of chance observa-
tions to useful purposes ; these arts gradually drifted to the early Egj^ptians.
For instance, it is related that Abraham came from Ur in Chaldea {Gen., IL 31),* and
he probably brought a higher civilization into Canaan, and also to Egypt. The
Egyptians developed and improved the Chaldean arts in the laboratories and
workshops attached to their temples. When the Babylonian empire ceased to
exist, the Chaldean nation was dispersed, and the priests were scattered over the
neighbouring lands, so that the term Chaldean became a by-word synonymous with
" a wise man from the East." The scholars also tell us that the Assyrian rab-mag
or the Semitic ma^— meaning a priest— has furnished the Latin and European
language with the terms fnagus, magic, and magician.
India.— India played no direct part in the development of Western science.
It is a tradition that Hermes the Egyptian predicted that naught of the histor}^ of
Egypt, but the letters engraved on stone, would survive. W hether this be true or
22 INORGANIC AND THEORETICAL CHEMISTRY
not, scholars are now mainly dependent upon the inscriptions on tombs and monu-
ments for their knowledge of the early Chaldeans and Egyptians. On the other
hand, what remains of Indian thought is recorded in their books — the Vedas, the
Charaka, and the Susruta — for the Indians were a literary people. According to
Max Miiller, there are many points in common between the early Greek and Indian
philosophers, and there is a historical possibility that the Greeks were influenced
by Indian thought travelling through Persia. From this point of view, it was only
when commerce had opened up the country that it became possible to recognize
the debt which European science owes to India, and to find that a great deal
formerly attributed to the Arabians was of Indian origin. The learning of Greece,
Persia, and India is said to have been taxed to help the sterility of the Arabian
mind.
In his History of Hindu Chemistry (London, 1902), P. C. Ray has shown that
Indian chemistry developed largely on independent lines — medicine, not the metals ,
was mainly emphasized. The contact between the Hindus and Persians is thought
to have given the latter a bias towards medicine which later showed itself in
the polypharmacy of the Arabians. Very fair accounts of the philosophical views
of the Hindus are available. H. T. Colebrooke,^ for instance, has shown that an
early philosopher Kanaka developed an atomic theory rivalling that of Lucretius
(60 B.C.). The Hindus also developed a five-element theory of the constitution of
the world, but the elements of the Hindus were not the same as the quintet — air,
earth, fire, wood, metal — of the early Chinese. The Hindu quintet embodied : Water,
the first thing created ; the sacred fire ; the unbounded cether ; the foster-mother
earth ; and the air which animates all living beings. The idea that water is the most
primitive element of all is found in many of the classical books of India.
The Vedic hymns, over 1000 B.C., personify the elements and natural phenomena
— for instance, they raise the active principles of plants to the dignity of gods.
The medical work Charaka, and the more recent Susruta, seem to be repositories
of information — chemical and therapeutical — which had accumulated between the
Vedic period and approximately the eighth century. Gold, silver, copper, iron,
tin, lead, as well as some varieties of brass were known. The writers also mention
a number of salts of the metals — e.g. alum, copperas, sodium and potassium carbon-
ates, and a few products of the mineral kingdom. The Hindus developed some
alchemical notions, but they directed their attention mainly to medicine. According
to P. C. Ray, the practice of chemistry between the twelfth and thirteenth centuries
was distinctly in advance of that of the same period in Europe. The Hindus
learned about zinc and mercury ; but, as L. Hoefer has pointed out, real progress
in chemistry in India and China was not possible so long as the preparation of the
mineral acids was unknown. These acids are incidentally described in works
dating from the sixteenth century.
The arts and sciences were largely cultivated by the higher classes, but, according
to P. C. Ray, when the caste system was established, the opinion that industrial
work tends to lower the standard of thought, which at one time threatened Europe,
seems to have likewise developed in India with disastrous results. The arts and
sciences were relegated to the lower castes, the spirit of inquiry gradually died, and
the artisan classes, guided solely by their mother wit and common sense, alone
kept up the old traditions. The withdrawal of the intellectual community from
active participation in the arts rendered India " unfit for the birth of a Boyle, a
Descartes, or a Newton, and her very name was all but expunged from the map
of the scientific world."
China. — So far as we can gather, the Chinese were civilized, and cultivated
the arts and crafts at a time when the European nations were barbarians. Some
scholars claim that there is strong evidence of a Western origin of Chinese civilization,
and that the first Chinese settlers came from a country in the far West which was
closely connected with the founders of Babylonian culture. The earliest docu-
mentary evidence of man's attempts to answer the question : From what are
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 23
all things made ? occurs in the Chinese work Shoo King, which is said e to contain
a document called The Great Plan purporting to have been given by heaven to Yii
the Great, about 2200 B.C., and which is considered by the scholars to be older than
Solomon's writings. Here reference is made to five elements— water, fire, metal,
wood, earth. The Chinese element wood was never recognized as an element in the
West. The early Chinese philosophers supposed that two elementary principles—
yang (male or active) and yin (female or passive) — were derived from
t'ai kih~the Great Origin of the Grand Cause. The two principles lur.^-'-.'
yang and yin gave rise to the five elements. Fire and wood belong '■'■'^•---
to the yang element ; water and metal to yin ; and earth is neutral. *^"'"''--
The union of the five elements produces yang and yin : and the union ^•'^'*'—
of these two principles produces the grand cause which is itself without Fio. 2.— Stupa
cause. The Chinese Buddhists symbolized the five elements by the ^onn of the
square (earth), circle (water), triangle (fire), crescent (air), and the gem ?^j^T^^?"
(aether) —as indicated in Fig. 2 . In the mediaeval European symbolism, j^\, ^ ^ y^,
the two latter figures are treated as one, and serve as the common ments.
symbol of air. It is said that all over those parts of Asia dominated by
Chinese civilization, stupas or monuments built in the general shape of the symbols
of the five elements are to be found — e.g. the gateway to the Buddhist monastery
of Pekin, etc.
The philosopher Lao-tze, founder of the Taou religion in the sixth century B.C.,
believed that a fine essence or spirit arising from matter may become planets and
stars ; and these speculations led to the search after the sublimated essence of
things. The Taouists sought some flux which would purge man from the dross of
animalism and leave the higher part of man's nature to be crystallized out and
sublimed into some stable and eternal form. They had no success in finding an
elixir of life, or philosopher's stone ; but they obtained a number of fairly pure
mercurial preparations. According to J. Adkins (1855), the earliest Chinese work
on alchemy now extant is the second-century treatise Chen tung chi, by Wei Peh
Yang ; and two centuries later, P'au P'on Tsze wrote many works on alchemy and
kindred subjects. Later still, a disciple of Lao-tze — Wei Poh Yang — wrote a book
The Uniting Bond in which reference is made to a red elixir which was probably
mercuric sulphide or vermilion, prepared from galena or lead ore — symbolized by a
white tiger— and mercury — symbolized by a blue dragon. The red elixir was
regarded as an elixir of life even though Wei Poh Yang appears to have been poisoned
when he attempted to practise his own philosophy. There seems to have been
some contact between eastern and western Asia during the seventh-century invasions
of the Mahomedans, and the teachings of the Taouistic alchemists penetrated
Arabia, and appear there as the philosopher's stone and the elixir of life.
W. A. P. Martin,7 in a chapter on Alchemy in China (1901), considers that the
alchemy of China is not an exotic but a genuine product of the soil of that country ;
alchemy is indigenous to China, and coeval with the dawn of letters. He under-
stands the words alchemy and chemistry to represent different stages in the progress
of the same science, and says that the skill of the Chinese in the chemical arts and
their knowledge of many chemical compounds give evidence of lives passed among
the fumes of the alembic. Whatever be the true facts, there can be little doubt
that the early Chinese practised the chemical arts somewhat extensively, and they
knew quite a long list of chemical preparations — nitre, borax, alum, corrosive
sublimate, arsenic, mortars, cements, oils, paper, sugar, etc. They appear to have
invented printing, the manufacture of paper, and gunpowder— or rather a kind of
Greek fire which was placed in vessels, ignited, and projected from a throwmg
machine. They were acquainted with various precious stones ; some of their
pottery has never been surpassed. Chinese porcelain seems to have originated
about the time of the Han dynasty— 206 B.C. to 220 a.d.— and it attained its highest
development under the Ming dynasty— 1368-1644. Glass was made m China in
the Wu Ti dynasty~422-455— and was probably derived by contact with Western
24 INOEGANIC AND THEORETICAL CHEMISTRY
nations. They knew about gold, silver, mercury, lead, copper, iron, zinc, nickel, and
various alloys. The method for making zinc was probably derived from India.
They seem to have had ideas about the transmutation of the base metals into gold ;
and they are credited with a knowledge of oxygen and the composition of water
as early as the eighth century. All this, however, exerted no direct influence on
the development of European chemistry, although there is much evidence to show
that indirect communication between Europe and China was possible — e.g. the
Arabian alchemist Avicenna is said to have been born at Bokhara on the borders of
the Chinese empire. From the time of Confucius, the Chinese made little progress
in the arts and sciences, while Europe rapidly grew in knowledge.
References.
* H. Kopp, Geschichte der Chemie, Braunschweig, 2. 3, 1844 ; Beitrdge zur GeschicMe der
Chemie, Braunschweig, 40, 55, 1869 ; M. Berthelot, Introduction a V etude de la chimie des
anciens et du moyen age, Paris, 1889 ; T. Bergmann, De primordiis chemice, Upsala, 1779 ;
E. Cullen's trans., Edinburgh, 1791 ; H. Boerhaave, Elementa chemio',, Lugduni Batavorum, 1732 ;
P. Shaw's trans., London, 1753 ; F. Hoefer, Histoire de la chimie depuis les temps les plus recules
jusgu'd notre epoque, Paris, 1842 ; J. C. Brown, A History of Chemistry, London, 1913.
2 O. Schrader, Sprachvergleichungen und Urgeschichte, Jena, 1907 ; T. Taylor, The Origin of
the Aryans f London, 1892 ; F. M. Miiller, Biographies of Words, and the Home of the Aryas,
London, 252, 1888.
3 G. F. Rod well, PM. Mag. (4), 35. 1, 1868 ; J. Beckmann, Beitrdge zur Geschichte der Erfind-
ungen, Leipzig, 1780-1805 ; A History of Inventions, London, 1814 ; P. Carus, Open Court, 15.
335, 412, 1901.
* F. Hommel, Der habylonische Ursprung der dgyptischen Kultur, Munich, 8, 1892 ; Z. A.
Ragozin, Chaldea from the Earliest Times to the Rise of Assyria, London, 1886 ; G. Radet, La
Lydie et le monde grec au temps des Mermnudes, Paris, 1893 ; I. P. Cory, Anx:,ient Fragments of
the Phoenician, Chaldean, London, 1836 ; V. E. Johnson, Chaldean Science, London, 1896 ;
H. V. Hilprecht, Excavations in Assyria and Babylonia, Philadelphia, 1904; A. H. Sayce, Babylonians
and Assyrians, London, 1900.
^ H. T. Colebrooke, Essays on the Religion and Philosophy of the Hindus, London, 1853 ;
Ia Mavilleau, Histoire de la philosophic atomistique, Paris, 1 895.
« J. fl. Gladstone, B. A. Rep., 448, 1883 ; J. Adkins, Journ. Roy. Asiatic Soc, 18. 1, 1856 ;
F. P. Smith, Amer. Chemist, 4. 46, 1873 ; A, Wylie, Notes on Chinese Literature, Shanghai, 1867 :
P. Carus, Chinese Philosophy, Chicago, 1896 ; Chinese Thought, Chicago, 1907 ; Monist, 15. 500,
1905.
7 J. Klaproth, Mem, Acad. St. Petersburg, 2. 476, 1810 ; C. W. Duckworth, Chem. News, 53.
260, 1886 ; H. Chatley, Journ. Alchem. Soc., 2. 33, 1913 ; W. A. P. Martin, The Lore of Cathay,
Edinburgh, 1901 ; The Chinese, their Education, Philosophy, and Letters, New York, 1901 ; P. M.
Cibot, Memoires concernant Vhistoire, les sciences, les moeurs, les usages des chinois, Paris, 1776-
1814; E. Soubeiran, Journ. Pharm.Chim.,{5), 13. 213, 1866; H. J. Holgen, Che7U. Weekblad,
14. 400, 1917; H. C. Bolton, Chem. News, 70. 53, 1894.
§ 9. The History of Chemistry in Egypt
Let us confess at once, without going round the subject, that practical chemistry took
its rise in the workshops of the smith, the potter, or the glass-blower, and in the shops of
the perfumer ; and let us agree that the first elements of scientific chemistry date no
further back than yesterday.- — -J. B. Dumas.
According to Diodorus Siculus' report i of his visit to Egypt — Bibliotheca
historica (c. 30 B.C.)— during the reign of Julius Caesar, the Egyptians regarded
Hermes Trismegistus as a man 2 to be esteemed above all others for his penetrating
genius in discovering everything that could be useful in life ; and it was the favourite
opinion of the Arabian and European alchemists in the Middle Ages, that this
Hermes laid the foundations of chemistry about the time of Moses. Hermes was
accordingly called the father of philosophy and of alchemy by the alchemists of
the Middle Ages — e.g. by Albertus Magnus (c. 1250), Koger Bacon (c. 1250), etc.
It was also said that before the time of Hermes, the transmission of knowledge
from one generation to another depended upon oral tradition, but Hermes invented
a system of recording events upon stone pillars in the same way that modern
writers employ parchment or paper ; consequently, engraved pillars were the
THE EVOLUTION AND METHODOLOGY OP CHEMISTRY 25
standard literature of the day. Some of those who now appear to be the more
credulous writers of early history, state that Hermes inscribed upon an emerald
the most essential secrets of alchemy, and presented this jewel to Sarah the wife
of Abraham ; and that after many subsequent adventures the stone was lost ;
they also state that a copy of the inscription survives. From the translations
which have been made of the supposed inscription it appears that even if the
inscription itself be not lost, its meaning has gone. The alchemists honoured Hermes
when they spoke of the hermetic sealing of a vessel.
Attempts have been made to identify Hermes Trismegistus with the Egyptian
king Siphoas or Memnon, who had the surname Hermes, and also with various
biblical celebrities — ^Adam, Cain, Enoch, Joseph, Moses, and Abraham.
Some writers maintain that Hermes Trismegistus is a fabulous personage,
and it is generally supposed that this Hermes was identical with the
Egyptian god Thoth — ^literally a pillar. Thoth is represented by the
Egyptians as an ibis-headed god with a pen in his hand, the tutelary
deity of wisdom and letters, Fig. 3. It is further said that the supposed
writings of Hermes really cover three successive epochs — the first
Hermes dealt with the period down to the deluge ; the second Hermes
was concerned with early traditions ; and the third Hermes embodied
the full-grown science of Egypt. The whole system of Thoths or pillared
literature was personified as Hermes Trismegistus — rpis, thrice ; /teyto-ro?,
greatest — meaning literally thrice great interpreter. It is theref ore ^'^- 3.—
easy to understand how Hermes might have been credited with being J^ q^
an extraordinarily prolific author. Thus, in his De mysteriis ^gyft (c. Thoth.
360 A.D.), lamblichus says that Hermes was the author of 36,525 books
— T. Bergmann (1779) laconically observes that, if so, the books must have
been very concise after the manner of those times, and that each book could
have contained but a few sentences. Indeed, in his Stromaia (c. 200), Clement
of Alexandria describes imposing celebrations in which the books of Hermes were
borne in processions.
Most of the writings attributed to Hermes appear to have been lost at the
destruction of the Alexandrian library ; a few passages are quoted from them by
Zosimus (c. 400) ; and copies of some dealing with burial rites and the future life
have been found buried with the mummies of kings and priests ; these have been
embodied in what is now called The Book of the Dead. In general, it has been said
that Egyptian thought was heavily hampered and severely restrained by a powerful
priestcraft ; that the people were haunted by dread and dismal shadows from the
underworld ; that they fostered an elaborate cult of the dead ; that their houses
were temporary abiding places ; while their tombs were their eternal homes.
Herodotus (c. 440 B.C.) believed that the early Egyptians were the wisest of men.
He said that they had three communities of priests— at Heliopolis, Memphis,
and Thebes. The priests were responsible for the preservation of such knowledge
as was considered worthy of being retained ; this knowledge was kept secret and not
divulged except to the elect. The sacerdotal secrets were in part described by
hieroglyphics on stone pillars, and on manuscript papyri, but the allegorical nature
of the symbols prevented them being read or understood by the unimtiate^.
According to lamblichus (c. 360), every discovery which was approved by the
priests was engraved, without the author's name, on stone pillars in the temples
Clement of Alexandria, Plutarch, and others say that the priests possessed still
more secret writings. No original record of the early writers is avadable, and our
knowledge of the practice of the Egyptian arts is gleaned from fragments in the
writings of Pliny (c. 23), Plutarch (c. 100), C. Galen (c. 190), etc.
About 332 B.C., while Egypt was under Greek rule through conquest by Alexander
the Great, the Greeks were received in the Academy of Alexandria ; and some ot
the Egyptian manuscripts were translated into Greek, and later on distributed oveT
Europe— Paris, for instance, among others, has one by Zosimus ; the bt. iMarK
26 INORGANIC AND THEORETICAL CHEMISTRY
manuscript is preserved at Venice ; and a number are reported to be at the Vatican
in Rome, the Sultan's museum in Constantinople, etc.3
The Rhind rmiihematical papyrus in the British Museum is the main source of
our knowledge of the early Egyptian mathematics. It is considered to be a copy
made about 1600 B.C., by an Egyptian priest, from a document seven hundred
years older. Researches near Memphis have given indications of Egyptian medical
practices 4500 B.C., and pictures of surgical operations of a date not later than
2500 B.C. have been found. The celebrated Georg Ehers' papyrus * was found in
the winter of 1872-3, near Memphis, in a terra-cotta vessel between the legs of a
mummy which was buried about 10 ft. deep. The papyrus is supposed to be a
copy of one of the six medical papyri of Hermes (c. 1550 B.C.) about the time of Moses,
and the text refers back to kings who reigned 3700 B.C. The papyrus is a kind of
materia medica and medical treatise ; it gives some directions to the medical
attendant of a sick person, and describes the necessary incantations and invocations
for the co-option of the help of the gods. The papyrus also contains references to
a number of metals and some compounds.
A portion of one of theearliest Egyptian manuscripts is preserved in the museum at
Leyden, and is known as the Leyden papyrus. It was found enclosed in the wrappings
of a mummy at Thebes, and is considered to have been written about the third
century. It was presented to the Netherlands by I. d'Anastasi, the Swedish consul
at Alexandria in 1828. The work contains over a hundred magic formulae, and
recipes for the preparation of alloys used in making various objects of the goldsmith's
art. It also has drawings of some chemical apparatus. It has been investigated
by C. J. C. Reuvens, M. Berthelot, etc.^ The grammatical errors and spelling have
led to the opinion that the papyrus must have been the memorandum book of an
uneducated artisan engaged in attempts to imitate gold and silver for fraudulent
purposes — e.g. the preparation of asem, an alloy of copper and tin, occupies a
prominent place among the recipes for imitating gold. The Royal Swedish
Academy of Stockholm also acquired a papyrus about the same time and from the
same source as the Leyden papyrus ; but the existence of the Stockhohn papyrus
seems to have been overlooked until about 1906. It was translated by C. 0.
Lagercrantz in 1913.^ It deals with the diplosis of silver, the imitation of precious
stones, and dyes.
These papyri are supposed to represent the class of books on the chemistry
of gold and silver which, according to Suidas' Lexicon (c. 1000), were burned by the
order of the Roman emperor, Diocletian, about 290 a.d., as a supposed punishment
for an attempted rebellion, and to prevent the Egyptians making gold, and so
acquiring wealth sufficient to enable them to oppose the authority of the Romans.
These incendiary forays on the books of a prohibited and feared art — alchemy —
were not infrequent in the early Christian era — witness The Acts of the Apostles
(19. 19). This helps to explain the paucity of the early records of Egyptian science ;
yet, in spite of various conquests of Egypt by the Persians, Babylonians, Greeks,
and Romans, the arts were cherished by the priests with more or less vigour until
the Saracen invasion of the seventh century, when every abode of learning, and
every monument of science was destroyed with a ruthless hand. In 642 a.d., the
famous Alexandrian Library, with its 700,000 books, was condemned to destruction
by Kaliph Omar, who, in refusing a petition to spare at least a part of the library,
is reported to have said : "If the books agree with the Koran, they are useless ;
and if they differ from it they are dangerous." A mania for pillage and destruction
with the idea of terrorizing the stricken inhabitants of a conquered territory, has
long been characteristic of the temper of invading barbarians in ancient and modern
times — witness the invasion of Europe by the Goths, the Vandals, and the Huns
early in the Christian era, and the more recent rape of Belgium and North France
by Teutonic hordes maddened with the lust of a world's conquest. Egypt never
recovered from the severe blows she received, and what was presumably the greatest
treasury of knowledge garnered by the ancient world, was used for kindling the
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY
9.7
fires of the baths of the invaders. l*robably a few volumes were surreptitioiwly
preserved ; others, said to have been saved from the plunderers, are probably
forgeries.
The Egyptians appear to have been acquainted with some six or seven metals
and with some alloys. 7 Various metallurgical processes for extracting or melting
metals have been depicted in their tombs, etc. — Fig. 4, for instance. The Egyptians
were well versed in the arts of glass-making, potting and the manufacturing of
precious stones and enamels ; they were familiar with the arts of dyeing, painting,
tanning, brewing, and baking ; they were acquainted with many poisons and their
antidotes, and with expressed and distilled oils. They were highly skilled in the use
of antiseptics — particularly in the embalming of the dead. Among the operations
employed in the arts and crafts of the Egyptians were : calcination, digestion,
decoction, distillation, expression, evaporation, fusion, fermentation, levigation,
and sublimation. So far as the available records go, there is nothing to show that
any results were obtained by experiments directed as deliberate questions to
nature. To know that liquids boil and evaporate, or that metals fuse and form
calces, may indicate an unconscious sagacity in observation, but it is not scientific
observation ; the early arts, said W. Whewell,^ were the parent, not the progeny
of science.
So far as we can learn, the disjointed knowledge of technical processes, so
jealously guarded by the Egyptian priests, was purely empirical, and it required
Fig. 4.— Gokl Washing, and the Fusion and Weighing of the Metal aa depicted in an early
Egyptian Tomb.
centuries of eflort before man learned to view these processes in a comprehensive
rational way. Evidences of the practice of these chemical processes are found in
ancient monuments and tombs of high antiquity — see, for example, R. Lepsius*
Die Metalle in den agyptischen Inschriften (Berlin, 1872)— and Fig. 4 represents the
washing, fusion, and weighing of gold as is reported to have been depicted on an
old Egyptian tomb. The records are too imperfect to form a clear idea of Egyptian
science, if they really had one such. There is some fragmentary evidence, more
or less confused by fictions, and disguised by personifications, that the Hermetic
writings assumed that all substances are produced from two elements: fire, the
spirit of the world ; and tnortuum malum, the inert matter of the earth— that is,
energy and matter ; although, according to Seneca's Qucesiiones natmales (3. 14,
c. 63 A.D.), the Egyptians adopted an extended form of the four-element theory in
which each element had an active (male) or passive (female) ioim—e.g. active air
was the ivind, and passive air the at7nosphere ; flame was active fire, and hfjht,
passive fire. According to Diodorus the Sicilian (c. 30 B.C.), the Egyptians taught
that by some internal changes, all bodies sprang from their seeds or atotn^,
and were changed, perfected, and then destroyed.
Consequently, the impression that Egyptian chemistry was mamly practical
recipe and unverified speculation, is well founded on pertinent evidence ; but
others claim that too little inside knowledge is available to justify speaking witn
any confidence. In any case, it will be clear that before the Christian era Egypt
must have been a kind of focus or centre which collected, assimilated, extended, and
developed knowledge derived from various Eastern sources ; otherwise expressed,
28 INORGANIC AND THEORETICAL CHEMISTRY
the rise of chemistry in Egypt can be compared with a river which drains a large
tract of territory, there is not one source, but many sources, each feeding a tributary
of the river.
Phcenicia. — It is fairly clear that the indefatigable merchant Phoenicians had
acquired some knowledge of the so-called chemical arts during their contact with
the Egyptians. The Phoenicians were famous for the manufacture of a purple dye
— Tyrian purple — the special boast of Tyre ; for the manufacture of glass — Sidonian
glass — particularly at Sidon ; for the weaving of fabrics of various kinds ; for
working in metals; and for the engraving of gems (// Chronicles, 2. 14). The
Phoenicians were great navigators, and it is supposed that they circumnavigated
Africa. Strabo says that they made a special study of astronomy and arithmetic.
Posidonius, a Greek writer of the first or second century B.C., made a special study
of Phoenician mining, and gathered his data from the remains of the Phoenician
mines in Spain.^
The biblical record. — The biblical records of the unfortunate Israelites show
evidence of the chemical arts and crafts employed by their Egyptian masters.
The Israelites must have carried much of this knowledge into Asia during their
exodus from Egypt under the leadership of Moses. Even from the beginning of
Genesis, we are told that Tubalcain (3870 B.C.), the eighth man from Adam, was a
worker in metals {Gen., 4. 22) ; good gold is said to have been obtained at Havilah
{Gen., 2. 11), and silver coins were in use at the time of Abraham {Gen., 23. 16) ; in
all, about six metals — gold, silver, copper, iron, lead, and tin — were known to the
IsraeUtes {Numb., 31. 22) ; Noah made wine from grapes {Gen., 9. 21) ; and vinegar
was in use {Numb., 6. 3) ; bricks were burned for the building of the tower of Babel
{Gen., 11. 3) ; weaving and dyeing were known {Exod., 26. 1) ; and oils, perfumes
{Exod., 30. 23), and butter {Gen., 18. 8) were manufactured.
The mechanical performance of operations essentially chemical in their nature
is not chemistry, otherwise the first man to light a fire, boil a rabbit, or roast a pig
was the father of chemistry. It is difiicult to see why the mere practice of these
arts should be taken to prove that the early artisans were chemists in all but name,
unless there is some collateral evidence of scientific procedure in the development
of the empirical crafts — roasting and boiling, baking and brewing, or potting and
dyeing. Moses' demonstration lo of the solubility of Aaron's golden calf {Exod., 32.
20) has been taken to show the profundity of the chemical knowledge he must have
acquired during his tuition by the Egyptian priests ; but, before the indignant
prophet can be credited with any profound knowledge of chemistry, more details
are required. Well might Francis Bacon, in his Novum Organwn (London, 1620),
protest against the vanity of the attempt to found science upon the scriptures.
In Exodus (31. 3) we are told that Bezaleel, the son of Uri, was endowed with
the spirit of the Lord, and with skill to work in metals and precious stones. These
hints of the early arts have been expanded by surmise and guess, and deformed by
fiction and fable. For instance, it has been said that man received his first knowledge
of the arts and sciences by divine or diabolic revelation. In his Chronicorum
canonum (c. 300), P. Eusebius tells us that the apocryphal Enoch was taught by
the angels, and transmitted his divine revelations orally, through Methuselah (c. 3300
B.C.) and his descendants down to Abraham. From the writings of Zosimus (c. 400),
it would appear as if there was once a race of amorous genii prone to fall in love
with women, for he says that the secrets of nature were revealed by such genii to
the daughters of men in return for love. The dowry was called Trapdoymv Oitav- the
divine tradition ; the first account of these revelations was called xvf^"^ (chema) ;
and the art itself, xvi^'-^ (chemia). Chema is thus an early tradition respecting the
operations of nature taught to mankind by angels, who appear to have been damnati
a Deo for their ill-timed loquacity. The credulous and imaginative O. Borrichius,
in his De ortu et progress^ chemim dissertntio (Hafniae, 1668), said that the angels
or demons here mentioned were the offspring of Seth and of Tubalcain, who had
been instructed by their progenitors in the mysteries of nature, and who profaned
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 29
their trust by communicating heavenly secrets to the daughters of Cain bv whose
charms they were seduced (6^en. 6 2-4). The gynecial myth cf the origin of chemistry
recalls the Jewish story of the fall of man, and also th6 Grecian legend of the Sibvl
who demanded both length of years and a knowledge of the divine arcana as the
price of her favours to Apollo the sun-god. Somewhat similar mvths are reported
to have been current among the Phoenicians, the Persians, and the Egyptians
They illustrate the extreme creduHty of man in the first of Comt^'s les trois etals'
when everything that is not understood is believed to have a supernatural origin '
Repeeences.
1 G. Booth, The Historical Library of Diodorus Sicvlus, London, 1721.
2 R. Pietschmann, Hermes Trismegistus nach oegyptischen, griechischen, und orientcUischen
Ueberheferungen, Leipzig, 1875; L. M6nard, Hermes Trismegistus, Paris, 1867; Alethophilo
Hennetis Trismegisti, Stuttgart, 1855. '
3 For lists of these manuscripts, see H. Kopp, Beitrdge zur Oeschichte der Chemie, Braunschweiir
2. 243, ] 869. *'
^ E. 0. von Lippmann, Abhandlungen und' Vortrdge zur Oeschichte der Naturwi^senschaften
Leipzig, 1. 1, 1913 ; H. C. Bolton, Amer. Chemist, 6. 165, 1875 ; F. H. Garrison, An Introduction
to the History of Medicine, Fhila,de\phia, 191S ; H. Schaeffer, Z)ie Alchemie ; ihr dgyptisch-gricch-
ischer Ursjnung und ihre weitere historische Entwicklung, Flensburg, 1887.
5^ C. J. C. Reuvens, Lettres a M. Letronne sur les papyrus bilingues et grecs du Musee d'ArUiquitis
de V Universite de Leide, Leide, 1 830 ; M. Berthelot, La chimie des £gyptiens d'apres les papyrus
de Leide, Paris, 1886 ; H. Kopp, Beitrdge zur Oeschichte der Chemie, Braunschweig, 1. 97, 1869.
^ C. 0. Lagercrantz, Papyrus Gtcbcus Holmiensis, Upsala, 1913.
' V. E. Johnson, Egyptian Science, London, 1891.
8 W. Whewell, History of the Inductive Sciences, London, 1. 253, 1857.
^ G. Rawluison, Phoenicia, London, 1888.
1" W. Herapath, Phil. Mag., (4), 3. 528, 1852 ; J. D. Smith, ib., (4), 4. 142, 1852 ; H. Kopp,
Beitrdge zur Oeschichte der Chemie, Braunschweig, 2. 400, 1869 ; J. Napier, Manufacturing Arts
in Ancient Times, with special reference to Bible History, Paisley, 1879.
§ 10. The History of Chemistry in Greece and Rome
They had visions. Oh ! They were as grand
As ever floated out of fancy land.
From the testimony of Diodorus the Sicilian (c. 30 B.C.), Clement of Alexandria
(c. 200 A.D.), and lamblichus (c. 350 a.d.), it would appear that the Greeks learned
the practice of the chemical arts and crafts largely from the Egyptians. i Diodorus
says in his Bihliotheca historica (c. 30 B.C.) :
Orpheus, Musaeus, Melampus, Daedalus, Homer, Lycurgus, Solon, Plato, Pythagoras,
Eudoxus, and Democritus the Abderite all went into Egypt, and they doubtless learned
there all those things which rendered them afterwards famous among the Greeks. For
thirteen years Plato and Eudoxus associated with those priests in Egypt who most excelled
in the knowledge of celestial things. They kept their knowledge in the greatest secrecy
for a long period and would not deign to impart it to any one. At length, subdued by time
and humble entreaty, they revealed some few things, but the greater part they concealed
entirely from the vulgar.
In the opinion of E. Zeller 2 there is little trustworthy evidence to support the
assumption that the philosophy of the Greeks was derived from Oriental or Egyptian
influences, although it is highly probable that it received some impulses from the
East ; but whatever the Greeks borrowed from foreign sources was clarified and
refined by the fire of their own genius. For example, it has been said that the
Phoenicians taught the Greeks the art of writing, but that it was the Greeks who
wrote.
The knowledge of the secret arts, and the prevailing opinions of the Egyptian
priests, as Herodotus (c. 440 B.C.) relates, must have been communicated in part to
many vagabond Greeks during their sojourn in Egypt from about 660 B.C. The
unrivalled Grecian artists surpassed their teachers in the beauty and elegance of
their aesthetic productions, and also in works dependent upon imagination and
30 INORGANIC AND THEORETICAL CHEMISTRY
fancy ; but artisans and craftsmen made much slower progress with the philoso-
phical Greeks than with the more practical Egyptians. An Alexandrian Society
is reported to have been formed among the Greeks in Alexandria about the third
century, but, so far as we can gather, the knowledge which they are supposed to
have acquired mainly from the Egyptians, was confused with metaphysical
fancies ; and its expression was obscured by ambiguous allegories and cabalistic
symbols — possibly aping the hieroglyphics of Hermes — so that their writings now
appear to us as if the authors tried to conceal their own ignorance in a cloud of
words and symbols. ,
The Greeks did not contribute much to the chemical arts, but they furnished
chemistry with a science of method applicable to all the sciences. The Egyptians
accumulated facts and invented useful arts ; the Greeks discovered the laws of
investigation, the principles of discovery, and the laws of thought. The most
important result of centuries of Grecian effort was consummated in the mighty
Organon of Aristotle (c. 320 B.C.). This organon of deductive and inductive method-
ology should have inaugurated the third of Comte's les trois etats, but it did not.
The facts had not been determined with sufficient accuracy. Isaac Newton could
not have discovered the gravitational law if accurate data had not been prepared
for him by J. Kepler and G. Galilei. Aristotle's organon came too early. In
any case, the method of investigation so gloriously established by Aristotle was
unproductive ; it was degraded, misunderstood, and perverted by his disciples,
who, instead of applying the great principles of the organon, worshipped their master's
opinions on a host of special subjects as if they were oracles divine. Thus, I. R.
Averroes, about the middle of the twelfth century, went so far as to say, " The
doctrine of Aristotle is the perfection of truth, for his understanding attained the
utmost limit of human ability."
The method of Aristotle was rediscovered and restated by Francis Bacon in his
Novum organum (London, 1620). The two methodologies are substantially the
same. To some. Bacon's organon appears to have inaugurated a kind of Lutherian
reformation in science ; rather did the Baconian organon grow tardily from seeds
planted by Aristotle and his predecessors in the unproductive soil of metaphysical
speculation. According to G. H. Lewes' Aristotle (London, 1864), the main
cause of the sterility of the method of Aristotle and Francis Bacon was their
failure to appreciate the need for unremitting verification, so well emphasized
by Roger Bacon (c. 1280) ^ — Bacon the First — in order to vindicate the
principles deduced from the available facts. The same idea was emphasized by
Albertus Magnus, about the same time as Roger Bacon :
A principle which does not agree with experimentali cognitione (experimental knowledge
acquired by the senses) is no principle, but rather the opposite.
Aristotle himself frequently emphasized the danger of relying on mere guesses as
if they were observed facts, but he so often departed from his own precepts that he
was frequently inveigled by the perils of his own speculations. The illustrious
Francis Bacon likewise completely failed in vigilance when he attempted to apply
his own method because he did not practise the very principles he had expounded
so well. Bacon the Second even went so far as to say that if his methods were
adopted, little would depend upon the acuteness of the intellect, for the varied
talents of all men would be reduced to one common level. He said :
Our method of discovering the sciences is one which leaves not much to the acumen and
strength of wit, but nearly levels all wits and intellects.
Although the principles of Francis Bacon's organon have been available for nearly
three hundred years, they have proved quite inadequate, and there are no signs
of this socialistic levelling, for the interval between mediocrity and talent is as great
as ever it was. Francis Bacon himself can scarcely be considered to have been a
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 31
scientific man, or even to have possessed the scientific instinct. Science may have
been about him but it certainly was not in him.
The Greeks seem to have generally emphasized subtilty in speculation and
debate rather than accuracy in observation and experiment. In theoretical work,
they had an overweening tendency to extreme abstraction, and they were careless
and credulous in observation ; otherwise expressed, they founded arguments too
confidentl} on unproved statements, and seem to have regarded logical consistency
of greater weight than accuracy in the statement of facts. This characteristic was
well summed up by the saying which Plato, in his Timceus, ascribes to the Egyptian
priest of Sais : "Ye Greeks will be always children ... for though wisdom falls
from your lips, your actions are weak and puerile ; " or as S. Brown expressed it :
" In the art of experiment, the Greek was as feeble as a child ; but in the sphere
of ideas and vast conceptions it is not a paradox to say that he was sometimes
stronger than a man."
The Ionian doctrine of one primal element.— The Theogony of Hesiod (c. 700 b.c.)
assumed that " the earth is the unmovable basis of the cosmos," but the poem is
rather a record of mythic cosmology, and anthropomorphism, characteristic of the
first of Comte's les trois etats, and it had no influence on the development of
philosophical opinions.'* Similar remarks apply to Pherecydes (c. 600 B.C.), who
followed Hesiod with an improved mythology. Pherecydes made a definite attempt
to distinguish between the material constituents of the universe — e.g. between the
earth and its atmosphere, and also between matter and force. He regarded force
as a mysterious power exerted by the god Zeus.
The Ionian philosophers — Thales, Anaximander, and Anaximenes — still further
substituted impersonal causes, acting uniformly and continuously, for personal
causes acting capriciously and arbitrarily. At this time, therefore, the Greeks
were in a transitional stage between the first and second of Comte's les trois etats.
Thus, the early philosophers of Greece soon recognized that a belief in superhuman
gods was not sufficient to explain the complex phenomena in the physical world.
They then promulgated hypothesis after hypothesis to explain how the universe
grew from some simple principle — earth, water, air, fire. The new explanations
proved just as unmanageable as those which regarded natural phenomena as the
work of supernatural agencies. These early speculations of the Greeks do certainly
testify to the vigour and activity of their questioning spirit,^ but their ardour and
confidence were untamed by labour or reverses. It required centuries of chasten-
ing discipline for man to learn that " he must acquire, slowly and patiently letter
by letter, the alphabet in which nature writes the answers to such inquiries."
The first of the seven wise men of Greece, Thales of Miletus (c. 600 B.C.)— a
contemporary of Solon — made one of the earliest protests against the personifica-
tion of nature by assuming natural phenomena to be produced by capricious designing
agencies— diabolic or divine. Three centuries later, Epicurus likewise protested
emphatically against referring natural phenomena to the deliberate interventions
of gods. Thales believed that natural phenomena are due to the operation of
invariable laws to be discovered by a proper appHcation of the human intellect.
According to Thales, all the various forms of matter are different manifestations of
one underlying essence or prima materia. The universe, to him, was made from
water, which he regarded as the primal element. The same idea occurs m many
of the sacred books of the Hindus— e.^r. The Institutes of Manu—&hout the nmth
century B.C. There is really nothing to show how Thales was led to make the
assumption. It has, however, been noted that it is typical of systematic thinkers
to reduce to one general proposition that characteristic which is possessed m common
by a number of simple facts ; and it is therefore hinted by Aristotle that 1 hales,
meditating on the constitution of the universe, saw that water or moisture is
omnipresent ; that Thales was impressed by the marvellous transformations ot
water in the form of rain and dew, snow and hail, river and sea ; and that the eartn
appeared to be floating in an ocean of water. All things also seemed to be nourished
32 INORGANIC AND THEORETICAL CHEMISTRY
by water, and he accordingly assumed that water is the sole primal element which
is convertible into all the manifold varieties of matter— mineral, vegetal, and
animal — -found in the world. W. Whewell, however, emphasizes his belief that the
opinions of the philosophers of this period are based on vague suggestions and
casual analogies, rather than on reasons which will bear examination. It is very
remarkable, said A. Comte (1839), that the most inaccessible problems, such as the
origin and cause of phenomena, should be the very ones which first occurred to
students of nature, while those which were within their reach were considered to
be unworthy of meditations serieuses.
Another Miletian, Anaximenes (c. 500 b.c), is considered to have been the pupil
of Anaximander, who, in turn, was the disciple of Thales. Anaximenes sought for
the first principle of things in the omnipresent yet invisible air, which he regarded
as the equivalent of life because all living beings were nourished by air. Air
embraces the whole world, said Anaximenes, and he regarded air as the one eternal
essence, more primitive than Thales' water. He called air to aTreipv — the infinite
— and considered it to be devoid of any material differentiation. Even as late as
the eighteenth century, some chemists accepted Anaximenes' air as the primordial
element. Thus S. Hales, in his Vegetable Staticks (London, 1727), supposed that
atmospheric air deprived of its elasticity entered in a solid form into the composition
of most substances, and that air is the universal bond in nature. G. E. Stahl also
wrote to the same effect in his Experimenta,ohservationes,animadversionesCCC nuynero
chymiccB et physicce (Berlin, 1731). We have no record how the lonians — Thales and
Anaximenes — accounted for the formation of the different forms of matter from
their primitive elements, since matter by itself can only be matter.
The Ephesian Heracleitus (c. 450 b.c.) expressed himself in such enigmatical
terms that he has been called the Obscure Philosopher.^ A few fragments of his
writings have survived. Heracleitus appears to have maintained that all ideas are
derived from sensations, and he was the author of the celebrated doctrine that all
things are perpetually in a state of motion or flux, and that there is no rest or quietude.
Strife between opposite tendencies is the parent of all things. All life is change,
and change is strife. The living and moving element in nature seemed to him to
be an setherial exhalation, or fire. All things in nature are formed of the principle
of fire, which, in turn, is composed of small indivisible parts, i/^ty/xara or atoms,
which are in perpetual motion. If all things are conceived to be in perpetual
motion or change, then all things are fire. Never-resting fire rules all. Every-
thing has arisen from fire by condensation or rarefaction, and all things resolve
themselves back into fire. This idea is but a modification of the water and air
elemental of the Ionian philosophers. Obviously, Heracleitus' elemental fire was not
ordinary fire ; he probably understood fire to mean that which by constant trans-
mutation causes all the varied changes seen in the universe, and which itself remains
unchangeable. This idea of a primum mobile comes as near to the modern doctrine
of energy as was possible with the facts then available.
The Grecian Hippocrates (c. 400 b.c.) was not exactly a follower of Heracleitus, although
there is a strong resemblance between the views of both. Hippocrates specialized in medicine,
and he has been called the oracle of medicine ; he expressly rejected the use of hypothetical
philosophy in medicine ; he did not altogether rely on empirical experience, but attempted
to formulate general rules and principles derived from experience and knowledge. From
the chemist's point of view, a small treatise, On airs, waters, and sites, is considered to be
the most interesting of the works attributed to him. As might have been anticipated it
contains many errors and inexactitudes.
The Ionian doctrine of one primitive element was abandoned by Anaxagoras
in a work On nature (c. 450 b.o.).'^ He assumed that every difference in the sensible
qualities of bodies is fundamental, original, and inalienable ; and that there are
so many elements as there are simple substances ; no means were known at that
time for breaking down the majority of substances, and they were accordingly
assumed to be simple or elemental. The number of elements was therefore supposed
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 33
to be very large, or infinite. By repeated subdivision, Anaxagoras argued that all
natural things could be resolved into ultimate particles which were later on termed
homoBomericB — ofjioios, like ; fxipos, a part. The homoeomerise were supposed to be
eternal, unchangeable, infinitely divisible, and capable of continuous extension.
Like homceomerise act on like, and so form matter. If the qualities of the homoeo-
merise are assumed to be developed only when the particles are in combination with
others, Anaxagoras' homoeomeriae are not very different from the atoms of Leucippus,
Democritus, and Lucretius. Anaxagoras' atoms are the same in kind as the
substance itself ; Leucippus' or Democritus' atoms are indivisible particles of one
kind of matter. Anaxagoras also introduced the idea of a motive principle, which
he called vov<s, as the cause of all changes. Democritus called this principle dvayicrj ;
Heracleitus, avaOvfiiacn^ or fire ; and Aristotle, at^^p, aether.
The four and five element theories. — The four and five element theories are
among the oldest attempts to classify the protean and multitudinous forms of
matter which make up the world. The five-element theory seems to have been
favoured by the Chinese and Hindu philosophers. The Greeks reduced the number
of elements to four. Diogenes Laertes (c. 250) tells us that the five-element theory
was first promulgated by the Pythagoreans, and that Empedocles (c. 500 B.C.) first
advocated the four-element theory as a consistent doctrine. Empedocles cited the
burning of wood in favour of his hypothesis. When wood burns, srrwke or air rises
upwards, and this is followed by flame or fire ; moisture or water is deposited upon
any cold surface in the vicinity ; and ash or earth remains behind. Empedocles'
simple statement seems to be the first record of a chemical analysis. Wood is
resolved into its supposed elementary constituents — fire, earth, water, air. True
enough, modern methods can probe much deeper, but Empedocles' analysis is excel-
lent for its time. The doctrine of the four elements thus appears as a methodical
deduction from facts observed during the analysis of wood, by burning it in air.
This analysis has been claimed as " the starting-point of chemistry in history."
Empedocles also formulated the germinating conception of chemical afl&mty,
for lie said that the cause of the various combinations and separations of these four
elements is love (<f>t\ia) and hate (vcikos), which are exerted as active forces pro-
ducing the union or decomposition of substances. The four elements of Empedocles
soon lost their material character, and grew into abstract principles. It was then
fancied that the whole world was compounded of four distinct principles or entities
—the earth typified all solids ; water, liquids ; air, the winds, clouds, and the breath ;
and lastly, fi/re, which was symbolized by the sun, and worshipped by many as
a god. Hence, in the writing of the alchemists of the Middle Ages, there is usually
a chapter devoted to this quartet — earth, water, air, and fire. In J. Lacinius'
alchemical treatise Pretiosa fnargarita novella de thesauro (Venice, 1546), fire is
symbolized by an angel, air by a bird, water by a dragon, and earth by a bull.
Aristotle added a fifth element, at%, aether, more divine than the others, and which
pervaded all things, and was in perpetual motion. Later, Aristotle's aether became
the quinta essentia—a. kind of primal matter, a divine subtle extract, the qumtessence
of the other four. The ancient Hindu philosophers also had a fifth element, which,
in their system, was wrongly supposed to be a medium for propagating sound, etc.,
and which, in consequence, had something in common with the modern concept
of an aether pervading all space. The Institutes ofManu regarded the subtle aetHeT
as being first created ; and from this, by transmutation, sprang air, which changed
into light or fire, and thence into water, and finally earth. f a o
Aristotle assumed that the one primitive quintessence of matter can act as a
vehicle or carrier for four primitive qualities : hot or cold, wet or dry. ii tnese
four qualities or elements are united with inert passive matter in pairs tne lour
primary forms of matter— air, earth, water, fire— are produced ; for i^fj^n^' J^^^
has hot and dxy ; t^a^er, cold and wet ; air, hot and wet ; ^nd eart/^ cold ana a^^^^
The different varieties of matter arise when different degrees of these ^ourelementa
qualities are impressed on matter. Aristotle denied that the four elements ot
34 INORGANIC AND THEORETICAL ^CHEMISTRY
Empedocles are really elements because they are mutually convertible one into
another. Empedocles' elements, however, may represent the four primary forms
of matter perceived by the senses, and into which the four qualities appear to be
resolvable :
For hot, cold, moist, and dry, four champions fierce.
Strive here for mast'ry, and to battle bring
Their embryon atoms. — J. Milton.
The alchemists of the Middle Ages supposed that the elements were formed
when the primal essence was clothed with three principles — tria 'prima — which they
called respectively salt, sulphur, and mercury. In the quaint words of Paracelsus :
Eisen, stahel, bley, smaragd, sappir, kieszling, nichts anders seind denn schwefel, salz,
und mercm*ius.
Salt represented the earth or the principle of fixity and solidity ; mercury represented
air and water, or the principle of liquidity and gaseity ; and sulphur typified fire or
the principle of combustion. Thus, said Paracelsus, " whatever fumes and evaporates
in the fire is mercury ; whatever flames and is burnt is sulphur ; and all ash is salt.'''
Albertus Magnus typified the three principles by arsenic, sulphur, and mercury, for
he supposed the metals were compounded of these elements.
The three principles of the alchemists were not substances or corpora, but rather
principia or qualities ; they were representative types of qualities or classes.
Sometimes the tria prima were confused with the four elements of the Greeks, and
it is difficult to understand clearly what was gained by the invention of the three
principles. The mystic alchemists went even further and imagined that all material
things were composed of a trinity : "A body and a soul held together by a spirit
which is the cause and the law." They believed the soul of matter to be the trans-
forming principle which they tried to extract in a pure form, and which they
expected would enable them to transform the baser forms of matter into the purer
forms, of which gold was the best type.
The four-element theory was demolished when water, air, and the earths were
decomposed into still simpler bodies ; and when fire was shown to be a manifestation
of energy. The term " element " was obviously not intended to be used in the same
sense as it is to-day. The four and five elements of the ancients were not con-
sidered to have aH independent natural existence, but to be derivatives of one
another ; the earlier notion of an element rather referred to the genesis of matter
than to its ultimate analysis, for the distinction between simple and compound
substances does not seem to have entered their minds. Whatever the idea involved
in the three, four, or five element theories, it was believed by many different races
in different parts of the globe ; it has pervaded the philosophy of all thinking races ;
it has been sung by the poets of every land ; and it has had a longer life than any
succeeding philosophy. The theory was living three centuries ago ; it is now dead.
The Greek philosophers. — Three gigantic spirits have dominated Grecian thought
— Pythagoras, Plato, and Aristotle. Each one in his turn exerted a profound
influence on his contemporaries, and on subsequent thinkers. Thomas Carlyle has
well said that all history revolves around certain famous personages. The records
of Pythagoras (c. 500 b.c.) and of his school are overgrown with myths and fictions ;
and, as with the records of other influential men of old, the older the records, the
greater the tendency to associate miraculous and extraordinary events with the
men's lives.^ The Pythagoreans formed primarily a moral, religious, and political
association, although the sect early gave a definite trend to philosophical thought.
The scholars are now mainly dependent upon more or less untrustworthy reports
for their knowledge of the physical tenets of the Pythagoreans. It is generally
agreed, however, that the Pythagoreans believed that number is the essence of all
things. It is difficult to gather what was meant by this high-sounding phrase, for
number appears to be merely a relation, or the expression of certain facts. One
section of the Pythagoreans — e.g. lamblichus — held number to be the substantial
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 35
element of corporeal things ; otherwise expressed, Uke Plato's ideas, numbers are the
eternal archetypes of thmgs ; another section— e.^f. Hippasus— held that all things are
formed, not out of number, but after the pattern of numbers— otherwise expressed
number is the pattern or model from which things are copied, meaning that all things
bear the same fixed relation that a series of whole numbers bears to unity ; or again
as expressed by Philo, or whoever wrote the Booh of Wisdom, " God ordained all
things in measure, number, and weight." According to E. Zeller's account of the
Pythagoreans, the idea must have arisen as man dimly realized the definite and
mathematical order in natural phenomena.
From the more or less legendary accounts of the Pythagoreans, ft appears that
they reduced all things ultimately to one incorporeal monad, and assumed that all
things are compounded of monads with dissimilar and opposite natures, the uniting
bond being harmony. Later writers — e.g. Ecphantus — appear to be in error when
they state that Pythagoras' monads were corporeal. The Pythagoreans attached
special importance to the number 4, the quarternion, which was said to be the source
and root of eternal nature ; and the later Pythagoreans — e.g. Philolaus — were fond
of arranging things in series of four. Philolaus considered that the elementary
nature of bodies depended upon their form, and it was assumed that the smallest
constituent parts of the earth had the form of a cube ; air, an octahedron ; fire, a
tetrahedron ; water, an icosahedron ; and the fifth dodecahedral element represented
the universe, and embraced all the others. The diagrams. Fig. 5, explain the idea.
The historical evidence has not enabled the scholars to decide whether Empodocles
adopted four from Pythagoras' five elements, or whether Pythagoras added a fifth
Tetrahedron Octahedron Icosahedron Cube
(Fire) (Air) (Water) (Earth)
Fig. 5. — ^Primitive Particles of Pythagoras' Elements
Dodecahedron
(Universe)
element to Empedocles' four. It is thought that the Pythagoreans probably
derived the five-element theory from the Hindus. According to Max Miiller, the
coincidences between the teachings of Pythagoras and Hindu learning are so
numerous as to make it highly probable that Pythagoras obtained his leading
tenets by contact with the Indians in Persia.
The celebrated Athenian pupil of Socrates, Plato, expounded his views on
natural phenomena in his Tiynoeus (c. 360 B.C.). He assumed that all things
and all phenomena are transitory and unreal, but the abstract idea of them is
alone eternal and real. Hence, the aim of philosophy is to discover the ideas or
abiding principles of which the phenomena of the material world are but the
image. I. Kant (1790) described Plato's hypothesis by the celebrated metaphor :
Just as a flying dove, feeling the resistance of the air, might wrongly
suppose it would be able to fly faster in airless space, so did Plato, feeling
the limits which the sensuous world opposed to his understanding, assume
that by abandoning the sensuous world, he would be more successful in the void
space of pure intellect. Plato asserted as an a priori truth that the principle of
matter was infinite, eternal, and deprived of all qualities ; that matter is converted
into bodies by being impressed with some occult moving power ; and that matter
may possess particular qualities— hotness, dryness, coldness, and wetness. He
considered that there are four elements— air, water, fire, earth— and assunied that
these elements can never be destroyed. The elements can be divided into infinitely
smaU particles incapable of further subdivision ; the ultimate particles of the
elements have definite forms analogous to those suggested by Philolaus, Fig. 5.
36 INORGANIC AND THEORETICAL CHEMISTRY
The differences between the various kinds of the same elements are due to differences
in the bounding planes of the constituent particles. Fire, air, and water can be
transformed into one another by the coalescence of the primitive particles into forms
peculiar to these substances. Earth cannot be converted into any of the other
three elements because its cubical particles have no mathematical relation with the
forms of the other three. ^
The influence of Plato's pupil, Aristotle, on the world of thought has been
rivalled only by the founders of the great religions. Aristotle lived between 384
and 322 b.c. Excluding his Organon, to which reference has already been made,
Aristotle's most interesting contribution to natural science is entitled Meteorology,
and it deals with astronomical, chemical, and geological subjects ; his views on the
constitution of matter are expounded in his Generation and Corruption. There is
a work on Physics containing unfruitful disquisitions on abstract space, motion,
infinity, etc., and also a kind of sequence to this work entitled The Heavens.
Aristotle also wrote some biological works. There has been some discussion as to
whether a work on Mechanical Problems attributed to him is really the one to which
he sometimes refers. lo
Greek was not a familiar language to the philosophers of the Middle Ages, and
Aristotle's writings in the original Greek do not appear to have been known in
Western Europe prior to the thirteenth century. Aristotle's works were translated
into Syriac, thence into Arabic, and carried to Spain by the Moors. About the
fourteenth century Latin translations, made direct from the Greek manuscripts, were
read in Europe, and soon got a remarkable hold on European thought. In a general
way, it has been said that although Aristotle professed to rely on experience and
induction as the sources of true knowledge, he often went astray ; his treatment
of natural philosophy displays a capable mind, hampered by unsuspected super-
stitious prejudices, wrestling with problems beyond its strength. Aristotle rejected
Plato's idea-hypothesis and Pythagoras' number-hypothesis. He supposed matter
to be capable of infinite division, and he objected to Democritus' idea of atoms,
although he admitted that matter may be made up of particles which are actually
though not potentially indivisible. Aristotle did not agree with Pythagoras'
and Plato's hypotheses that the elemental monads have definite geometrical forms.
He said that the attempt to bestow an intrinsic figure on the elements — Fig. 5 — is
absurd, an element cannot have one. Elementary matter must be formless and
amorphous, ready to take on any form according to circumstances, but itself possess-
ing no particular form.
A famous disciple of Aristotle, Theophrastus (c. 372-287 B.C.) of Lesbos, suc-
ceeded his master at the Lyceum.ii Theophrastus wrote two works on botany
which were standard even throughout the Middle Ages ; a history of physics, a
work on natural science, and several other fragments — some writings ascribed to
him are no doubt spurious. Theophrastus followed the philosophy of Aristotle
rather closely in his Treatise on Fire (c. 315 B.C.) — Trepl nvpos. Theophrastus removed
fire from the list of elements ; and he recognized that air plays an important part
in the maintenance of a flame, and in the development of plants. In a fragment
On Odours, he adds that the odour of a substance is due to its volatility. The more
important parts of Theophrastus' writings, from the chemists' point of view, dealt
with minerals — irepl KiOoiv. Here he mentions coal, cinnabar, orpiment and
sandarach (arsenic sulphides) for the first time, and he also describes the prepara-
tion of white lead, red lead, verdigris, colcothar, chrysocolla, etc.
A number of writings and fragments of Archimedes of Syracuse (287-212 B.C.)
has been preserved. 12 They deal with some mechanical and hydrostatical problems.
The discovery of the celebrated principle of Archimedes — if a solid be weighed in
air, and then immersed in water, the apparent loss of weight is equal to the weight
of a volume of water equal to the volume of the solid — is described in M. P. Vitru-
vius' De architectura, published near the beginning of the Christian era. Al-Khazini's
account in the twelfth century lacks the piquancy and interest of that of Vitruvius.
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 37
It has been remarked that although Archimedes had fairly entered upon the right
path of his department of experimental science, no further advance was made for
nearly two thousand years, when Galilei and Stevinus took up the work. The
celebrated Hero flourished about 117 b.c. In his work on Pneumatics, he described
the principal physical properties of air known to the ancients, and indicated some
ingenious mechanical contrivances operated by means of rarefied or compressed air ;
ho also wrote a treatise on the properties of reflected light ; and two treatises on
the mechanical powers.^^
The main contributions of the Greeks to chemical science are the prima Tnateria
hypothesis ; the four-element theory ; the atomic theory ; the idea of the trans-
mutation of matter from one form to another by some agent or principle ; and more or
less vague notions of an active principle causing combination or change. There is
also Aristotle's work on methodology. i* The Greeks were not generally guided by
observation and experiment either in founding or in verifying their hypotheses.
Consequently, their great conceptions were wondrous feats of the imagination ; but
Lord Macaulay would have none of it, for, in his essay on Lord Bacon (London, 1837),
he claimed that the Greeks aimed at the stars, and through no want of skill the shot
was thrown away. The arrow was indeed followed by a track of dazing radiance,
but it struck nothing. Their philosophy began in words and ended in words.
Rome. — The Romans acquired some knowledge of the chemical arts after they
had conquered the Egyptians and the Greeks. The Romans had no philosophy of
their own, but they borrowed ideas and learned lessons at the feet of conquered
Greece.15 War was the strength of the Romans, and they favoured the arts and
crafts which made good soldiers. The works of art which the Romans acquired as
loot from conquered nations attracted much attention, and stimulated some of their
artisans to imitate these productions. The Romans, however, displayed but little
inventive genius, and it is probable that what successes they obtained were largely
due to the work of imported craftsmen. The early Romans developed a code of
civil law which has been a pattern for succeeding nations. The doctrine of the
supremacy of law inculcated by the Romans probably exerted some influence on
man's subsequent attitude towards external nature, and some confusion has resulted
from the assumption that a law of nature represents an obligation on the part of
natural phenomena analogous to the obligations of a people to its civil law. The
modern view of a law of nature is very difierent from this.
The poem of Lucretius (95-52 b.c), De rerum natura, is much admired, and is
intimately associated with the history of the atomic theory. It has been considered
curious that, with the exception of a few fragments and letters, the works of the
three founders of the Grecian atomic theory — Leucippus, Democritus, and Epicurus
— should have been lost, and that we have to rely upon the Roman's eloquent poem
for a clear and concise account of Epicurus' doctrine. There is nothing to show if
Lucretius added anything new to what he found in Epicurus' two works — Concern-
ing nature, and On atoms and voids — which have been lost.
The principal writings dealing with the physical arts and crafts of the Romans
are the works of Vitruvius, Galen, Dioscorides, Varro, Seneca, and Pliny. M. P.
Vitruvius was an engineer and architect in the service of the Roman state at the
time of Augustus— near the beginning of the Christian era. He wrote the cultured
work, De architectural^ in which he gives many indications of the learning of his time,
viewed more particularly from the point of view of the practical application of
theoretical knowledge. Another celebrity— Dioscorides— was born in Asia Minor
and flourished in Rome about 75 a.d.i^ contemporaneously with Pliny. His Z)c
tnateria inedica was a standard work for many years. In this book, Dioscorides
describes the art of distillation for the first time, although Aristotle (c. 320 B.C.)
seems to have had this operation in his mind when he wrote in his Meteorology (2. -) :
Sea water is rendered potable by evaporation ; wine and other liquids can be submitted
to the same process, for, after having been converted into vapours, they can be condensea
back into liquids.
38 INORGANIC AND THEORETICAL CHEMISTRY
Dioscorides also describes other chemical operations — e.g., the extraction of mercury
from cinnabar, by heating a mixture of the latter with carbon. He also mentions
lime-water, zinc oxide, blue vitriol, white lead, etc.
Another Greek, C. Galen (131-201 a.d.), regarded himself as a disciple of Hippo-
crates. He practised medicine in Rome about 160 a.d. He was an experimental
physiologist, and wrote on human anatomy, physiology, and botany. M. T. Varro
(116 B.C.-28 A.D.) wrote on agriculture, law, mensuration, etc. A. Seneca wrote
a work, Qucestiones naturales (c. 63 a.d.), which appears to have been largely
drawn from Aristotle's Meteorology.'^^ It deals mainly with astronomy, meteorology,
and physical geography ; and it was the authority on science in the Middle Ages
up to the fourteenth century, when it was largely supplanted by Aristotle's works,
which then became accessible in Europe through the Latin translations of the Greek
texts. Caius Plinius Secundus, or Pliny the Elder,2o wrote the Historia nuturalis
about 77 A.D. It deals with an enormous variety of subjects and is a congested and
uncritical compilation from credible and incredible authorities and popular beliefs.
E. Gibbon, in his Decline and Fall of the Roman Empire (London, 1789), called it
" an immense register of the discoveries, arts, and errors of mankind." The works
of but a few of the authorities quoted by Pliny are known.
The prosperity of the Roman Empire — which included England, France, Spain,
and all the countries about the littoral of the Mediterranean Sea — was on the wane
about 180 A.D. The Romans, satiated with conquest, became indolent and corrupt,
and their intellectual activity slackened. Their empire was invaded by the un-
civilized northern races — Goths, Vandals, and Huns. The destructive impulses of
the invaders led to the complete disintegration of the empire ; and about the fifth
century, culture and civilization in Rome were crushed in a few dark years. Many
of the records of science, literature, and art were deposited in monasteries, where
they were preserved as sacred trusts until civilization again revived in Western
Europe. Fortunately, however, Constantine transferred the Roman capital to
Byzantium (Constantinople) in the fifth century, and the New Rome maintained a
continuity of government and of civilization until the raid of the fourth crusaders
in 1204. It is considered that more destruction and damage to ancient records
were wrought in the sack of Constantinople by these Crusaders than by the Mahom-
edan conquest in 1453.21
References.
1 G. Grote, A History of Greece, London, 1. 355, 1869.
2 E. Zeller, A History of Greek Philosophy, London, 1. 26, 1881 ; S. H. Butcher, Some Aspects
of the Greek Genius, London, 1891 ; L. von Schroder, Pythagoras und die Inder, Leipzig, 1884 ;
M. B. St. Hil&ire, Premier memoire sur le Sankhya, Paris, 1852; R. Gar be, Monist, 4. 176, 1894 ;
W. Jones, Works, London, 3. 236, 1799 ; H. T. Colebrooke, Miscellaneoiis Essays, London, 1837.
' Roger Bacon, Opera inedita, London, 1860; Opiis majus, London, 1773; G. H. Lewes,
The History of Philosophy, London, 2. 77, 1871 ; B. R. Rowbottom, Journ. Alchem. Soc, 2. 75,
1914 ; S. Brown, Essays, Edinburgh, 1858; J. E. Sandys, Boger Bacon, London, 1914.
* E. Zeller, A History of Greek Philosophy from the Earliest Period to the Time of Socrates,
London, 1. 211, 266, 1881 ; F. Bacon, De principiis atque originibus, London, 1612 ; J. Burnet,
Early Greek Philosophy, London, 1908.
* W. Whewell, History of the Inductive Science, London, 1. 20, 1857 ; J. H. Bridges, Essays
and Addresses, London, 143, 1907.
* F. Lassalle, Die Philosophic Herakleitos' des Dunkeln, Berlin, 1858; T. Gomperz, Sitzber.
Akad. Wien, 997, 1886 ; Greek Thinkers, London, 1. 69, 1901 ; J. Burnet, Early Greek Philosophy,
London, 143, 1908 ; G. Gladisch, Herakleitos und Zoroaster, Leipzig, 1859.
' T. Gomperz, Greek Thinkers, London, 1. 223, 1901 ; J. Burnet, Early Greek Philosophy,
London, 290, 1918.
® E. Zeller, A History of Greek Philosophy from the Earliest Period to the Time of Socrates,
London, 1. 306, 1881 ; J. Miiller, Naturwiss. Ver. Innsbruck, 23. 1897 ; T. Gomperz, Greek Thinkers,
London, 1. 99, 1901 ; J. Burnet, Early Greek Philosophy, London, 319, 1908.
' E. Zeller, Plato and the Older Academy, London, 1876 ; T. H. Martin, fjtudes sur le Timee de
Platon, Paris, 1841 ; J. S. Konitzer, Ueber Verhdltniss Form und Wesen der Elementarkorper n/ich
Plato's TimcBus, Neu Ruppin, 1846 ; E. O. von Lippmann, Journ. prakt. Chem., (2), 76. 513, 1907 ;
AbJiandlungen und Vortrdge zur Geschichte der Naturwissen-scJuiften, Leipzig, 2. 28, 1913; F. W. Bain,
On the Realization of the Possible and the Spirit of Aristotle, London, 1899. -
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 39
1" E. T. Poselger, Aristotle's Mechanische PrdbUme {Quoestiones mechanica), Hanover, 1881.
^^ E. Zeller, Aristotle and the Earlier Peripatetics, London, 2. 348, 1897.
12 T. L. Heath, Archimedes, Cambridge, 1897; Archimedes, Opera, Basil, 1544; (Euvres
Paris, 1807.
13 W. Schmidt, Hero's Werke, Leipzig, 1899 ; B. Woodcroft, The Pneumatics of Hero of
Alexandria, London, 1851 ; T. H. Martin, Hero, Paris, 1854.
1* E. Zeller, Aristotle and the Earlier Peripatetics, London, 1897 ; G. H. Lewes, Aristotle, a
Chapter from the History of Science, London, 1864; C. Daubeny, An Introduction to the, Atomic
Theory, Oxford, 1850 ; M. B. St. Hilaire, La physique d'Aristote et la science contemporaine,
Paris, 1863 ; T. E. Jones, Aristotle's Researches in Natural Science, London, 1912 ; J. Lorscheid,
Aristotles' Einfluss auf die Entwicklung der Chemie, Miinster, 1872 ; E. O. von Lippmann, Arch.
Geschichte Naturwiss. Technik., 233, 1910 ; Abhandlungen und Vortrdge zur Geschichte der Natur-
uissenschaften, Leipzig, 2. 64, 1913.
15 A. Terquem, La science romaine a Vepoque d'Aiiguste, Paris, 1885.
1^ J. Gwilt, Vitruvius, London, 1826 ; M, H. Morgan, Vitruviu^, London, 1914 ; A. J. Brock,
Galen, 1916.
1' E. 0. von Lippmann, Zeit. angew. Chem., 18. 1209 ; 1905 ; Abhandlungen und Vortrdge zur
Geschichte der Naturwissenschaften, Leipzig, 1. 47, 1906.
1^ H. Kopp, Beitrdge zur Geschichte der Chemie, Braunschweig, 217, 1869 ; E. 0. von Lippmann,
Chem. Ztg., 26. 629, 1189, 1911 ; Abhandlungen und Vortrdge zur Geschichte der Naturwissen-
schaften, Leipzig, 2. 157, 162, 1913.
1* J. Clarke, Physical Science in the time of Nero, being a translation of the Quoestiones naturales
of Seneca, London, 1910 ; A. Terquem, La science romaine a Vepoque d^Auguste, Paris, 1885.
^^ E. 0. von Lippmann, Zeit. an^ew. Chem., 6. 383, 1893 ; Abhandlungen und Vortrdge zur
Geschichte der Naturwissenschaften, Leipzig, 1. 1, 1906.
21 E. Pears, The Fall of Constantinople in the Fourth Crusade, London, 1885 ; J. B. Bury, The
Roman Empire, London, 1910 ; H. Gelzer, Byzantin Kulturgeschichte, Tiibingen, 1909.
§ 11. The History of Chemistry in Syria, Persia, and Arabia
Very little advance in culture could be made even by the greatest man of genius if he
were dependent for what knowledge he might acquire merely on his own personal observa-
tions. Indeed it might be said that exceptional mental ability involves a power to absorb
the ideas of others, and even that the most original people are those who are able to borrow
the most freely. — W. Libby (1917).
About the third, fourth, and fifth centuries, the Neo-platonic schools at Alexan-
dria and at Athens included Ammonius Saccas, Plotinus, Porphyry, lamblichus,
Proclus, etc. These schools cultivated mysticism and magic. As with the Pytha-
goreans, they taught that the air is full of spirits and demons which control health
and disease, and natural phenomena in general. It was said :
God rules the world. He has demons imder his control, some of which govern animals,
some vegetables, and others minerals. . . . One demon governs the liver and another the
heart.
When animals or vegetables were destroyed by fire, the gases which escaped were
supposed to be subtle spirits returning to the air. With beliefs like these, natural
phenomena could be investigated only by contact with the supreme divinity, and
this could be attained only by certain mysterious ceremonies involving the use of
secret symbols, incantations, and prayers. A knowledge of these ceremonies was
regarded as a divine gift particularly reserved for the priests and the mitiat^d.
Somewhat similar ideas were later incorporated in the mystical forms of alchemy
of the Middle Ages.
In the period between the first and fifth centuries, alchemy attracted the
attention of many learned men, and authentic writings on alchemy began to appear.
The first, Zosimos of Panopolia, lived in the third century, and most of his writings
seem to have been lost. Some fragments attributed to him have been collected
from Greek papyri, and he is often quoted by later alchemists. Zosimos described
various forms of apparatus and furnaces, minerals, and alloys and he frequently
refers in more or less obscure language to the transmutation of the metals J^rag-
ments of the writings of Zosimos, Africanus, Synesius, Olympiodorus, (pseudo)
Theophrastus, (pseudo) Democritus, and several other Greek alchemists i-a bout
40 INORGANIC AND THEORETICAL CHEMISTRY
150 in all — were preserved in European museums — Venice, Rome, Paris, Munich,
etc. — whither they drifted after the conquest of the Turks in 1453. The essays
reproduced in M. Berthelot's Collection des alchimistes grecs are all composed in an
enigmatical style with obscure chemical terms used in many different ways ; they
discuss magical and astrological formulae ; and give citations from mythical authors.
The writers were acquainted with many ores, minerals, earths, salts, and animal
and vegetable substances ; there is no evidence of a scientific classification ; and the
writers were in ignorance of the mineral acids and their important derivatives.
They were chiefly concerned with the operations of solution, distillation, and heating.
The conquests of Rome brought the Orient and the Occident, the East and the
West, into close communication. At the beginning of the Christian era, Alexandria
was the asylum of Eastern traditions, the centre of medical, alchemical, and philoso-
phical culture ; and the sanctuary of the world's learning. The Roman depreda-
tions in the fourth century led to a rapid decline ; and as a result of the Mahomedan
conquest of Egypt in the sixth and the seventh centuries, the Alexandrian philoso-
phers and teachers were scattered, and some refugee Byzantine alchemists travelled
to Constantinople ; others settled in Persia and Syria, where they introduced the
Greek and Egyptian philosophies. Some of the writings of the Greek philosophers
were translated into Syrian.
In the seventh century, the Arabians overran Syria and Persia ; and the Syrian
schools languished and died. The Arabians then began to cultivate those very
arts which they had done so much to destroy. Syrian scholars were employed by
the rulers for positions demanding wisdom, knowledge, and judgment. Copies of
two Syrian manuscripts are preserved in the British Museum ; one is translated in
M. Berthelot's La chimie au tnoyen age (Paris, 1893). It contains various technical
recipes, discussions on magic and mystic doctrines, the elixir of life, the adulteration
of gold, and descriptions of some chemical apparatus. An Academy was founded
at Bagdad about 800 a.d., and the Arabians began to collect and translate books
from various countries — East and West. The works of Aristotle were translated
from the Greek into Syrian, and re-translated from Syrian into Arabian. Conse-
quently the alchemists of Arabia derived their ideas and knowledge from those of
Syria ; the Syrians in turn were largely dependent on the Greek works of the
pseudo-Democritus, Zosimos the Panopolite, the pseudo-Cleopatra, and others who
flourished at Byzantium. Historians generally consider that the Greek writers of
this period, in turn, derived their ideas from the Egyptians. ^ In any case, the
Arabians, like the Greek writers of the Alexandrian school, imparted mysticism into
their versions of Hellenistic philosophy, so that there was a partial reversion to
the first of Comte's three states. Instead of regarding natural phenomena as the
workings of natural law, they were inclined to consider them to be subject to the
capricious wills of superior intelligences, and creatures of an imagined demonology.
As a result, physical scdence reverted to magic, astronomy to astrology, and philo-
sophy to theosophy. The alchemical operations were described in mystic language.
Hence too arose the philosopher's stone, the elixir of life, etc. The Arabians had
a bias in favour of medicine and pharmacy rather than metallurgy, and they appear
to have interpreted the alchemical writings from the Egyptians, in terms of medicine
and pharmacy — a bias possibly derived from the Hindus. Consequently, the
philosopher's stone of the Alexandrian school became the Arabian elixir of life.
The reputation of one Geber, an Arabian writer of the eighth century, loomed
mightily in the alchemical world about the later half of the Middle Ages. He is
credited with having been the first to give chemical knowledge a systematic form by
publishing the first extant system of chemistry. It is very true that the ideas
expressed in these writings are the earliest to stand in historical continuity with those
of the present day. This fact has invested the writings of Geber with a special
interest, and this int-erest is only quickened by a knowledge of their contents and
style. The fragmentary information which is available respecting Geber is most
disappointing ; there is no agreement among the^historians concerning his birth-place,
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 41
his parents, his social or political relations, his rank, the events of his life or
his death.3 There are, however, quite a number of Latin treatises alleged to be
translations of Arabic texts of Geber's writings. For example, up to recent years,
the Summa perfectionis magisterii was credited to Geber, and it was said to have
been the first work exclusively devoted to chemistry. The book is an attempt to
summarize what was then known or believed with respect to chemical operations
and processes. It is, however, disfigured by unintelligible matter which has wrongly
led some to the idea that the term " gibberish " for unintelligible words, is a tribute
to Geber's style of writing.
The Latin Geber was acquainted with alum, copperas, saltpetre, sal ammoniac,
aqua fortis, oil of vitriol, aqua regia, etc. ; he described the action of mercury on
gold, and of sulphur on red-hot iron ; and he supposed that there are three elementals
— mercury, sulphur, and arsenic. The metals, said the Latin Geber, are compound
bodies which are extracted from their earthy ores when the latter are mixed with
carbonaceous materials and heated in a furnace in the absence of air. It seemed
to him as if the calx got something from the furnace and so became a metal. Ac-
cording to the Latin Geber, the metals are compounds of the same substances —
mercury and sulphur — united in different proportions. Geber also accepted as
dogmas of his faith, the transmutation of the metals, and the influence of the
planets on the metals — although he said :
It is as impossible to transform the metals into one another as it is to turn a bull into a
she-goat ; for it has taken nature thousands of years to make the metals, and we cannot
hope to effect the transformation when we rarely live a hundred years.
Many grave doubts have arisen as to the genuineness of the Latin writings which
have been attributed to the eighth-century Geber. M. Berthelot has compared
the texts of the Latin works, and translated the known Arabic texts preserved in
the Museums at Paris and Ley den. He has also compared these works with those
of contemporary writers. The style and standards of the Latin and Arabian works
are altogether different ; and, as a result, Berthelot concludes that the Latin works
attributed to Geber were the composition of one or more writers about the thirteenth
century, who forged the name of the Arabian Geber to crown the book with veneration
and respect. The Latin version of Geber is not to be regarded as a translation
from Arabic texts. The Latin versions, on which Geber's reputation rests, are some-
times called the thirteenth-century works of the pseudo-Geber to distinguish them
from some Arabic texts which were probably the work of an unimportant Geber, or
of some writer between the eighth and eleventh centuries. The Arabian Djaber
(Geber) is reputed to have been the pupil of a Khaled ben Yezid ibn Moaouia, the
first Mahomedan writer on alchemy. M. Berthelot's translation of the works of
the Arabian Geber show that Geber use(f the hydrostatic balance ; attempted to
classify minerals ; discussed the changes in volume which occur when substances
are heated ; and stated that he had seen many persons ignorantly attempting to
manufacture gold and silver by wrong methods, and added : "I perceived these
workers were divided into two categories, the dupers and the duped. I had pity
for both of them." m j •
A debate among the Arabians as to the possibility of alchemy is described m
the writings of the Arabian E. S. Avicenna (980-1037). The doctrine was defended
by A. M. Rhases (840-940), or Rhazes,whose writings are often quoted by meditevai
alchemists. The Arabic physician Avicenna wrote on chemistp^ and medicine, and
he also wrote commentaries on the works of Aristotle. Judging from the reports
of his Porta elementorum and his Dictiones, his philosophical ideas closely followed
those of Aristotle ; his medical work. The Canon, was mainly a compilation ol
Hippocrates and Galen ; and his general knowledge was but little in advance ot
the Greeks. Notwithstanding this, Avicenna's medical works were long revered
as a code of science ; but they sank into almost complete obhvion about the end
of the seventeenth centurj^ Similar remarks apply to the commentaries ot
42 INOKGANIC AND THEORETICAL CHEMISTRY
I. R. Averroes (1126-1198) upon the works of Aristotle ; in fact, it was mainly
through the commentaries of Averroes that Aristotle's scientific work became
known in Europe in the Middle Ages.
There is a very important treatise, The hook of the balance of wisdom, written
in the twelfth century by the Arabian optician and physiologist Al-Khazini, or
AlhazanA It contains a memoir on the use of the balance for the determination
of specific gravities, and is supposed to have been based upon a work by Abu-r-Raihan
written about 1000. Al-Khazini said :
The water-weight of a body visibly changes according to the difference between the waters
of different regions in respect to variety and density, together with incidental difference due
to variety of seasons and uses. ... In winter one must operate with tepid, not very cold,
water on account of the inspissation and opposition to gravity of the latter, in consequence
of which, the water- weight of the body conies out less than it is found to be in summer. . . .
The temperature was apparently estimated by the distance a kind of hydrometer
sank in water. The specific-gravity bottle was described, and an improvement on
the floating hydrometer of Pappus (c. 400 B.C.) indicated. Gravitation seems to
have been regarded as a force directed to the centre of the earth, and which diminished
proportionally with the distance ; it remained for Newton to show that it diminishes
as the square of the distance. Both Abu-r-Raihan and Al-Khazini compiled tables
of the specific gravities of various solids and liquids with which they were acquainted ;
and the numbers agree closely with those adopted to-day.
During the period of the intellectual darkness which prevailed in Europe after
the decline and fall of the Roman Empire, the torch of learning was borne by the
Arabians, but there is little to show that the Arabian alchemists — Avicenna, Anven-
zoar, Averroes, etc. — who flourished between the eleventh and thirteenth centuries
— did much to extend the chemical knowledge which they derived mainly from their
contact with the Egyptians, Greeks, and Hindus. The Arabians borrowed freely ;
but they showed little genius for independent thought. In his posthumous A
History of Chemistry (London, 1913), J. C. Brown sums up by saying : "far from
crediting the Arabians with being the originators and improvers of chemistry, as
stated by E. Gibbon (1789),^ much of their knowledge was not understood, and they
involved it in mystical confusion which hindered the progress of science for cen-
turies ; " and W. Whewell, by saying :
The Arabians cannot claim in science or philosophy, any really great names, they pro-
duced no men and no discoveries which have materially influenced the course and destinies
of human knowledge, they have tamely adopted the intellectual servitude of the nation
which they conquered by their armies ; they joined themselves at once to the string of
slaves who were dragging the car of Aristotle and Plotinus.
About the eighth century, the Arabians amalgamated with the European settlers
in Egypt, and under the name Moors, crossed into Spain, where they founded
Academies at Cordova and Granada. These Moorish universities flourished between
the eighth and eleventh centuries, and furnished the schools of Europe with many
learned teachers. The power of the Moors in Spain was destroyed with the conquest
of Granada by the Christians under Ferdinand and Isabella in 1492. The Arabian
centre of learning at Bagdad was captured in the eleventh century by the Turks, a
tribe which separated from the Mongols in the sixth century, and settled in Asia
Minor. The Turks gradually extended their power westwards, and formed the
Ottoman or Turkish Empire under the leadership of Othman (born 1258). The
Turks crossed into Europe in 1356, and about a century later, 1453, captured
Constantinople. The learned men congregated in that city then slowly drifted
westwards with their manuscripts and learning.
The authors of the earlier Arabian alchemical books were directly or indirectly
associated with the famous schools of Alexandria, the last resting-place of the secrets
of the Egyptian priests. There can be no doubt that the chemical arts were well
developed in old Egypt. The Egyptian origin of the term chemistry would harmonize
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 43
with the prefixing of the article al (the) to the word Khem (Egypt) when the Arabians
overran Egypt, and thus learned many of the secrets of the temple laboratories of
the Egyptian priests. No doubt also the contact of the Arabians with Persia made
them acquainted with some chemical knowledge derived by the Persians from
India. The Arabians also learned from the Grecian philosophers through the Syriac
translations. The learning derived by the Arabians from East and West was
probably distorted, modified, and adapted to suit their own particular dogmas,
and carried to Europe partly by the currents of returning crusaders, and partly
by the Moors via North Africa and Spain.
The origin of the term chemistry.— Chemistry had no special name prior to the
sixth century, before which it was variously known as the art of Hermes, the Hermetic
art, the Sacred art, the Occult art, or the Black art. Many have tried to trace the origin
of the name chemistry, and the quest has led etymologists to suggest several
different hypotheses ; accordingly, the student has the choice of a number of
plausible guesses at his disposal.^
(1) The various attempts which have been made to make the root a Greek word have
not been veiy successful. H. Barbarus, in his Compendium acientice naturalia (1547), and
A. Libavius, in his Alchymia (Francofurti, 1606), consider it possible that the term is derived
from x^l^os- — -a juice or menstruum- — 'in reference to the use of various solvents by the
early alchemists ; or from x^'w — to fuse or melt ; and J. A. Quercetanus, in his De pris-
corum medicina (c. 1600), uses the term halchymiam for a fused salt — iAs, salt ; €ind
Alexander the Aphrodisian (c. 200 a.d.) speaks of the use of x^'f^ opyava — a kind of crucible
for melting substances. While this derivation of the word was in fashion, alchemy was
spelt alchymy, and chemistry, chym,istry ; but this spelling was dropped when it wsus
recognized that the Greeks had neither the name chemia nor the science ; it was only
near the beginning of the Christian era that the new science began to attract attention in
Europe. The scholars tell us that the word alchemy does not occur in Greek writings
earlier than the third or fo\u*th century, when J. F. Matemus mentioned the acientia
alchemice in an astrological work entitled Mathesis, written about 337 a.d. He says, in
the jargon of astrology : "If man be bom in the house of Mercury, he will devote himself
to astronomy ; if in Venus, he will be fond of singing and pleastu-e : if in Mars, he will
apply himself to arms- ; if in Jupiter, he will follow religion and law ; if in Saturn, he will
devote himself to alchemical knowledge. ..." Zosimos of Panopolis (Egypt), a writer
possibly contemporaneous with, or possibly earlier than, Mattmus, refers to x'/M^a.
chemia — or xvi^^'ta, chemeia — as the art of making gold and silver. We are also" told that
the term was seldom or never used by subsequent writers before the ninth century, but
thereafter somewhat frequently.
(2) It has been argued that the word is derived from the Hebrew word Chatnan or
haman, meaning a mystery or secret, in which case, chemistry would mean the secret art ;
and Zosimos (c. 400) considers that chemistry shoxild be treeisured as a religious secret to
be known and jealously guarded by the priestcraft. S. Bochart (c. 1660) favours a de-
rivation with a similar connotation, for he refers the word to the Coptic kema or kemo,
obscure or hidden, or the Arabic chem/i, to hide. Hencje the old designation the occult
science, and the Arabic book of secrets called Kemi.
(3) It has been suggested by S. Bochart, in his Oeographios sacrce (Cadomi, 1646), that
the word may be derived from Noah's son Cham, whom he thinks was identical with
Zoroaster the 'founder of the Magi. According to Diodorus Siculus' Bibliotheca historica
(c. 30 B.C.), the word chemistry is derived from the name of an Egyptian king named
Chemnis or Chemhes ; and, according to H. Goring's De hermetica n/iedicina (Helmestadii,
1648), the god Chemnis was worshipped in the city of Thebes, which was famous for its metal
and colour industries.
(4) Plutarch, in his De Iside et Osiride (c. 100 a.d.) implies that the word comes- from
the Egyptian Kham or Khem (Psalms, 105. 27)— meaning black or dark— because the same
word was applied to the country of Egypt. The term thus refers to the art of the black
coimtry, or the Egyptian art. The trend of opinion seems to favour this suggestion.
References.
1 H. Kopp, Beitrdge zur Geschichte der Chemie, Braunschweig, 1. 9^' ^869; 2. 1, 1869 .
M. Berthelot and C. E. Ruelle, Collection des anciens alchimistes grecs. Pans, 1887-8.
2 H. W. Schaefer, Die Alchemic, Flensburg, 1887 ; M. Bemiam, SenUnttis sacro medvcw,
Hamburg, 1640 ; A. J. Pernety, Les fables igyptiennes et grecques detmleesjt reduites au mime
principe avec une explication des hieroglyphes et de la guerre de Troye, Pans, 1 /o8. xj^^*^,
« H. ^mgst^ll,Literaturgeschichte der Araher, Wien, 2. 185, 1851 ,;^ 3. 293, 1851 ; F. Hoefer,
Histoire de la chimie, Paris, 1. 308, 1842 ; H. Kopp, Geschichte der Chemie, Braunschweig, 1. 61,
44 INORGANIC AND THEORETICAL CHEMISTRY
1843 ; T. Thomson, The History of Chemistry, London, 1. 119, 1830 ; K. C. Schmeider, Geschichte
der Alchemie, Halle, 86, 1832 ; M. du Fresnoy, Histoire de la philosophic hermeiique^ Paris, 1. 29,
1842 ; 0. Sprengel, Histoire de la medecin, Paris, 2. 263, 1815 ; J. Ferguson, Laboratory, 1. 71,
1867 ; Bibliothcca Chemica, Glasgow, 1. 299, 1906; Geber, Works, Gedani, 1682.
* J. W. Draper, A History of the Intellectual Development of Europe, London, 2. 45, 1876 ;
H. C. Bolton, Am^r. Chemist, 6. 413, 1876 ; N. Khanikoflf, Joum. Amer. Oriental Soc., 1. 1859 ;
J. J. Clement-MuUet, Joum. asiatique, (5), 11. 379, 18.
5 W. Whewell, History of the Inductive Sciences, London, 1. 211, 1857 ; E. Gibbon, The History
of the Decline and Fall of the Roman Empire, London, 1789.
« G. Hoffman, Ladenburg's Handworterbuch der Chemie, Breslau, 2. 516, 1884 ; A. F. Pott,
Zeit. deut. morg. Ges., 30. 6, 1876; E. Wiedemann, ib., 32. 575, 1878; C. Schorlemmer, Chem.
News, 40. 309, 1879 ; R. A. Smith, ib., 42. 68, 244, 1880 ; E. O. von Lippmann, Chem. Ztg., 38.
685, 1914.
§ 12. The History of Chemistry during the Middle Ages. Alchemy and
Medico- or latro-chemistry
The applications of chemistry to various kinds of industries are all buried in the tombs
of many generations of artists who have left no other traces of their existence than a few
of their productions.- — P. Lacroix (1869).
The Middle Ages are sometimes taken to extend from about the seventh to the
seventeenth centuries. During the fourth century Western Europe was ravaged
by Teutonic barbarians— the Goths and the Vandals. The Koman Empire trans-
ferred its capital to Byzantium (Constantinople), on the banks of the Bosphorus,
where Greek metaphysics mingled with Oriental mysticism ; and intellectual
Europe there managed to exist until the Turkish conquest of Constantinople in
the fifteenth century. The traditions of the Greek philosophers were preserved
in the schools of Alexandria and Byzantium,^ and there was a succession of real
though feeble students of philosophy, physical and natural science, mathematics,
and medicine. Byzantium thus kept alive the thought and knowledge of the
ancient world during a period when Western Europe was submerged in turmoil
and strife.
During the fifth century, the Huns, under Attila, devastated the fairest provinces
in the West about the time the Anglo-Saxons were conquering England. Natural
science could make no progress under these turbulent conditions ; and ignorance
and superstition prevailed in the West. There was a gradual infiltration of ideas,
knowledge, and art from Byzantium, the Greco-Roman Empire, into Western
Europe between the fifth and the fifteenth centuries. The fall of Byzantium
(Constantinople) in 1453 led to the westward migration of the scholars of the
Eastern Empire. Europe also gained some hints of the chemical lore of the
Arabians from the returning crusaders ; and after the Moors had carried Arabian
literature into Western Europe vid Spain in the tenth century, some progress was
made. The works of the Grecian and Egyptian writers were not directly known in
the West until after the thirteenth century, although Latinized versions of Arabian
translations, preserved in the Mahomedan libraries in Spain, were available.
This gave rise to the erroneous impression that chemistry originated in Arabia.
Some Latin translations of the Arabic writings were collected and printed in
the seventeenth century — for instance, the Theatrum chemicurn (Argentorati, 1613-
22) and J. J. Manget's Bibliotheca chemica curiosa (Genevse, 1712). M. Berthelot
found in these works whole passages taken from the older Greek alchemists. The
meaning of the original writings seems to have been distorted and perverted during ^^
the many translations and re-translations ; as a result, the mediaeval chemists oi^H
alchemists started their work with mutilated and incoherent descriptions of the ^
technical and philosophical works of the Greeks and the Egyptians ; and the literary
productions of the alchemists of this period are characterized by much obscurity,
either in unconscious mimicry, because their mutilated models were similarly
tainted, or else to hide their real meaning from a hostile community, or from the
vulgar. It was said : "A profound secret should not be revealed in the vulgar
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 45
tongue, the true adept can sufficiently comprehend the mystical language, and it
would not be right that it should be understood by the people."
Historians tell us that the tardy growth of science in the early Middle Ages was
largely due to the constitution of society. The chief elements of feudal society were
the barons and the priests. The barons were perpetually at war, and the study of
natural science and philosophy was eminently distasteful to them. The priests were
often men of great learning, but they devoted their energies mainly to theology. They
possessed great power over society, for on them devolved both the spiritual and the
temporal teaching of thepeopJe. Except inrare cases, the priests did not devote special
attention to the physical and chemical sciences. Aristotle's works were considered
sufficient for all purposes, and speculations in reference to natural phenomena were
discountenanced, and in some cases forbidden. Ignorance appeared to be a sacred
duty. It was generally thought impious to attempt to draw aside the veil enshroud-
ing nature's mysteries, and man shrank from all inquiry into the perplexed ways
of the universe. What a reversion from the intellectual fearlessness with which
the Greek un weary ingly interrogated nature, and wrestled with her secrets ! What
a contrast with Euripedes' hymn (c. 450 B.C.) :
Happy is the man who has learned to search into the reasons of things, and to discern
the deathless and ageless order of nature — whence it arose, how, and why.
The alchemical SChooL — The most celebrated alchemists during the twelfth and
the thirteenth centuries were Albertus Magnus, Thomas Aquinas, Roger Bacon, Arnold
Villanovanus, and Raymond LuUy. Their works serve as milestones indicating the
state of alchemy at that period. Very few important additions to chemical knowledge
were made, since the general tendency of the age was towards magic, sorcery, and
the* transmutation of the base metals into gold. Albertus Magnus and Thomas
Aquinas were Dominican friars ; Roger Bacon was a Franciscan monk ; and
Arnold Villanovanus was a university professor at Barcelona. Some of the works
attributed to these men are no doubt spurious.
Some religious orders sought to spread a knowledge of the arts and sciences,
but they unfortunately also attempted to control the progress of science in pre-
determined channels ; and the promulgation of hypotheses, or the discovery of facts
which did not harmonize with accredited authorities, or orthodox beliefs, was re-
garded as a serious offence against the State or Church. The students of alchemy
were believed to be magicians, and were supposed to be in communication with
beneficent or malignant spirits ; and, although Albertus Magnus denied this
assumption when he declared that " all those stories of demons prowling in the
regions of the air, and from whom secrets of futurity may be ascertained, are absurdi-
ties which can never be admitted by sober reason," yet, the fear and dread of magic
took complete possession of the popular mind ; even " the church service books gave
agonizing petitions for averting these dire influences, and prescribed impressive
exorcisms for thwarting the occult powers." 2 The first step to be taken by a student
of nature was thought to be to league himself with Satan by bartering his soul fo^
knowledge and occult power, and whenever a mediaeval thinker appeared to be
inspired by a love of knowledge and freedom of thought, the disease was ascribed
to diabolic agency.
One of the oddest and oldest tricks of the human mind, in ancient and modern
times, is to invoke spirits, in time of need, to explain ill-understood phenomena
In accord with the beliefs and customs of the times both Roger Bacon and
Arnold Villanovanus were prosecuted for being in league with demons, and in
1317, the inquisition of Tarragona condemned the writings of Arnold to be burned
on account of their heretical sentiments. Both Albertus ^^^gnus and Ihomas
Aquinas were astute enough to escape the severe persecution which betell man>
of their brother monks who studied the alchemical arts. ^
It must be confessed that the authorities probably had some justification tor
their attitude against the " unholy quest of alchemy," just as to-day it is necessary
46 INORGANIC AND THEORETICAL CHEMISTRY
to limit the activity of fortune-tellers, etc., by legislation. In the fifteenth century
severe interdicts against the practice of alchemy were issued in the Roman provinces,
in England, and elsewhere ; indeed, Duke Frederick I of Wiirttemberg is said to
have kept a special gallows for hanging the alchemists 3 — but the alchemists still
continued their labours.
The views of the eminent German alchemist, Albert of Bollstddt, or Albertus
Magnus (1193-1280) ,4 were mainly derived from those of Aristotle. The alchemical
writings attributed to Albertus Magnus have been shown by the scholars to be in
the main compilations from Arabian sources, although he introduced several
novelties. Albertus Magnus specially studied the union of sulphur and the metals ;
and, like the Arabian Rhases, he considered the metals themselves to be compounds
of difierent proportions of the three principles or elementals : arsenic, mercury, and
sulphur. Sulphur, said he, " blackens silver and burns the metals on account of
the affinity which it has for these substances.'- The term affinity was thus used for
the first time to designate the unknown cause of chemical action. Silver was
supposed to be the metal most closely allied to gold, so that he considered the trans-
mutation of silver into gold would be the easiest to realize. Albertus Magnus knew
how to separate the noble from the base metals by fire, and how to separate gold
from silver by aqua regia. Some suppose that the treatise on alchemy ascribed to
Albertus Magnus is spurious. The canonized scholar, Thomas Aquinas (1225-1274),5
was a pupil of Albertus Magnus. It has been said that while the master was a
student of nature and philosophy, the pupil was a student of man and society.
Both are considered to have excelled as exponents of theology rather than as students
of natural science. From the little knowledge which is available concerning the
alchemical labours of Thomas Aquinas, he would appear to have been particularly
attracted by the action of mercury on the metals — lead, tin, etc. — and he applied
the term a7nalgam to the liquid or paste which is formed when these metals are
opened up with mercury.
Among the foremost in substantial knowledge in the thirteenth century stood
Roger Bacon (1214-1294). He saw far beyond his age ; and his reputation among
his contemporaries was so great that he was styled Doctor Mirabilis. His knowledge
was thought to be uncanny; his insight was mistaken for wizardry. Roger Bacon's
knowledge of physical science was probably derived from Arabian and Greek sources,^
for no new principle has been traced to Bacon himself. S. Vogl has pointed out that,
during a great part of his life, Roger Bacon was practically without the means of
prosecuting experimental research ; and he was thwarted in his aspirations at
every turn by his superiors. It is therefore not surprising that he failed to enrich
science by any striking original discoveries. Nevertheless, his critical examination
of the science of his time was conceived in a broad philosophical spirit which showed
that he had made a great advance in the methodology of science. Bacon was
not exactly an admirer of Aristotle, for he said :
If I had all the books of Aristotle in my power, I would cause every one of them to be
burnt, because studying them is only a loss of time, and a cause of error, and a multiplica-
tion of ignorance, beyond what can be explained.
In his Ofus majus, R. Bacon emphasized very clearly the importance of scientia
experimentalis, which, in his opinion, is the mistress of all the sciences — domina est
omnium scientiarum. Indeed, he actually claimed an equal rank for observation
and experiment. True enough, towards the end of the Grecian epoch, there dawned
an era of experiment, for C. Galen experimentally investigated the nerve system,
and C. Ptolemy, the refraction of light ; consequently the experimental method
was not a new thing. R. Bacon's merit lies in having explicitly indicated the im-
portance and bearing of experiment as a universal instrument of research. The
more important scientific works of Roger Bacon are : the Opus majus written in
1266, and the supplementary Opus minor, with its introductory Opus tertiurn,, com-
pleted within a year of the publication of the Opus majus. The last-named work
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 47
contains very little about alchemy, but much more occurs in the two subsidiarv
works. • ^
Alchemy, said R. Bacon, falls into two divisions— speculative and operative
Operative alchemy includes the practical and industrial processes pursued, with
more or less wisdom, by men who have a definite purpose in view. Alkimia
speculativa treats of the transformations of matter from its simplest to its most
complicated form, and in this sense the problem of R. Bacon's speculative alchemy
approaches that of modern chemistry. Roger Bacon was necessarily ignorant of
the fundamental truths of chemical science, and he could do little more than compile
a number of empirical facts. He believed that air is the food of flame, for if a lighted
lamp be placed in a closed vessel, the flame is extinguished. Like Albertus Magnus
he supposed the best and basest of metals to differ only in the relative proportions
of their constituent parts — mercury and sulphur — and their degree of purity. He
also devoted special attention to the^ properties of saltpetre and gunpowder. Arnold
Villanovanus (1234-1312) 7 specially studied distillations, ^nd he prepared many
essential oils — turpentine, rosemary, etc. The fanatical Raymond Lully (1235-
1315) enjoyed an ephemeral reputation ; he led a turbulent restless life, and although
an enormous number of books have been attributed to him, it is certain most are
spurious.8 There is also the probability that there are two different Raymond
Lullys — one the fanatic, one the alchemist. Lully is reputed to have made spirit of
wine which he called aqua vita ardens, and he seems to have rectified it by distillation
from potassium carbonate.
This quintet may be taken to represent characteristic types of alchemists during
the twelfth century. Arnold Villanovanus ascribed any successes which he
obtained in his experimental work to the favourable position of the planets and
stars, and to suitable prayers ; these conditions seemed to him to be more im-
portant than a mastery of the controllable conditions under which the operations
were performed. This was rather unsatisfactory because no science is possible if
the phenomena under consideration are subject to the capricious wills of beneficent
or malignant spirits, for science postulates that natural phenomena are but linka
in an endless chain of cause and effect, and that in experimenting " the same
antecedents are invariably followed by the same consequents." The intellect of
man now began to assert its claim for independent thought ; and a general yearning
for progress was apparent. Learning revived in Italy, the land whence it had been
almost blotted out of existence a thousand years before. A few literary societies
appeared during the fifteenth century, and in the sixteenth century these societies
became quite numerous. Their chief work was the study of the philosophy of
Plato, and the development of the Italian language. Scientific societies were also
founded.
The invention of printing, about the middle of the fifteenth century, gave an impetus
to the pursuit of literature. There was also a spirit of social unrest. The voyage of Colum-
bus opened up the New World for those who sought new fields of discovery, fortune, or
adventure. Martin Luther's revolt was inaugurated in 1517 by the posting of his thesis
upon the church door at Wittemburg. That versatile genius Leonardo da Vinci ( 1 452-1 51 9),
whose compendious manuscripts were so long thought to be written in secret script because
written backwards, has been but recently appreciated, and his notes in part transcribed tind
edited. He was a pioneer of the modern spirit of investigation and practised the inductive
method a century before Francis Bacon. The foundations of astronomy, mechanics, and
physics were laid about this time ; Nicolas Copernicus had published his De revoltUionibus
orhium ccelestium in 1543. During the next fifty years fuller and more accurate data were
compiled by Tycho Brahe (1546-1601). About 1608, the astronomical Don Quixote,
Johann Kepler (1571-1630), published voluminous works » which have been styled " a most
singular medley of soimd thoughts and vmmitigated nonsense." Kepler, however, did submit
his ridiculous conceptions to the test of observation, and rejected those which did not stand
the trial. Among the wildest of guesses on the motions of the planets and their satelhtes,
he discovered those truths which have long been known as Kepler's laws. Galileo Galilei
made important experiments on the laws of motion, towards the end of the sixteenth
century ; and a century later, Isaac Newton demonstrated the aU-embracmg law of
gravitation in his epoch-makmg Philosophia naturalis prmcipta mathematica (London, 1685).
48 INOKGANIC AND THEORETICAL CHEMISTRY
A Latin compilation on technological chemistry, entitled : Compositiones ad
tingenda, was published towards the end of the eighth century, and about the
tenth century one entitled : Mappce clavicula. These works contain recipes for
industrial processes closely resembling those of the ancient Greek papyri. The
term vitriol for impure ferrous sulphate was used for the first time in the eighth-
century work. Reference is made to the use of the hydrostatic balance in the
analysis of alloys of gold, and this has been taken to show that the knowledge of
this instrument did not pass through Arabian channels to Western Europe, but
came direct from the writings of Archimedes of Syracuse (287-212 B.C.), which
were carried west by the fugitives from Constantinople after its capture by the Turks
in U53.
A large number of alchemists — P. Bonus, N. Flamel, Isaac of Holland, G. Ripley,
T. Norton, T. Charnock, E. Kelley, John Dee, M. Sendibogius, M, Maier, J. Boehme,
T. Vaughan — who wrote under the nom de plume, Eupenius Philathes — and another
— who wrote under the pseudonym, Erenaeus Philathes — laboured with some skill,
between the fourteenth and seventeenth centuries, although the alchemical school
was perhaps at its zenith in the fifteenth century. About this time there were three
different types of alchemist. The first or bookish type spent his time commenting
upon, elucidating, or unconsciously obscuring the views of the earlier writers ; this
type might also include the mystical chemists who hinted at a secret doctrine of
a spiritual order. The second or mercenary type hoped to find unlimited riches
when he had succeeded in converting the base metals into gold ; and the third or
investigating type sought to discover the properties and combinations of the metals,
and the best means of extracting them from their ores. The last formed the
prototype of the modern chemist, although representatives of all three types still
survive. The majority of the alchemists were diligent experimenters, and although
they worked in a stupendous chaos of phenomena, their indefatigable zeal will
long be remembered for the multitude of primary facts which they discovered,
even though the names of the discoverers are forgotten. The alchemists crystallized
and calcined, digested and distilled, filtered and fused, just as chemists do to-day.
Auguste Comte i^ has said that it is difficult to understand how the early investi-
gators could have had the energy and perseverance to discover the chief chemical
phenomena had they not been constantly incited by unbounded hopes arising from
their chimerical notions of the constitution of matter. The alchemists were indeed
stimulated and guided in their work by a logical system of hypotheses. For
instance, they accepted the older prima materia hypothesis of the ultimate constitu-
tion of matter. The changes which were observed in the different forms of matter
appeared as the outer clothes of an unchangeable all-pervading essence. The
qualities of the elements, not their essences, are changeable ; some of these qualities
are more easily removed than others, thus the four elements were regarded as
firmly clinging coverings, while heat and cold, moistness and dryness, were more
easily removed. The different varieties of matter were the different vestments
or wrappings of the one universal entity, the quintessence of things. The universal
essence was regarded as the perfect thing — The One Thing. This one thing was
given many different names — e.g. the stone of wisdom or the philosopher's stone,ii
a term which, according to M. Berthelot, appeared in alchemical writings about
the seventh century, although the central idea is much older.
The property of matter which enabled it to withstand the action of fire was
attributed to its possessing the quality of fixidity later symbolized by salt ; if it
possessed the principle of volatility — later symbolized by mercury — the substance
would volatilize ; if it possessed the principle of combustibility — later symbolized
by sulphur — the substance would burn ; the principle of redness gives matter a
red colour ; and so on. To the Romans, lead and tin were differently coloured
varieties of the same metal, and called dark and light lead respectively. Thus, the
variations in the different forms of matter were supposed to depend on the qualities
or principles with which it was endowed. The chemical properties of matter were
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 49
but dimly recognized even in the Middle Ages ; and the differences between bodies
were considered to depend essentially on their physical qualities. Hence, it was
assumed that the properties of a body could be modified by the abstraction or
addition of qualities and forms. It was argued that just as the hardness, colour
fusibility, and other properties of certain metals can be altered, so must it be possible
to change all the properties of one metal into those of another, and thus produce
a veritable transmutation. Consequently, the alchemists believed in the transmu-
tation of the metals.
The idea of transmutation occurs in the pre-Christian Greek writings, but the
idea of transforming the base metals into gold developed near the beginning of the
Christian era when the Egyptian goldsmiths seem to have carefully studied the
diplosis — 8t7rA(oo-t? — or doubling of gold ; in other words, the art of increasing
the weight and bulk of gold by adulteration with cheaper metals. In M. Berthelot's
Collection des anciens alchimistes grecs (Paris, 1887-8) quite a number of works on
this subject are cited — one by Moses (not the law-giver of Israel, though possibly
by one who adopted this name as a nom de plume) is entitled Trcpi StTrAwo-ews xP^a-ov,
or The diplosis of gold, is preserved in the collection of alchemical writings at Venice ;
another by Cleopatra (not the celebrated queen), entitled KXcoTrarpa? xP^a-oTroaa
{c. 50 B.C.) or The chrysopoeia of Cleopatra, is in the collection at Ley den ; etc.
The last-named manuscript deals with the preservation of beauty ; with weights
and measures ; and with the making of gold. In the Collection des anciens alchimistes
grecs there are drawings of digesters, aludels, alembics, and a variety of apparatus
for distillations, and of water baths i^ for heating in the laboratory. It may be
added that the water bath was in use, 500 B.C., in Egypt, and was caUed the hath of
Isis : the name was later altered to the bath of Mary — or the bain marie, as it is
still called in France — after an Egyptian Jewess, Mary, a writer on alchemical
subjects.
In the opinion of M. Berthelot 13 the idea of alchemy, as a method for trans-
muting the base metals into gold, was a development from the fraudulent practices
of the goldsmiths in Egypt as an accidental accretion to chemistry, either from a
misreading or misunderstanding of ancient manuscripts. As a result, the working
recipes for adulterating gold were regarded as directions for the transmutation of
the metals. This is shown by the fact that some of the Egyptian papyri — e.g. the
Leyden papyrus — contain elaborate prescriptions for the falsification of the precious
metals, and these recipes reappear later obviously copied as formulae for the trans-
mutation of the base metals into gold. Hence H. Kopp could say : Die Geschichte
der Alchemic ist die Geschichte eines Trrtums.
According to the transmutation hypothesis, the baser metals were diseased and
imperfect ; gold was the most perfect of the metals. The process of transmutation
consisted in healing and ennobling the diseased metals. It was postulated that a
stone of wisdom, or philosopher's stone, could be found which would heal the
diseased metals, for, said W. Salmon, in his Bibliotheque des philosophes chimiques
(Paris, 1672-8), the philosopher's stone is " the universal medicine for all imperfect
metals, it fixes that which is volatile, purifies that which is impure, and gives colour
and lustre more brilliant than nature herself." This hypothesis is quite legitimate,
but the questions which the alchemists asked from nature appear to have been too
profound; they could not understand her responses. The idea of a universal
'medicine for diseased metals was extended and the philosopher's stone was invested
with all kinds of mystic properties by extravagant visionaries. The Arabian
pharmaceutists supposed it to have the power of elevating man's diseased and sickly
body into a state of golden health, and thus arose the idea of an elixir of life i* or
elixir vitce—oi universal medicine capable of curing all curable diseases, and which
later developed into an elixir of immortality. Still later, in the old age or dotage
of alchemy, the alchemists sought a philosopher's stone which would preserve
health, raise the dead, rejuvenate the old, make cowards brave, etc. The en-
thusiastic visionaries gave still further play to their fancies, and Paracelsus miagmed
VOL. I. *
50 INORGANIC AND THEORETICAL CHEMISTRY
an alkahest or universal irresistible solvent which would dissolve every substance
with which it came into contact ; there was also the perpetual lamp which would
burn for ever ; 15 perpetual motion ; etc. The series of facts which nature revealed
to the first experimenters in chemistry were so unlike anything already known that
the ordinary principles of belief were shaken or subverted ; and their mind became
so exceedingly credulous that J. Play fair, in an essay On the progress of mathematical
and physical science (Edinburgh, 1853), could say that one who professed to be in
search of truth ever wandered over the regions of fancy in paths more devious and
eccentric.
The medico-chemical or iatro-chemical school.— In the sixteenth century,
alchemy received an impetus in another direction — medicine. Philip Hoehener,
who, on commencing his professional career, styled himself PhiUppus Aureolus
Theophrastus Paracelsus Bombastus, was born at'Zurich in 1493,i6 and he seems to
have developed the amazing arrogance, insolent presumption, and swelling vanity now
implied by the term " bombast." It has been pointed out that it is not generally
the calm, cautious, common-sense men who do the new and great things of the
world, for it seems to require vigorous impulses and certain extravagances of
character to institute drastic reforms. W. Ostwald, in his Grosse Manner (Leipzig,
371, 1909), attempted to arrange men of genius in two classes which he called respec-
tively romanticists and classicists. The classification is based on mental reaction
velocity — or mental temperature, so to speak. The romanticist has a high and
the classicist a low mental reaction velocity. The latter is inclined to be phlegmatic
and melancholic, and the former sanguine and choleric. The romanticist with his
agile mind reads everything, he is interested in everything and everybody, and,
as a result of his enormous consumption of facts, he writes a great deal. On the
other hand, the classicist works more silently and more alone, and he writes com-
paratively little. W. Ostwald would undoubtedly have classed the wayward erring
Paracelsus among the romanticists. Paracelsus seems to have combined in himself
the personality of two men : there is the daring reformer and incessant observer,
and there is also the mystic hypnotized by conceit who claimed that he was privileged
above all others, and received knowledge direct from God or by inspiration from
the Divine.
The works of Paracelsus embrace many subjects—chemistry, botany, philosophy,
physics, astrology, theosophy, magic, and most important of all, medicine. His
style is generally clear, and characterized by energy and vigour, but suffused with
mysticism. Paracelsus maintained that each disease has its own specific symptoms
and cause, and must be combated by specific remedies— every disease, said Paracelsus,
must have a remedy. The development of this idea led to his being called the
Luther of medicine since, previously, all diseases were considered to result from an
excess of phlegm, bile, or blood. Paracelsus introduced many new remedies, and
he directed the attention of medical men to the importance of chemical preparations
and medicines ; he taught that the direct object of chemistry is not to make gold,
but to cure disease ; and he gave a bias to the quest for the essences or quintessences
of things — e.g. he investigated the active principles of plants which he used medicin-
ally in the form of tinctures, extracts, essences, etc. — and thus he prepared tincture
of opium or laudanum.
There is little evidence to show that Paracelsus contributed any important
discovery to chemical science. There are, however, references in his writings to
zinc and bismuth which he characterized as bastard metals because, though resem-
bling the metals in general appearance, they lacked the characteristic ductility and
malleability of the seven metals known from ancient times. Paracelsus prepared
arsenic acid by the action of nitre on arsenious oxide ; he discriminated between
the alums and vitriols by showing that the former had an earth and the latter a
metal as base ; he prepared copper amalgam by the action of mercury on copper
precipitated from its sulphate by iron ; he noted the development of a gas during the
action of oil of vitriol on iron ; he used an infusion of nut-galls for detecting iron
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 51
in mineral waters ; he mentioned the bleaching action of the fumes from buming
sulphur on red roses ; and he described the separation of hydrochloric from nitric
acid by means of silver. Paracelsus promulgated some astounding, even childish
hypotheses on the slenderest of evidence, so that the wildest vagaries were
promulgated by the followers of the mystic Paracelsus during the succeeding
century. The great merit of Paracelsus lies in his having undermined faith in the
traditions which had previously corrupted and demoralized the thought and works
of most of the earlier alchemists.
Soon after Paracelsus' degraded death in 1541, the alchemists seem to have
parted ways. The palseo-alchemical school— TraAaios, ancien1>— still pursued the
transcendental and ever-vanishing images of alchemy which could not be brought
into harmony with the inflexible world of fact. It is characteristic of a science in
its early stages, said S. Brown (1843) and A. N. Whitehead (I916),i7 to be both
ambitiously lofty in its aims, and trivial in its handling of details. This statement
is very true of the mediaeval alchemists, and " their successors still tried to scale to
heavenly heights ; but their vitality was gone and they degenerated into fanatical
inanities of no historical significance ; and their compilations are usually mystical
anonymities fathered on to the potentates of old." The neo-alchemical school —
v€ops, new — soon renounced the unattainable sublimities of the earlier alchemists,
they dropped the Arabian al, and sagaciously pursued the sober and attainable
aims of a truer chemistry. They sought knowledge, not gold ; they confined their
attention to phenomena and reactions which could be realized experimentally ;
and they assiduously devoted themselves to the discovery of primary' facts, without
dissipating much energy on attempts at transmutation. In fine, they were
undoubtedly the working chemists of their day, and they laid the foundations of
experimental chemistry.
Masses of information were rapidly accumulated by George Agricola (1491-1555)
— the father of metallurgy, and author of the painstaking De re metallica (Basil,
1556), 18 on mining and metallurgy ; by Andreas Libavius (1540-1616), i^ the dis-
coverer of tin tetrachloride or liquor fumans Libavii ; and by AngeloSala (1575-
1640), 20 who severely criticized the old mystic hypotheses, and who would have
chemists cease from trifling with sublimities. To the alchemists who professed
to extract from antimony a mercury which would effect the great transmutation,
A. Sala said : " Show me only one drop of your wonderful mercury and I will
believe you ; but meanwhile I am deaf to your nonsensical claims." A. Libavius
proved that the acid obtained by distilling alum and green vitriol (ferrous sulphate)
is the same as that obtained by burning sulphur with saltpetre ; he studied the action
of nitric acid on sulphur ; and prepared artificial gems by tinting glass with metal
oxides. A. Sala specially studied ammonia ; and he synthesized ammonium
chloride by treating ammonium carbonate with muriatic acid. A. Sala recognized
that iron is not changed to copper when dipped in a solution of blue vitriol, for he
saw that the copper comes from the blue vitriol. Paracelsus had given a bias to
alchemy which led its followers to study diligently the preparation of medicines
rather than pursue an emasculated alchemy in the quest for the unattainable. The
new school of medico-chemists and pharmaceutists made a mistake in attempting
to explain the changes and processes which occur in the human organism by fanciful
hypotheses founded upon their ignorance of the facts. Paracelsus himself seems to
have made the childish assumption that a demon named Archseus resided in the
stomach, and changed bread into blood, etc.
The talented J. B. van Helmont (1577-1644) of Brussels, began his career an
enthusiastic alchemist, and ended a worthy chemist ; he also speciabzed m medicme,
and helped to carry on the medical reform inaugurated by Paracelsas. Conse-
quently, his posthumous collected works ^i—Ortus medicin(B (Amsterdam, 1648)—
appear to be both alchemical and chemical. J. B. van Helmont is particularly
noted for distinguishing clearly between air and gases ; for his work on carbon
dioxide which he did not distinguish sharply from sulphur dioxide, ammonia, and
52 INORGANIC AND THEORETICAL CHEMISTRY
nitrogen peroxide ; for wholeheartedly advocating Thales' doctrine that water is
the 'prima materia out of which all things are made — although Paracelsus had some-
thing to say in the same direction ; and for his denying the elemental nature of fire
which he considered was not a material substance at all. J. B. van Helmont is also
noted for first using melting ice and boiling water as fixed points in thermometry ;
for his use of the term saturation to signify the combination of an acid with a base ;
for emphasizing the imperative claims of the balance for a premier place in the
chemical laboratory ; and for showing that although a metal can enter into many
combinations, yet it does not lose its own peculiar nature since it can always be
again separated unchanged— no metal can be obtained from a salt if it is not already
present therein. The clear recognition of this fact was a necessary condition for
progress in chemistry. It was previously supposed that a change in the appearance
of a metal constituted a veritable transmutation. It was not until the chemical
properties had been studied that it became possible to realize that the differences
between the various kinds of matter depend on differences in their chemical com-
position, and are not produced solely by the addition or abstraction of certain
qualities or principles.
The famous J. R. Glauber (1604-1668) was a laborious and diligent chemist
who studied the preparation and properties of several salts — e.g. he prepared blue
vitriol by the action of sulphuric acid on verdigris ; various acetates by the action
of wood vinegar on alkalies, earths, or metals ; ammonium sulphate, or as he called
it secret sal ammoniac, by the action of sulphuric acid on sal ammoniac ; ammonium
nitrate which he called nitrum flammaris ; etc. J. R. Glauber prepared nitric acid
by distilling a mixture of nitre and alum or sulphuric acid ; and hydrochloric acid
by distilling common salt with sulphuric acid. The term muriatic acid for this
acid was also coined by him. The residue in the last-named operation is known
to this day as Glauber's salt, or sodium sulphate, which J. R. Glauber regarded as a
most wonderful salt — sal mirahile — for he ascribed to it extraordinary curative
properties when used as a medicine. He said :
This salt is the beginning and end of all things, and it increases and exalts their powers
and virtues ; it is the true universal medicine ; not that I would have any man persuade
himself, that in these words I would assert immortality, for my purpose tendeth not thither,
seeing that I am not ignorant there is no medicine against death.
J. R. Glauber 22 also studied the products of the distillation of bones, and of wood.
He described the preparation of pyroligneous spirit or wood vinegar — acetum
lignorum — by the destructive distillation of wood, and stated that it could be made
as virtuous as wine vinegar — acetum vini — by re-distillation. He also noted the
preservative action of wood tar. J. R. Glauber recognized the law of chemical
exchange — double decomposition — in the action of sulphuric acid on common salt,
and of potassium silicate on gold chloride. He said that the potash of the silicate
neutralizes the acid of the gold salt, so that the silica and gold are both deprived
of their solvents, and are precipitated.
F. Sylvius de la Boe (1614-1672), C. Glaser (1615-1673), 0. Tachen (1620-1690),
Robert Boyle (1627-1691), J. Kunckel (1630-1715), N. Lemery (1645-1715), J. K.
Dippel (1673-1734), and many other interesting chemists flourished during this
period. Their work will be discussed more specifically later on. Most of these
men believed in the alkahest, the philosopher's stone, and in the transmutation of
the metals. Their faith may have been largely founded upon J. B. van Helmont's
assurance that he had verily witnessed the transformation of mercury into gold.23
Fortunately, these men were indefatigable workers, and did not fritter away much
time on fantastic fictions. Said S. Brown (1851) : It is never the originators of a
great but useful scientific error, nor yet its true believers, but it is the indolent
perpetuators, who will not move to the music of a new fact and the new time, that
are ridiculous, shifty, ambiguous, and not respectable.
Some important works, written under the worn de plume Basil Valentine,
probably in the sixteenth or seventeenth century, were for a long time wrongly
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 53
supposed to have been the work of a fifteenth-century Benedictine monk, before
Paracelsus. On account of the many parallel statements in the writings of
Basil Valentine and Paracelsus, J. B. van Helmont and others assumed that the
latter was indebted to the former for many of his ideas and facts. The truth is more
probably the direct converse of this, and the imposition of Basil Valentine as a pre-
Paracelsian writer has been called " a seventeenth-century hoax." Anachronisms
in the supposed writings of Basil Valentine show that these could not have been
written so early as the fifteenth century. In common with the later views of H. Kopp,
J. Ferguson, K. Sudhoif, M. Berthelot, F. Strunz, C. W. G. Kastner, etc., W. Hommel
says that all the evidence indicates that the name was a pseudonym for Johann
Tholde who, in 1603-1604, first published the works of Basil Valentine, and pretended
that he had translated them from an old Latin manuscript which he had discovered. 2*
The writings are characterized by some clearness, particularly when describing the
results of experiments. The masterpiece, Triumph-Wagen des Antimonii, published
at Leipzig, in 1624, seems to include almost all that was known about antimony up
to the seventeenth century. Basil Valentine precipitated gold from its solution by
the addition of mercury ; copper from its solution by means of iron ; and iron from
its solution by potash ; he obtained metallic mercury by the distillation of corrosive
sublimate with chalk ; and he is sometimes regarded as one of the founders of
analytical chemistry. A. number of other works are attributed to the same writer.
Special attention should be directed to Robert Boyle, who, more than any
previous worker, emphasized the importance of the science or, as he called it, the
philosophy of chemistry. 25 He has accordingly been called "the father of chemistry,"
although the same cognomen has been applied to several others. R. Boyle claimed
that those who had previously studied chemistry regarded it as a means of pre-
paring medicines or improving the metals, while he considered the art neither as
a physician nor as an alchemist, but rather as a philosopher. Chemistry, he claimed,
had been too often practised by illiterate arrogant impostors who wrote in a language
which could scarcely be understood by a philosopher.
Without seeking the grand elixir, chemistry may greatly promote om: knowledge of the
works of nature. It is certain that some meliorations of metalline and mineral bodies may
be made, useful medicines prepared, and various productions serviceable in particular
trades may be obtained by means of chemistry, and therefore this subject may be studied
to advantage.
R. Boyle further claimed that he had a larger view in cultivating the science— no
less a purpose, indeed, than the general advancement of natural philosophy.
Chemistry is eminently conducive to extend the empire of mankind by enlarging our
views, and giving us a command of nature. Just as the Bologna stone would never become
luminous unless it were chemically prepared, so many natural bodies would never afford
light to philosophy xmless it be struck to them by chemical operations.
In his remarkable Sceptical Chymist (Oxford, 1661), Robert Boyle introduced the
modern conception of an element, and dropped the four principles or elements of
the peripatetic school, and the prima tria of the alchemists. In 1 660, Boyle designed
a new air pump based upon that of 0. von Guericke. Between 1660 and 1672,
R. Boyle tried the effect of a reduced pressure upon the properties of many substances,
and he made many experiments on the elasticity of gases. He demonstrated what
is now known as Boyle's law ; he showed that air expanded by heat (1662) ; he
studied the action of alkalies on vegetable tinctures (1663) ; and attempted a classi-
fication of substances into acids, bases, and salts (1680). He also studied the cal-
cination of metals in sealed vessels (1673), and assumed that during the calcmation
"a subtle fluid is able to pierce into the compact and solid bodies of metals
imparting to them " no despicable weight." Robert Boyle had a clear conception
of the ponderable character of air, for he several times attempted to determine its
weight, and showed that the weight of a bladder of air appears to be greater in
vacuo than in air.
54 INORGANIC AND THEORETICAL CHEMISTRY
References.
^ C. Krumbacher, Oeschichte der byzantinischen Literatur, Miinchen, 1897.
* A. D. White, A History of the Warfare of Science and Theology, London, 1896 ; P. Lacroix,
Les arts au moyen dge et a Vepoque de la renaissance, Paris, 1869.
» V. Weech, ZeU. Geschichte Oherrheins, 25. 468, 1873.
* F. Pouchet, Histoire des sciences naturdles au moyen dge. Alberius Magnus et son ^poqu£,
Paris, 1853 ; W. J. Townsend, The Great Schoolmen of the Middle Ages, London, 1881 ; Albertua
Magnus, Opera Omnia, Lugduni Batavorum, 1653 ; Theatrum chemicum, Argentorati, 2. 23,
423, 1615; 4. 809, 825, 841, 1617; H. Fronober, Die Lehre von Materie und Form nach Albert
dem Grassen, Breslau, 1909.
^ R. B. Vaughan, St, Thomas of Aquin : His Life and Labours, London, 1871-2 ; Theatrum
chemicum, Argentorati, 3. 267, 278, 1617 ; 4. 960, 1619 ; 5. 806, 1661 ; Secretu alchemice. Colon,
1679 ; Thesaurus alchemice, Lugduni Batavorum, 1602 ; De esse et essentia mineralium, Venetice,
1488.
^ S. Vogl, Die Physik Roger Bacon, Erlangen, 1906 ; Theatrum chemicum, Argentorati, 2.
377, 1615 ; 5. 834, 1619 ; Fr. Rogeri Bacon, Opera quoedam hactenus inedita (J. S. Brewer),
London, 1859 ; H. G. Bridges, The Opus majus of Roger Bacon, Oxford, 1897 ; H. F. Wiistenfeld,
Oeschichte der arabischen Aerzte, Gottingen, 1840 ; E. Charles, Roger Bacon, sa vie, ses ouvrages,
ses doctrines, d'apres des textes inedits, Bordeaux, 1861 ; R. Bacon, Thesaurus chemicum, Franco-
furti, 1620; Fr. Rogeri Bacon, Ordinis minorum, Opus majus (S. Jebb), London, 1733;
R. Adamson, Roger Bacon: the Philosophy of Science in the Middle Ages, London, 1876;
J. E. Sandys, Roger Bacon, London, 1914.
' Arnoldus de Villanova, Opera Omnia, Lugduni Batavorum, 1520 ; Theatrum chemicum,
Argentorati, 1. 128, 1613 ; 2. 108, 1615 ; 3. 128, 137, 1617.
* P. 0. Keicher, Raymundus Lullus und die Grundzuge seines philosophischen Systems aufgezeigt
als ein Reaktionsversuch gegen die arabische Philosophic, Miinster, 1 908 ; Theatrum chemicum,
Argentorati, 3. 165, 295, 1617; 4. 1, 135, 171, 515, 1619; Opera alchemia, London, 1673;
Opera omnia, Argentorati, 1677,
* J. Kepler, Nova astrcmomia seu physica codestis tradita commentares de motibus stellce martis,
Prague, 1609; Harmonices mundi, Linz, 1019; G. Galilei, Discorsi e dimostrazioni matematicJie,
Ley den, 1638.
^° A. Comte, Cours de philosophic positive, Paris, 3. 7, 1864.
^1 M. M. P. Muir, The Chemical Essence and the Chemical Elements, London, 1 894 ; J. C.
Draper, Amer. Chemist, 5. 1, 1874 ; H. Kopp, Quart. Journ. Science, 5. 21, 1868.
^2 E. 0. Lippmann, Abhandlungen und Vortrdge zur Geschichte der Naturmssenschaften, Leipzig,
2. 185, 1913.
^3 M. Berthelot, Les origines de Valchimie, Paris, 1885 ; Introduction a V etude de la chimie des
ancient et du moyen dge, Paris, 1889 ; La chimie au moyen dge, Paris, 1893 ; M. Berthelot and
P. E. Ruelle, Collection des anciens alchimistes grecs, Paris, 1887-8; H. Kopp, Die Alchemic in
dlterer und neuer Zeit, Heidelberg, 1886 ; Beitrdge zur Geschichte der Chemie, Braunschweig,
1869; Veber der V erf all der Alchemic und die hermetische Gesellschaft, Giessen, 1847; G. P.
Nenter, Berichte von der Alchemic, Niirnberg, 1727 ; E. A. Hitchcock, Remarks upon Alchemy
and the Alchemists, Boston, 1857 ; H. S. Redgrove, Alchemy ; Ancient and Modem, London,
1911 ; G. Letz, Die Alchemic, Bonn, 1869; E. 0. von Lippmann, Entstehung uvd Ausbereitung
der Alchemic, Berlin, 1919.
1* J. Gildenemeister, Zeit. deut. morg. Qes., 30. 534, 1876 ; S. Brown, Essays, Edinburgh, 1858.
15 C. H, Bolton, Monthly Journ. Science, (3\ 9. 715, 1879.
1* A. M. Stoddart, Life of Paracelsus, London, 1911 ; A. E. Waite, The Hermetic and Alchemical
Writings of Paracelsus the Great, London, 1894 ; F. Hartmann, The Life of Philippus Theophrastus
Bombast of Hohenheim known hy the name of Paracelsus, London, 1896 ; J. M. Stillman, Monist,
27. 390, 526, 1917 ; F. Mook^ Theophrast^is Paracelsus— eine kritische Studie, WUrsburg, 1876 ;
R. Netzhamraer, Theophrastus Paracelsus, Einsiedeln, 1901 ; J. Ferguson, Encyc. Brit., 18. 236,
1885 ; H. Magnus, Paracelsus der Ueberartz, Breslau, 1906 ; M. Neuburger and J. Pagel, Handhuch
der Geschicht der Medizin, Jena, 3. 403, 1905 ; S. Brown, Essays, Edinburgh, 131, 1858 ; W. Luzi,
Das Ende des ZeiUdters der Alchemic und der Beqinn der iafrochemischen Periode, Berlin, 1892;
R. Browning, Paracelsus, London, 1835.
" A. N. Whitehead, B. A. Rep., 355, 1916 ; S. Brown, Essays, Edinburgh, 1. 131, 1868.
1* G. Agricola, De re metallica, London, 1912.
*• A. Libavius, Alchemia, Francofurti, 1595 ; Opera chymica, Francofurti, 1604.
^ A. Sala, Opera medico-chymica omnia, Rothomagi, 1650.
21 M. Meslens, Note historique sur J. B. Hehnont, Paris, 1874 ; J. B. Tan Helmont, Works,
London, 1664.
2 2 J. R. Glauber, Works, London, 1689.
2' G. de Mengel, Journ. Alchem. Soc., 1. 49, 1913 ; K. C. Schmieder, Oeschichte der Alchemic,
Halle, 1832.
2* M. Berthelot, Introduction a V etude de la chimie des anciens etdu moyen dge, Paris, 279, 1889 ;
K. Sudhoff, Beitrdge aus der Geschichte der Chemie dem Geddchtniss von G. W. W. Kahlbaum, 254,
1909 ; F. Strunz, Theophrastus Paracelsus seine Leben und seine Persohnlichkeit, Leipzig, 30,
I
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 55
1903 ; K. C. Schmeider, Geschichte der Alchetnie, Halle, 1832 ; C. W. Kestner, MedicinMches
GeUhrten- Lexicon, Jena, 1740; C. W. G. Kastner, Beytrdge zur Begrnndung eines wissen-
schaftlichen Chemie, Heidelberg, 1807; J. M. StiUmann, Pop. Science Monthly, 81. 591, 1912-
W. Hommel, Zeit. angew. Chem., 32. 73, 1919 ; H. Kopp, Geschichte der Chemie, Braunschweig*,
1. 74, 1843; Ansichten uher der Aufgabe der Chemie, Braunschweig, 110, 1875; Beitrdge zur
Geschichte des Chemie, Braunschweig, 3. 112, 1875 ; Die Alchemie, Heidelberg, 1. 29, 1886 ; J. Fer-
guson, Bibliotheca Chemica, Glasgow, 1906; B. Valentine, The Triumphal Chariot of Antimonv
I^ndon, 1893 ; P. S. Wellby, Journ. Alchem. Soc, 2, 91, 1914.
25 Robert Boyle, Works, London, 1744; The Philosophical Works, London, 1725; The
Sceptical Chymist, Oxford, 1661 ; H. B. Dixon, B. A. Rep., 594, 1894; T. E. Thorpe, Essays in
Historical Chemistry, London, 1, 1894.
§ 13. The Evolution of Ideas regarding the Nature of Calcination
Let all the greatest minds in the world be fused into one mind and let this great mind
strain nerve beyond its power ; let it seek diligently on the earth and in the heavens ; let
it search every nook and cranny of nature ; it will only find the cause of the increcised
weight of the calcined metal in the air.^ — Jean Rey (1630).
The principle operations of the earlier chemists were performed by fire, and
one of the many names applied to chemistry in its early days was Pyrotechnia —
TTv/o, fire ; rcxi^^y, art. Calcination has always been one of the most important
operations in the chemical laboratory. Paul de Canotanto,i about the middle of
the fifteenth century, defined this operation as involving " the incineration of the
metals, or the destruction of the igneous principle."
The term calx (calcis) is the Latin word for lime, but the meaning was extended by the
alchemists to anything produced in the same way as quicklime — namely, by roasting to a
powder or friable substance. The operation of heating or roasting was called calcination.
Consequently, as the Latin Geber expressed it in his De alchemia, calcination is the pulveri-
zation of a thing by fire by the deprivation of the humidity consolidating its parts— in
illustration, the ash of wood, the oxide of a metal, and the ignited residue of a substance
dissolved in acid Were all calces. The alchemists regarded the calx as the purest and most
refined residuum of a substance which remained after the coarser parts had been dispelled
by heat.
It was probably known very early that limestone loses weight during its con-
version into a calx, and it came as an incredible surprise to find that an
increase in weight occurs when the metals are converted into calces. Near the end
of the fifteenth century— November, 1489— P. Eck de Sultzbach 2 was probably
the first to demonstrate experimentally that when a metal is calcined in air, the
resulting calx — or cineris fixi, as he called it — is heavier than the original metal.
He also showed that an amalgam of silver and mercury increased in weight 50 per
cent, when heated for eight hours in air. The increase in weight, which many later
observers also noticed, seems to have puzzled the earlier chemists. P. Eck de Sultz-
bach attributed the increase to the union of a spirit (gas) with the metal ; and,
as will soon appear, he was nearly right. No notice seems to have been taken of
P. Eck de Sultzbach's surmise, and many probable and improbable explanations of
the increase in weight, and of the change in the appearance of the metal, were made
during the sixteenth and seventeenth centuries. Two sets of hypotheses now
struggled for existence. j j r
One set of hypotheses assumed that the metals are naturally compounded of
a substance lighter than air which buoys up the metals, so to speak, against gravi-
tation ; during calcination this component is driven from the metal and the caJx
remains. Thus, H. Cardan, in his book De suUilitate (Basil, 1553), stated :
The metal during calcination dies, for the celestial l^eat^aior ccKi.^t.^--w^ich gav^^^^^
life and rendered it light, is dissipated, and the metal consequently becomes heavier during
calcination.
Paracelsus expressed a similar idea a short time previously : ^^^''^^'''t^u'lhU
separates watery moisture, fat, natural heat, odour, and whatever else is combus ible
Accordingly, terms like terra damnata and caput mortuum were applied to the rebidues
56 INORGANIC AND THEORETICAL CHEMISTRY
left after the spirit had been driven from the metals by calcination, and the residua
were often symbolized pictorially by a skull and cross-bones.
In another set of hypotheses, it was assumed that something ponderable is
absorbed by the metal. R. Boyle attributed the increase in weight to " the
arresting of igneous corpuscles/' and N. Lemery,3 to the assimilation of corpuscles
de feu by the metal. In R. Boyle's -essay. Fire and flame weighed in the balance
(London, 1672), a number of experiments are described showing the actual gain in
weight which occurs when metals are calcined in air ; thus, an ounce of copper
filings gained 49 grains in two hours, and an ounce of lead gained 28 grains in the
same time. R. Boyle inferred that *' glass is pervious to the ponderous parts of
flame " because tin or lead are partially calcined when heated in hermetically sealed
vessels ; and he stated that the increase in weight arises from the assimilation of
the " extinguished flame " by the calx. It is rather remarkable that R. Boyle did
not attribute the increase in weight to the action of the air on the heated body,
because, shortly afterwards, in an essay entitled Suspicions about some hidden
qualities of air (London, 1674), apparently following R. Hooke's experiments, q,v.j
R. Boyle suggested that " air contains some odd substance, either of a solar, astral,
or other foreign nature ; on account whereof the air is so necessary to the sub-
sistence of flame ; " and he further added that " this substance is not improbably
a volatile nitre akin to that which seems so necessary for the maintenance of other
flames." In opposition to H. Cardan, Boyle also says that the calx of a metal must
be the metal plus, not minus, something acquired during calcination, and not its
terra damnata. J. Kunckel (1677), J. J. Becher (1690), J. Romberg (1700),* and
others also attributed the increase in weight of a metal during calcination to the
absorption of what J. Kunckel called particulce ignicB. In an analogous manner,
0. Tachen (1666) & assumed that the increase is due to the absorption of an acid
existing in the flame, and he found that when lead burns to red lead, it increases
its weight one-tenth, and returns to its former weight when reduced to the metallic
state. H. Boerhaave (1732) ^ must have suspected that something was wrong,
since he kept mercury at a slightly elevated temperature for fifteen years in order
to find if there was any increase in weight due to the absorption of the alleged fire
particles ; and, in opposition to Boyle's hypothesis, no increase due to this cause
could be detected. He also demonstrated that the weight of certain metals —
e.g. silver — was the same whether at ordinary temperatures or at a red heat.
A. Cfesalpin, in his De metallicis (Romse, 1596), summarily dismissed the subject
by assuming that the increase in weight is due to the deposition of soot in the interior
of the metal during calcination, and others supposed the increase was due to the
retention of the vapours of charcoal, or the volatile salt of charcoal or the matter
removed from the calcining vessel. J. Hartmann, in his Praxis chymiatrica (Lipsise,
1625), showed that the increase could not be due to the assimilation of soot, or the
vapours of charcoal, because antimony increased in weight when heated in the focus
of a burning lens with sunlight ; and N. le Febvre ^ supposed that when the metal
is calcined by means of a burning glass, the increase in weight is due to the absorption
of the matter of light, which J. Mayow called particulce niti-o-aercB, and which were
supposed to be derived, not from the air but from the sun, which he regarded as a
chaos of these particles.
Jean Rey appears to have been the first to critically examine the different
hypotheses which had been proposed to explain the increase in weight which occurred
when the metals are calcined. J. Rey's work was published in an obscure pamphlet
entitled Essays de Jean Rey, docteur en medicine, sur la recherche de la cause pour
laquelle Vestain et le plomb augmentent de poids quand on les calcine (Bazas, 1630),^
which at that time does not seem to have attracted much attention from those
interested in the subject, since the discovery of the pressure of air, shortly after-
wards, diverted the minds of investigators away from a study of the chemistry of
air.
1. The facts. — In order to clarify the mind, the facts must be reviewed.
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 57
Investigators of nature, said D. Sennert us in hmEpitotne naturalis scienli(B{Oxioid,
1664), are warned not to look for the causes of phenomena before there is a complete
agreement as to the facts. Four things are present during the calcination of the
metal in air : (1) The containing vessel or crucible ; (2) The metal being calcined ;
(3) The air ; and (4) The source of heat. Again the metal weighs more after the
calcination than it did before.
2. The hypotheses. — In applying the inductive method of investigation to these
facts, it is necessary to review every rational explanation consistent with the facts, and
to examine each hypothesis rigorously and impartially, since, as emphasized above, it
is necessary to show that the explanation finally selected is alon£ consistent with
the facts. This extension of the inductive process might be called the method of
exhaustion ; its importance was recognized by Epicurus (c. 300 b.c.).^ It is a
mistake to confine the attention to one hypothesis, because that might seriously
limit the range of the inquiry. The mind unconsciously assimilates evidence in
favour of a pet hypothesis ; and a pet hypothesis is apt to grow from a favoured
child to a tyrannical master. Four plausible hypotheses may be suggested to
explain the cause of the increase in weight : (1) the gases, etc., from the source
of heat unite with the containing vessel ; (2) the air unites with the containing
vessel ; (3) the gases from the flame penetrate the crucible, and unite with the
metal ; and (4) the air unites with the metal.
3. Testing the hypotheses by experiment.— By heating the crucible alone,
without the metal no change in weight occurs. This blank, dummy, or control
experiment shows that neither the first nor the second hypothesis will account for
the increase in weight of the metal. The third hypothesis can be tested by heating
the crucible and the metal out of contact with the air. There is then no change
in the weight of the metal. The third hypothesis is therefore untenable. This
method was not practicable for the early chemists, and hence J. Key employed
a less decisive test. It might be expected that if the results depend upon the
absorption of the flame gases, different residts must be obtained by using different
sources of heat — sun-glass, etc. — but the same results are obtained. in every case,
and accordingly, the third hypothesis is probably wrong.
4. The conclusion.— Key thus examined all the previously suggested explana-
tions, and rejected them one by one ; the remaining unchallenged factor was air.
The sole invariable antecedent of a phenomenon is probably its cause. Hence,
unless something has been overlooked, it is concluded that when metals are calcined
in air the increase in weight is due to the fixation of air by the metal, and not to the
absorption of furnace gases, nor to variations in the weight of the vessel in which
the calcination is made. The idea was not far from F. M. A. de Voltaire's mind lO
a century later, for in 1737 he said :
II est tr^s possible que 1' augmentation du poids soit venue de la mati^re r^pMidue
dans I'atmosphere, done dans toutes les autres operations par lesquelles les matieres
calcinees acquierent du poids cette augmentation pourrait aussi leur Hre venue ae la
meme cause, et non de la matiere ignee.
Similar remarks apply to R. A. Vogel's Experimerda chemicorum de in^emento
ponderis corporum quorundam igne calcinatorum examinat (Gottingen, 1753) made m
ignorance of J. Key's work. . . ,
J. Key attempted to explain how air alone could produce an increase m tne
weight of a metal during calcination. J. Key imagined that when air is heated, it
separates into a heavier and a lighter part, and that when a metal is calcined m air,
the lighter part of the air is distilled off, and the denser portion-/ air epats- alone
attaches itself to the metal and forms an ash or calx. J. Key did not prove th^
subsidiary hypothesis, viz. that only a part of the air attaches itself to the meta^
to form a calx. The increase in weight which occurs during calcination was com-
pared to the wetting of sand with water-most of the water can be drained away,
but a little remains adherent to the sand :
58 INORGANIC AND THEORETICAL CHEMISTRY
The condensed air becomes attached to the calx, and adheres, little by little, even to
the smallest of its particles. Thus the weight increases from the beginning to the end.
When all of it is saturated, it cannot take up more.
J. Rey's explanation proved to be fallacious. The great merit of J. Rey's work
lies in his demonstration that air is a ponderable fluid ; and the analogy between
air and a liquid regarded as ponderable fluids enabled him to grapple with an
intangible body, and to reason on that which from its subtlety had hitherto eluded
the grasp of the philosophers of all previous ages.ii
5. Confirmatory experiments. — S. Hales i^ and J. Juncker also explained the
increase in weight by assuming that particles of air were absorbed by the metal,
and S. Hales showed that when " 1922 grains of red lead is heated there arises
34 cubic inches of air." He did not consider it necessary to test the gas expelled
from the red lead since he assumed that it was elemental air. J. Rey's idea that the
increase in weight which occurs when a metal is calcined in air is due to the fixation
of air by the metal, was confirmed by the work of P. Bay en is early in 1774. Bay en
showed that mercurial calx owes its " calcined state " to its intimate combination
with an elastic fluid, the weight of which, in adding itself to that of mercury, " con-
stitutes the cause of the observed increase in weight " of the mercury during cal-
cination. The experiment was made by reversing J. Rey's procedure and heating
the calcined mercury until it decomposed into the original mercury and an elastic
fluid. The mercurial calx and the revived mercury were weighed before and after
the calcination :
Mercurial calx . .^. . . . . .576 grains
Revived mercury . . . . . . . . 518 „
Difference . . . . . . . . 58 „
P. Bayen added : "I cannot state positively that the 58 grains represent the true
weight of the elastic fluid, liberated from the 576 grains of mercurial calx, but
clearly everything leads to that conclusion."
J. Rey also made the interesting unforeseen observation that " nature, in her
inscrutable wisdom, has set limits which she does not overstep " ; in other words,
however long a metal may be heated in air, a definite weight o£ each metal can
combine with only a definite maximium amount of air. Students to-day regularly
repeat J. Rey's experiments on the metals, under various guises, as class exercises —
Table I. for example.
Table I. — -Action of Air on the Calcination of the Metals.
Metal.
Weight of metal
(gram).
Weight of calx
(gram).
Increase in weight
(gram).
Ratio weight air
absorbed : metal
used.
Magnesium
Zinc ....
Aluminium .
Copper
Tin ....
1
1-658
1-246
1-890
1-252
1-269
0-658
0-246
0-890
0-252
0-269
1-52
4-06
112
3-97
3-72
Hence, one gram of the
(Absorbed air). Magnesh
1 1-52
absorbed air is
im. Zinc.
4-0(i
respectively eq
Aluminium.
112
uivalent to
Copper.
3-97
Ti
3-72
grms.
6. Anticipation of new phenomena. — A good hypothesis ought to predict
phenomena which have not been observed, and to foretell the results of new
experiments ; because, if the hypothesis be true, it ought to include all other cases.
A hypothesis which is not illogical and which does not contradict known facts
is to be judged by its usefulness. The end justifies the means. G. J. Stoney has
expressed the idea neatly : "A theory is a supposition which we hope to be true ;
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 59
a hypothesis is a supposition which we expect to be useful. Fictions belong to the
realm of art ; when allowed to intrude elsewhere, they become either make-believes
or mistakes." When the consequences of a hypothesis are logically deduced, a
good hypothesis should not only explain, but it should anticipate new facts. Key's
hypothesis can be used to predict new results. In his Memoire sur la calcination
de retain dans les vaisseaux fermes, et sur la cause de Vaugrmntalion du poids
qu'acquiert ce metal pendant cette operation (1774), A. L. Lavoisier i* wrote :
Thus did I at the beginning reason with myself. ... If the increase in weight of a metal
calx (calcined in a closed vessel) be not due to the addition of fire matter, nor of any other
extraneous matter, but to the fixation of a portion of the air contained in the vessel, the
whole vessel after calcination must be heavier than before, and must merely be partly
void of air, and the increase in the weight of the vessel will not occur until after the air
required has entered.
A. L. Lavoisier confirmed this inference experimentally on November 12, 1774,
although the gifted Russian chemist, M. W. Lomanossofi,!^ had come to the same
conclusion in 1756, eighteen years before A. L. Lavoisier.
Rbfebences.
^ Paul de Canotanto, Theoria ultra estimationem peroptima ad coqnitionem totiua alhimia
veritatis. Manuscript No. 7159 at the Bibliotheque royalo, Paris — vide F. Hoefer, Histoire de la
chimie, Paris, 1. 444, 1842
2 P. Eck de Sultzbach, Theatrum chemicum, Argentorati, 4. 1007, 1622 ; G. F. Rod well,
Chem. News, 8. 113, 186, 246, 1863; 9. 14, 26, 50, 242, 1864; 10. 74, 195, 208, 1864; 11. 38,
74, 160, 291, 1865 ; 12. 62, 74, 293, 1865 ; 14. 51, 1866 ; 16. 29, 43, 1869.
' N. Lemery, Cours de chimie, Paris, 1675.
* J. Kunckel, Chymische Anmerkungen de principiis chymicis salihus, acidis,alcalibu8, Wittem-
berg, 1677 ; J. J. Becker, Physica subterranea, Franckfurt, 1690 ; J. Homberg, Mem. Acad.,
64, 1700.
5 O. Tachen, Hippocrates chemiciis, Venice, 210, 1666.
^ H. Boerhaave, Elementa chemice, Lugduni Batavorum, 1732.
' N. le Febvre, Traicte de la chymie, Paris, 1660 : J. Mayow, De sal-nitre et spiritu nitro-aereo,
Oxford, 1669.
8 Alembic Club Reprints, 11, 1895 ; R. P. Beraud, Dissertation sur la cause de VaugmerUation de
poids que certaines matieres acquierent dans leur calcination, Haye, 1748.
* E. Zeller, The Stoics, Epicureans, and Sceptics, London, 424, 1870.
10 F. M. A. de Voltaire, Mem. Acad., 169, 1737.
11 G. F. Rodwell, Chem. News, 10. 208, 1864.
12 S. Hales, Vegetable Staticks, London, 1. 288, 1727 ; J. Juncker, Conapectus chemtca themettco-
practicce, Halle, 1749.
i» P. Bayen, Journ. Phys., 3. 135, 281, 1774.
" A. L. Lavoisier, (Euvres, Paris, 2. 103, 1862.
i« A. Smith, Journ. Amer. Chem. Soc, 34. 109, 1912 ; Ostwald's Klassiker, 178, 1910.
§ 14. The Evolution of Ideas regarding the Nature of Burning
step by step we cross great eras in the development of thought ; there is no sudden
gigantic stride ; a theory proceeds by slow evolution until it dominates or is destroyed.
— G. F. RoDWELL (1869). ,.^ ^ ., , .
Slowly, gradually and laboriously one thought is transformed into a different tnougni,
as in all likelihood one animal species is gradually transformed into a new species. J"^^
ideas arise simultaneously. They fight a battle for existence not otherwise than diet tne
Ichthyosaurus, the Brahmin and the horse. Thoughts need their own time to ripen, grow,
and develop. — E. Mach.
The beautiful fiction of Greek mythology, as related by ^Eschylus, teUs how
Prometheus stole fire from heaven, and gave the sacred gift to man as the most
useful of all his necessaries. To many ancient worshippers, fire was a thmg divme,
the supreme manifestation of God himself, and it soon became the one visible
symbol of God. Even to-day the sacred fire exists among the races of the iiaikans.
Accordingly, the Zoroastrian fire worshippers called their god the one fire, or the pure
fire ; i and the sun was worshipped first as an emblem of the deity— Hre— ana
afterwards as itself a god.2 Fire thus came to be the first and most potent oi an
60 INORGANIC AND THEORETICAL CHEMISTRY
the elements, and it is easy to understand how Heracleitos regarded subtle fire as
the sole primal element from which all things were created ; and how fire was
canonized by Pythagoras and Empedocles as one of the four indispensable and all-
sufficient components of the universe.
Some of the early philosophers promulgated a dynamical theory of heat and fire.
Epicurus (c. 300 B.C.) regarded heat as a result of the rapid motion of minute spherical
particles which insinuated themselves in the pores of the densest substances ; cold
was likewise produced by angular particles moving more slowly. Lucretius (c. 80
B.C.) similarly referred heat to the motion of primary particles which penetrated
every material thing. H. Cardan ^ (1557) spoke of a ynotus ignis and a motus
caloris. R. Fludd (1617), F. Bacon (1620), A. Kircher (1644), and others have
propounded views which amount to a denial of the elemental nature of fire, since
they virtually assumed that heat is a violent motion of the particles of bodies, or
that fire is air which has been made to glow by the vehement collision of its particles,
and that the heat so generated changes combustible matter into flame.
Rene Descartes, in his Principia fhiloso'phice (Amsterdam, 1644), assumed that
originally all matter consisted of square particles endowed with two kinds of
motion : a rotation of each particle about its own centre ; and a rotation of groups
of particles about a common centre. The angles of the particles were abraded by
collisions producing three kinds of particles which he called elements : (1) Materia
primi elementi, or fine dust, which he also called materia subtilis, or materia coslestis,
because the sun, stars, and fire were supposed to be composed of this material.
(2) Globuli secundi elementi, or rounded particles which were supposed to make up
the atmosphere and everything between the stars and the earth. (3) Particulce
tertii elementij or particles which retain some of their angles and are partially
rounded ; these were assumed to make up the earth and all terrestrial bodies. The
particles of the materia coelestis were supposed to be in far more rapid motion than
the other particles. The different forms of matter were supposed to be determined
by the relative proportions and motions of these three elements ; and every natural
phenomenon, the result of the conduction of motion from one body to another.
Fire, according to R. Descartes, consisted of the third element rapidly agitated by
the tnateria coelestis ; and the particles of combustible bodies were supposed to be
peculiarly adapted to receive the motions of the materia coelestis. It was all a
transmission of motion, not substance. N. Lemery adopted the main tenets of th^
Cartesian theory in his famous Cours de chimie (Paris, 1675) :
I understand by igneous corpuscles — corpuscles ignees — a subtle form of matter which
having been thrown into rapid motion, still retains the capacity of impetuous motion
when it is enclosed in grosser matters ; and when it finds bodies which by their texture or
figure are easily put in motion, it draws them about so strongly that their parts develop
heat by being rubbed violently against one another. . . . The particles of sulphur, for
instance, are very susceptible to motion . . . and it seems probable that fire is only violent
motion of minute bodies about their common centre.
Flame, said R. Descartes, is directed upwards because it contains much materia
coelestis which is lighter than air, and the cause of lightness in bodies generally.
Descartes' materia coelestis approximates to the modern conception of an aether
more subtle than air, and filling the interstices between the molecules of air with a
continuous series of globules which pervade the pores of glass, and of the densest
substances without interruption ; and propagating light by communicating impulses
from one molecule to another so as to produce a kind of pressure without locomotion.
Isaac Newton * postulated a similar aether " pervading and lurking in dense
bodies, but not yet sufficiently manifested by experiments." R. Hooke introduced
the notion of vibratory impulses in this medium, and the idea was elaborated by
C. Huygens and T. Young into the undulatory theory of light which is now generally
accepted. The communication of the vortex motion of the materia coelestis to the
atoms is thus described by R. Boyle : ^
The restless agitation of the materia coelestis wherein the particles of air swim, so whirls
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 61
them round that each corpuscle endeavours to beat off all others from coming within the
little sphere requisite to its motion about its own centre . . . their elastic power is made
to depend upon the vehement agitation which they receive from the fluid sether Imateria
coelestis) which swiftly flows between them.
Several early observers noticed that fire cannot subsist without air. Theo-
phrastus,6 for instance, in the fourth century B.C., in his treatise On Fire, noticed
that air plays an important part in the maintenance of flame ; Hero of Alexandria
(c. 117 B.C.) demonstrated this by placing a lighted lamp in a closed vessel, and
showing that under these conditions the flame was extinguished— Hero said that the
fire consumed and rarefied the air ; and from a similar experiment in the thirteenth
century, Roger Bacon inferred that aer est cihus ignis—aii is food of flame— in agree-
ment with Theophrastus— 315 B.C.— who said, " It is not at all irrational to believe
that flame is maintained or supported by an aeriform bo4y." Near the beginning
of the sixteenth century (c. 1500), Leonardo da Vinci 7 clearly recognized that air
is necessary for the sustenance of the flame of a burning candle, for he said : " There
is smoke in the centre of the flame of a wax candle because the air which enters
into the composition of the flame cannot penetrate to the middle. It stops at the
surface of the flame and condenses there." Leonardo da Vinci also showed that
air is necessary for respiration ; and that air is not an element because one part
of it alone is concerned in combustion. R. Fludd 8 noticed in 1617 that when a
candle is burnt in a glass vessel over water, the water rises in the vessel as the air
is consumed, for " air nourishes fire, and in nourishing consumes it." H. Cardan
also, in his De rerum varietate (Basil, 1557), classified different substances as corti-
hustihle or incomhustiUe. Flame, said he, is nourished by a ga,s— flatus — which wiU
ignite a glowing splint, and which exists in saltpetre. H. Cardan was here ver}'
near to the discovery of facts which in the hands of A. L. Lavoisier produced une
revolution iintnense dans la science.
After his discovery of the air-pump in 1650, one of the first experiments tried
by 0. von Guericke ^ was to ascertain if a candle would continue burning in an
exhausted receiver, and it was found that owing to the want of air the flame of a
lighted candle expired more quickly under the exhausted receiver of an air-pump
than when the receiver was not exhausted ; fire, said Guericke, consumes air. In
his first treatise on pneumatics, New experiments, fhysico-mechanical, touching the
spring of air (London, 1660), R. Boyle mentions several proofs that combustion
cannot proceed in a space void of air ; and in 1672, R. Boyle, in an essay On the
difficulty of pr^eserving flatne without air (London, 1672), showed that when placed
under the receiver of an air-pump, the flame of burning gas, derived from the action
of an acid on iron, is suddenly enlarged on exhausting the air, and finally is ex-
tinguished ; and he showed that sulphur does not burn if heated in vacuo. These,
and other experiments on similar lines, clearly showed that air is necessary for
combustion.
Robert Hooke outlined a theory of combustion in his Micrographia (London,
1665). He noticed the similarity in the actions produced by air and by saltpetre,io
and hence suggested that air is mixed with a substance which is like, if not identical
with, that which is fixed in saltpetre, and that only this portion of air is required
to support combustion and respiration. A similar conclusion had been hinted at
by R. Fludd,ii who said : " The substance of saltpetre is nothing but air congealed
by cold." Again, in his Lectiones cutleriance (London, 1674-9), Robert Hooke
assumed that burning is produced by the solvent action of the surrounding air
which is dissolved by the burning body much as water dissolves salt. He said :
Air is a menstruum that dissolves all sulphurous bodies by burning, and without air,
no such dissolution will follow, though the heat applied be never so great which was
particularlv tried by charcoal enclosed in an iron case with a screw stopper, which though
violently heated yet the coke was not burned nor wasted when taken out. . . . Ihat
shining transient body we call flame is but a mixture of air and volatile sulphurous parts
of combustible bodies which are acting upon each other as they ascend. . . . The action
is performed with so great violence and does so minutely act, and rapidly agitate the
62 INOKGANIC AND THEORETICAL CHEMISTRY
smallest parts of the combustible matter, that it produces in the diaphanous medium of
air the action or pulse of light.
J. Mayow (1669)12 subjected the guess or hypothesis of Hooke to the test of
observation. The following experiment is a more refined form of one made by
J. B. van Helmont, circa 1640 : —
J. Mayow arranged a candle in water so that the wick was between 9 and 10 cm. above
the surface of the water. A glass cylinder, A, Fig. 6, was lowered over the burning candle
so that the level of the water inside and outside the cylinder was
the same. A small syphon, B, was used for the purpose. Im-
mediately the cylinder was in position, the syphon was removed.
The flame of the candle soon expired, and water rose in the
jar. Some gas still remained in the jar, but it could not be air
because one of the characteristic properties of air is to support
the burning of the candle, and the flame of the candle is ex-
tinguished in the residual gas. Mayow obtained analogous
results by confining a mouse under the jar. The mouse died,
and the water rose in the jar.
Hence, Mayow inferred that air contains two kinds of
F 6— Ma ow's Experi Pa^*ticles, One of which — the nitro-aerial particles —
ment on Combustion " ^ withdrawn and destroyed by the burning candle.
J. Mayow also stated :
Though the particles of air are very minute, and are vulgarly taken for an element of
the greatest simplicity, it appears to me necessary to judge them to be a compound. . . .
It is manifest that the air is deprived of its force by the respiration of animals much in the
same manner as by the deflagration of flame.
Mayow does not seem to have quite grasped the idea that the nitro-aerial particles
which support combustion actually combine with the burning body, although he
correctly inferred that air was a mixture containing nitro-aerial particles as one
constituent. The nitro-aerial particles were indiscriminately called fire-air, nitre-air,
and nitro-aerial spirit. Mayow' s observations appear to show that air is a mixture
of two gases one of which is withdrawn during combustion, and the remaining gas
does not support combustion. Stephen Hales i^ also noticed that in the combustion
of phosphorus under a bell- jar, white fumes are produced and air is absorbed. When
the experimenters of the seventeenth century spoke of the destruction of the
elasticity of a portion of the air, they meant that some of the air was lost —
presumably by absorption by the confining liquid, etc.
Some modern commentators consider that J. Mayow's nitro-aerial spirit repre-
sented oxygen, and his aerial spirit, nitrogen. It has been said that J. Mayow's
nitro-aerial particles were made to explain too much, for he applied them to all
sorts of phenomena — e.g. the formation of acids, fermentation, the production of
nitre, calcination, combustion, and respiration— rather is this a tribute to J. Mayow's
genius. J. Mayow considered the nitro-aerial particles to be fixed as the acid
component of nitre because the effects produced by nitric acid and by the burning
glass on antimony were the same. He extended his views to other substances —
particularly the acidification of sulphurous and fermenting substances by exposure
to the atmosphere — and thus inferred that his nitro-aerial particles are the active
agents in combustion and acidification. When J. Mayow regarded these same
particles as the principle by which metals increase in weight when calcined in air ;
the principle by which vegetables germinate and. grow ; and by which the blood
changes its colour in the lungs during respiration, he seems to have generalized
with far greater precision from a few facts than the greater part of the next
generation did from many.i^
J. Mayow, however, did mix some fantastic hypotheses with his eminently logical
interpretations of ingenious experiments, and in some cases the relevant matter is
mixed with so many irrelevancies, that it is difficult to tell which is which unless
his statements are interpreted in the light of what is now known to be true. To-day,
J. Mayow's brilliant reasoning would be accepted as a logically conclusive proof of
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 63
the existence of oxygen as a distinct substance ; but his demonstration was a
century ahead of its time. Instead of his unique experimental talents
being encouraged by his contemporaries, they were damped by the coldest of
receptions. His work was evidently above the heads of his contemporaries The
historian of science, said G. F. Rodwell,i5 should endeavour to grasp the precise
mode of thought of the man of whom he writes, to think as he thought, to view
the phenomena in the light of the age in which he lived, and then to reason on them
as he reasoned. Evidently, then, E. Hooke and J. Mayow got very near to the
present-day theory of combustion, but unfortunately, the latter's ingenious experi-
ments had very little, if any, influence on the subsequent development of chemistry,
because the lowering clouds of the phlogiston hypothesis appeared as a grey after
dawn and gradually darkened the sky of chemistry until the chemical world appeared
to be enveloped in an impenetrable fog. For another century more trust was placed
in phantasms of the imagination than in facts obtained by precise observations.
It must be added that in the Far East, the Chinese philosopher Mao-Khoa, who flourished
about the eighth century, is said to have had a fairly clear idea of the composition of air,
and of the part played by oxygen— which he called yin—m. combustion and respiration!
This historical information, however, played no part in European discoveries since it is
but a comparatively short time ago that Mao-Khoa's views were reported, and im-
familiarity with the language and literature has prevented many examining the claims of
_^ the Chinese scholar to a proud place in the history of chemistry.
* References.
1 T. Stanley, History of ChaUaick Philosophy, London, 1662 ; V. Titelbach, Open CouH, 15.
143, 1901.
2 Malachi, 4. 2 ; / Chronicles, 21. 26 ; // Chronicles, 7. I ; / Kings, 18. 38 ; Exodus, 3. 38 ;
19. 18 ; Deuteronomy, 4. 12.
' H. Cardan, De rerum varietate, Basil, 1557 ; R. Mudd, Utriusque cosmi majoris scilicet et
minoris metaphysica, physica atque technica historia, Oppenheim, 1617 ; F. Bacon, Novum organum,
London, 1620 ; A. Kircher, Ars m/igna lucis et umbrcs, Rome, 1644 ; G. F. Rodwell, PhU. Mag.,
(4), 35. 1, 1868.
* Registry Book of the Roy. Soc., 5. 67, 1675-9 ; Letter from Newton to Halley, 1686 ; Letter
from Newton to Boyle, 1678 ; Isaac Newton, Opticks, London, 1717.
5 R. Boyle, New Experiments, Physico-mechanical, touching the Spring of Air, London, 1660.
* Theophrastus, Uepl irvpSs, Paris, 1567.
' J, B. Venturi, Notice de quelques articles appartenant a Vhisloire naiurellede la chimie, iiris de
Vessai stir les ouvrages de Leonard de Vinci, Paris, 1797 ; M. Libri, Histoire des sciences mathe-
^matiques en Italic, Paris, 3. 27, 1838-41 ; E. O. Lippmann, Leonardi da Vinci als Gelehrter und
Techniker, Stuttgart, 1900 ; E. Muntz, Leonardo di Vinci, London, 1898.
^ R. Fludd, Utriusque cosmi majoris scilicet et minoris metaphysica, physica atque technica
historia, Oppenheim, 1617.
' 0. von Guericke, Experimenta Magdehurgica, Amsterdam, 1672; G. Berthold, Wied Ann.,
54, 724, 1895.
i» R. Bathurst and N. Henshaw, Aerochalinos, or a Register for the Air, London, 1677.
^^ R. Fludd, Utriusque cosmi majoris scilicet et minoris metaphysica, physica atque technica
historia, Oppenheim, 1617.
12 J Mayow, De sal-nitro et spiritu nitro-cereo, Oxford, 1669 ; Tractatus quinque medico-physici,
Oxford, 1674 ; Alembic Club Reprints, 16, 1907 ; J. B. van Helmont, Orius medicince, Lugduni
Batavorum, 84, 1656.
13 S. Hales, Vegetable Staticks, London, 1727.
14 W. V. Harcourt, Phil. Mag., (3), 28. 478, 1846 ; J. B. Cohen, CAcw. WorU, 3. 247, 1914 ;
A. Smith, Journ. Amer. Chem. Soc., 34. 109, 1912 ; G. D. Yeates, Observations on the Claims of
Moderns to some Discoveries in Chemistry and Physiology, London, 1798.
15 G. E. Rodwell, Chem. News, 14, 25, 1866.
§ 15. The Phlogiston Theory
During the greater part of the eighteenth century, the doctrine of phlogiston was not
only the lamp and guide of chemists but it remained the time-honoured and highest
generalization of physical chemistry for over half a century. — S. P. Langley.
Phlogiston died as an old king,— once infinitely dominant, somewhat tyrannical, not
always just ; now deposed, decrepit, utterly senile, forsaken by all.— W. Odling.
Up to about the middle of the fourteenth century, combustion was explained by the
64 INORGANIC AND THEORETICAL CHEMISTRY
aid of the assumption that all combustible bodies contained a common element,
the essence of fire, that is, an inflammable principle which enabled them to burn.
This obviously means little more than saying that substances burn because they
are combustible. The idea of a subtle fire innate in matter has pervaded philosophy
from the earliest times. Zeno (c. 450 B.C.) called it drcKveKov -n-vp — barren fire ;
Heracleitos (c. 450 B.C.), dvaOvfiiai^ ; Lucretius (c. 80 ac), suUilis ignis, coelestis
ignis, or tenuis ignis ; Paracelsus (c. 1500), sideric sulphur ; H. Cardan (c. 1553),
color coelestis ; and R. Descartes (c. 1664), materia coelestis. The alchemists of the
Middle Ages variously styled it elemental fire, astral fire, sulphurous principle, or
materia ignis.
The empyrean i element of the ancient Greeks was consecrated under the
classical name phlogiston by the hierophants of a newer chemistry. The word
phlogiston is derived from the Greek <f>\oyL^oi, to inflame, and is related to ^Aeyw,
to burn, and <^Ao^, flame. In some cases phlogiston was believed to resemble
that subtle fiction we now call cether, and J. Juncker,^ in 1744, called it materia
igtiea OBtherce. J. Kunckel (1676) thought that the inflammable principle must be
sulphur, and wrote ubi ignis el color, ihi sulphur — where there is fire and heat there
is sulphur. Virtually all chemists of this period attributed the combustibility of
a substance to the presence of sulphur. There were many sulphurs — e.g. the sulphur
of wood (carbon), the sulphur of wine (alcohol), etc., and Robert Boyle in his essay
On the difficulty of preserving fiame without air (London, 1672), called the fume or
gas which is evolved when an acid acts upon iron the volatile sulphur of Mars ; and
in his essay On the producihleness of chemical principles (London, 1680), he speaks
of the sulphur of the chemist as being a combustible and inflammable principle.
Twenty-five years after the appearance of R. Descartes' Principia, and about
the time of J. Mayow, J. J. Becher began to publish the chemical side of a theory
analogous in many respects with the physical theory of Rene Descartes. The most
important work of J. J. Becher is his Physica suhterraneo (Lipsise, 1669), and the
three supplements dated 1671, 1675, and 1680 respectively — J. J. Becher's term
suhterraneo is probably equivalent to the modern inorganic. J. J. Becher advocated
the importance of experiment in chemical science. He rejected the four-elements
and the quintessence of the ancients, but he did so only to promulgate four elements
of his own devising — fire ; the earthy principle ; the combustible element ; and
the metallic one. This enabled him to classify material substances into fiery or
imponderable bodies, earth, combustibles, and metals. The combustibles and metals
were later grouped together, and his system was simplified into fire, the first kind
of substance ; earths, calces, and acids, the second ; and combustibles and the
metals, the third ; otherwise expressed, J. J. Becher's triad included fire, the
products of combustion, and combustibles. It was not the custom, in J. J. Becher's
time, to keep one specific technical term for one specific thing. He seems to have
used the terms vitrifiable earth — terra lapida or terra vitrescihilis — inflammable
earth — terra pinguis — and mercurial earth — terra fluida or terra mercurialis — almost
in the same sense that the alchemists spoke respectively of salt, sulphur, and
mercury. He regarded his three elements as three varieties of sulphur ; vitrifiable
earth was called fixed sulphur, tnercuriol earth, or volatile sulphur ; and infiam-
mahle earth was indiscriminately called combustible sulphur, sulphur adustible,
sulphur ardens, or phlogistic sulphur. J. J. Becher said :
Combustible sulphur is the innate heat of the metals. . . . The base metals contain an
inflammable principle which by the action of fire goes into the air, leaving behind a metal
- calx.
This recalls the hypothesis promulgated by H. Cardan a century earlier. J. J.
Becher supposed the different forms of matter to be compounds of one or more of
these elements differently arranged with or without water. G. F. Rodwell,^ in a
valuable article On the theory of phlogiston, says that J. J. Becher never used the
word ^Xoyia-rov as a noun to designate the matter or principle of fire. It was
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 65
reserved for J. J. Becher's disciple G. E. Stahl, near the beginning of the eighteenth
century, to employ the term phlogiston for the materia ignis of the early writers.
Towards the end of the seventeenth century, G. E. Stahl sketched an outline
of the theory of phlogiston in his Zymotechnia fundamentalis (Franckfurth, 1697) ;
and in his Specimene Becheriano (Franckfurth, 1702), he elaborated J. J. Becher's
Physica subterranea, a work which he rated very highly, and vaunted it to be opus
sine pari — a work without a peer ; primum ac princeps — first and foremost ; liber
undique et undique primus — a book everywhere supreme ; etc. G. E. Stahl also
wrote his Fundamenta chymice dogfnaticce et experimentalis (Norimberga?, 1723) as
a text-book of phlogistic chemistry, and he described in his Experimenta, observa-
tiones, animadversiones, CCG numera, chymicce et physicw (Berlin, 1731), a nimiber
of experiments in support of the theory, and answers to some questions. Like the
ancients, G. E. Stahl believed in the existence of two kinds of fire : (i) ordinary
visible fire, or mundane fire, or gross earthy fire which he called ignis or flame ;
and (ii) pure, subtle, invisible fire, materia ignis, or phlogiston, which became ignis
only when associated with material particles which assimilated its motion. It
therefore follows that J. J. Becher's terra infiammibilis, terra pinguis, combustible
sulphur, sulphur ardens, or phlogistic sulphur ; and G. E. Stahl's phlogiston are
new names for an old time-honoured principle. The dominant functions of Des-
cartes' materia coelestis were conferred on phlogiston, and some new properties were
added. Phlogiston, said G. E. Stahl, is the materia aut principium ignis, non ipse
ignis. Although the real nature and properties of phlogiston were unknown, its
existence was pure conjecture, yet G. E. Stahl did not hesitate to speak very
definitely about this creature of the imagination. He said :
Phlogiston is a very subtle matter capable of penetrating the densest substances ; it
neither bums, nor glows, nor is visible ; it is agitated by a rapid motion- — igneo motu — and
it is capable of communicating its motion to material particles adapted to receive it. The
particles when endowed with this rapid motion constitute visible fire. . . . Fire is an
aggregate of a great number of particles in vehement motion. The maieria of fire is phlogis-
ton— a thin all-pervading medium composed of movable particles — the forina is the motion
itself ; the materia is passive, the forma is active. The motion of phlogiston is gyratoriua
seu verticillaris and not progressive. . . . Heat is an intestine motion of the particles of
matter.
G. E. Stahl taught that in the act of combustion, phlogiston, an intrinsic con-
stituent of every combustible body, was set at liberty. Oxidation was said to be
due to the escape of phlogiston ; deoxidation or reduction to the absorption of
phlogiston. When a metallic oxide was heated with a substance rich in phlogiston
— e.g. charcoal or reducing agents generally — the charcoal supplied the calx
or metallic oxide with phlogiston, and reproduced a compound of phlogiston with
the metallic oxide which was the metal itself. Metals were thus supposed to be
compounds of phlogiston with their calces or oxides. The noble metals were sup-
posed to have their phlogiston so firmly fixed that nothing can take it from them.
While the base metals are turned to calces when roasted in air, the royal metals
remain intact during the fiercest trial. If phlogiston escaped, the metalUc oxide or
calx remained. The idea is symbolized
Metal ^ I-hlogiston 4- Metal calx or oxide
The body from which phlogiston escapes, when no longer capable of supporting
combustion, was said to be dejMoqisticated, and conversely, the body- solid, liquid, or
gas -with which the phlogiston' was combined, or by which it was absorbed, was
said to be phlogisticafed. Apparently overlooking the theories of R. Hooke (1664)
and J. Mayow (1674), which were developed while Stahl was in the nursery,
M. E. Chevreul * claimed that on doit a Stahl la premiere explication de la combustion.
The phlogistians are said to have been most assiduous in collectmg instanttw
convenientes, but very reluctant in accepting instanti(B inconstanti(E. G. E. Stahl, by
denying that the calx of mercury weighed more than the mercury from which it was
VOL. I. *"
66 INORGANIC AND THEORETICAL CHEMISTRY
derived, sacrificed fact to theory. Phlogistic chemistry was thus established in
opposition to facts which at first sight appeared to carry its own refutation, for if
the calcination of a metal be attended by the expulsion of phlogiston, the calx
should weigh less than the metal. When the facts that the loss of phlogiston is
always associated with a gain in weight, and vice versa, became too insistent, and
could no longer be denied, G. E. Stahl, in his Fundamenta chymice (Norimbergse,
1723), frankly evaded the difficulty by introducing another perplexity. He said :
The fact that metals when transformed into their calces increase in weight, does not
disprove the phlogiston theory, but, on the contrary, confirms it, because phlogiston is
lighter than air, and, in combining with substances, strives to lift them, and so decreases
their weight ; consequently, a substance which has lost phlogiston must be heavier than
before.
Thus, the phlogistians said that phlogiston also embodied the principle of levity,
and conferred a negative weight upon bodies. Consequently, when phlogiston is
associated with matter, the weight is lessened, just as inflated bladders lessen the
water-weight of a swimmer.
It may not seem rational to postulate the existence of a substance weighing
less than nothing. It will be observed, however, that the assertion, all 7natter
is heavy and possesses weight, is one way of saying that the attraction of gravitation
exists between all masses of matter. This is by no means a self-evident principle,
because it is just as easy to conceive of two masses of matter repelling one another,
and easier still to imagine two masses of matter neither attracting nor repelling one
another. Thus, G. B. Airy ^ said : "I can easily conceive that there are plenty of
bodies about us not subject to this mutual action, and therefore not subject to the
law of gravitation." Hence, the assumption of a phlogiston weighing less than
nothing is not so silly as is sometimes supposed. If phlogiston be a principle of
levity, however, with a negative gravity, it would not be attracted but rather repelled
by other substances. Consequently, in order to explain Ifow phlogiston becomes
fixed in combustible bodies, it would be necessary to invent another force stronger
than gravitation. It is quite true that no form of matter with a negative gravity
has been detected, and accordingly, it is assumed that a form of matter weighing
less than nothing does not exist, and that, other things being equal, an increment
in weight is necessarily an effect of an increment of matter.
The era of phlogiston presents serious claims to be regarded as the period when
chemistry began to take shape as a definite science. It represented a definite attempt
to group diverse chemical phenomena about a rational principle which seemed
adequate to embrace the then known facts. The doctrine of phlogiston was
invented to render chemical phenomena intelligible to the mind ; it was founded
on fact ; and it owed its value in the minds of a race of eminently practical chemists,
to the facts which it represented. New facts soon began to accumulate which
could not be explained in terms of the original simple hypothesis, and auxiliary
hypotheses were framed in quick succession ; these made the theory contradictory
and unmanageable. In his Reflexions sur le phlogistique (Paris, 1783), A. L.
Lavoisier ^ said :
Chemists have turned phlogiston into a vague principle, one not rigorously defined,
and which consequently adapts itself to all the explanations for which it might be required.
Sometimes this principle has weight, sometimes not ; sometimes it is free fire, sometimes
it is fire combined with the earthy element ; sometimes it passes through the pores of
vessels, sometimes the vessels are impervious to it ; it explains both causticity and non-
causticity, transparency and opacity ; colours and their absence ; it is a veritable Protean,
changing in form each instant.
A. L. Lavoisier's explanation of the increase in weight which occurs when lead
is calcined, seems so obvious that it is now difficult to appreciate the difficulty as
set forth by P. J. Macquer (1769) 7 :
The phenomenon is un vrai paradoxe chimique. While it is easy to prove the fact, it
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 67
is difficult to find a satisfactory explanation. The phenomenon is outside all physical ideaa
which we have formed, and it is only in the future that a solution of the difficulty can be
expected.
The chief difficulties encountered in the application of the theory of phlogiston
were : (i) The increase in weight which occurs when metals are calcined ; (ii) The
necessity for the presence of air during combustion ; (iii) The change of mercury
calx into a metal without the addition of phlogiston. In the latter case, P. j.
Macquer, indeed, used the fact that mercury calx can be converted into a metal
by merely heating it jper se in the absence of a body containing phlogiston, to argue
that the mercury calx is not a real calx, but merely a substance which has acquired
par r action dufeu Fapparence d'une chaux metallique. P. J. Macquer endeavoured
to remove the objection by assuming that phlogiston is light and that during com-
bustion, light and air mutually precipitate one another ; during the calcination
of a metal, the air unites with the metal and disengages phlogiston ; and during
the reduction of a metal calx, light unites with the metal and liberates air. C. W.
Scheele ® supposed that heat, light, and inflammable air were compounds of air
and phlogiston which are convertible into one another by the addition or subtraction
of phlogiston — -inflammable air was assumed to contain most, and heat least
phlogiston. During calcination, the metal either attracted air by means of its
phlogiston and thus formed heat, or else communicated phlogiston to the air, and
attracted heat from the fire ; in either case the air remained in the calx and im-
parted an overplus of weight. When a calx is reduced by inflammable air, heat,
or light, the latter is decomposed and the phlogiston remains united to the reduced
metal. The fact that oxygen supported combustion better than air led to the
hypothesis that air contains more phlogiston than oxygen, which was hence called
dephlogisticated air. At one time H. Cavendish (1766) ^ assumed that inflammable
air is itself the phlogiston of the ancient chemists, and that a certain amount is fixed
in all combustible bodies. Inflammable air, i.e. hydrogen gas, was accordingly
called phlogisticated air. This hypothesis substituted a definite tangible material
for a vague principle, but many of the properties of G. E. StahFs phlogiston were
utterly at variance with those of hydrogen, and the hydrogen hypothesis completely
failed.
About 1770, it had been definitely proved that there is an increase in weight
during the conversion of a metal into a calx by calcination of the metal in air. The
fact was qualitatively explained, somewhat clumsily, by the phlogiston hypothesis
which was based upon the subtilis ignis of the ancients, or the materia coekstis of
R. Descartes. R. Hooke, J. Rey, and J. Mayow had recognized that air somehow
plays an important part in the process of calcination and combustion, but while
their ideas on the general principle were clear, the details were somewhat hazy and
indefinite.
References.
1 Empyrean — the highest heaven where the ancients supposed pure fire subsisted.
2 J. Juncker, Conspectus chemice theoretico-practicce, Magdeburg, 1744 ; J. von Lowenstern
Kunckcl. Nfdzliche Ohservationes, Hamburg, 1676.
3 G. F. Rodwell, Phil. Mag., (4), 35. I, 1868.
* M. E. Chevreul, Campt. Rend., 59. 977, 1864.
5 G. B. Airy, Gravitation, London, 1885; R. Hare, Jmer. Journ. Sctence, (1), 42, 200, 184-i;
W. Whewell, Cambridge Phil. Soc, 7. 197, 1842.
« J. B. Dumas, Lecons sur la philosophic chimique, Paris, 161, 1837.
' P. J. Macquer, Mem. Acad., 153, 1769 ; Mments de chimie pratique. Pans, 1 /5I.
8 C. W. Scheele, Chemische Abhandlungen von der Luft umf. dem Feuer, Ixsipzig, 1782.
« H. Cavendish, Phil. Trans., 56. 141, 1766; R. Kirwan, An Essay on Phl^tston and the
Constitution of Acids, London, 1789 ; Phil. Tram., 72. 236, 1782 ; W.^Nicholson, The Controversy
between Kirwan and the French Academicians on Phlogiston, London, 1 /87.
68
INORGANIC AND THEORETICAL CHEMISTRY
§ 16. Lavoisier's Experiments on Combustion and Calcination
Nature is ever making signs to us, she is ever whispering to us the beginnings of her
secrets ; the scientific man must be ever on the watch, ready at once to lay hold of nature's
hint, however small ; to listen to her whisper, however low.— M. Foster.
The beginning and end of every exact chemical process is weighing. — W. Nicholson (1808) .
In 1772, Antoine Laurent Lavoisier began to publish accounts of a brilliant
series of investigations which in a few short years banished phlogiston completely
from chemical science. Chemistry had grown too great to be governed by the
mystic phantom — phlogiston. In his Opuscules physiques et chimiques (Paris,
1774), A. L. Lavoisier first showed that phosphorus and sulphur increase in weight
and absorb large volumes of air when they are burnt, and he obtained similar results
with lead and mercury in closed vessels. A. L. Lavoisier pursued the subject
further in a Memoire sur la calcination de F Stain dans les vaisseaux fermcs, et sur la
cause de F augmentation de poids qu'acquiert ce metal pendant cette operation (1774).
He found that the vessel containing the air and tin did not increase in weight,
although part of the air was absorbed. When the flask was opened, air rushed in,
and the increase in the weight of the vessel was found to be equal to the increase in
weight which the tin alone had suffered. Hence, A. L. Lavoisier concluded, with
J. Rey, that the increase in the weight of the tin was solely due to an absorption
of the air in which the calcination had occurred. There was not sufficient air in
the flask to saturate all the tin, and yet some air always remained as a residue.
Hence, A. L. Lavoisier concluded further that only part of the air can combine
with the metal during the calcination ; he also found that the increase in the
weight of the tin during calcination is equal to the decrease in the weight of the air.
Hence, it seems as if air contains at least two constituents, only one of which is
absorbed by the heated metal. This inference was tested by an important ex-
periment described in his Traite elementaire de chimie (Paris, 1789).
The mercury was confined in a glass retort with an S-shaped neck which dipped under
a bell-jar in a trough of mercury, as illustrated in Fig. 7. The air in the retort was in
commimication with the air in the bell- jar. The level of
the mercury in the bell-jar was adjusted at a convenient
level, and its position " very carefully marked with a strip
of gummed paper." By means of a charcoal furnace, the
mercury in the retort was heated- — not quite to its boiling
point (357°). A. L. Lavoisier said : " Nothing of note
occurred during the first day. The second day I saw little
red particles swimming over the surface of the mercury, and
these increased in number and volume during four or five
days ; they then stopped increasing and remained in the
same condition. At the expiration of twelve days, seeing
that the calcination of the mercury made no further pro-
gress, I put the fire out." The red particles were identified
with the calx of mercury now called red oxide of mercury,
or mercuric oxide, and then called mercurius calcinatus
per se.
After making allowance for variations of tempera-
ture and pressure, A. L. Lavoisier found that when
mercury was calcined with a given volume of air in a
closed vessel, 50 cubic inches of air were reduced to
, . , between 42 and 43 cubic inches ; the difference, 7 to 8
Fig. 7— a. L. Lavoisier s ^^^-^ inches, that is, one-fifth or one-sixth of the total
Experiment on the Com bus- , f ,i • i i i i ,i ,
tion of Air. volume 01 the air, was absorbed by the mercury, formmg
the red calx of mercury. The air which' remained
in the retort was not absorbed by the excess of hot mercury ; it was rather
less dense than ordinary air ; it extinguished the flame of a burning candle
immersed in the gas ; and a mouse was quickly suffocated when placed
in the gas. Hence, A. L. Lavoisier first called the gas moufette atmospherique, and
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 69
later azote, " from the a privative of the Greeks, and ^a>^, life." In France, the
gas is still called azote, though in Britain it is called nitrogen.
By collecting the red mercury calx, and re-heating it in a suitable retort
(probably to 400°), Lavoisier obtained between seven and eight cubic inches of a gas
which had been previously removed from the air by the hot mercury. The gas was
exactly analogous in properties with the dephlogisticated air, discovered on August
1st, 1774, by Joseph Priestley, by heating mercurius calcinatus per se by means of a
burning lens. When a burning candle was immersed in the gas, the candle burnt
with eclat eblouissant—hlmdmg brilliancy — as Lavoisier expressed it ; a smoulder-
ing splinter of wood burst into flame when plunged in the gas ; and the gas did not
suffocate a mouse like azote. A. L. Lavoisier first called this gas Vair eminemment
respirable, pur, ou vital, and afterwards oxygen. The latter term is its present-day
designation. In this manner, A. L. Lavoisier proved that atmospheric air is
made up of two gases— oxygen and nitrogen — of different and even opposite
natures, the oxygen alone combines with the metal during calcination, and is the
cause of the increase in weight.
A. L. Lavoisier further showed that the sum of the weights of the mercury and
oxygen obtained by heating mercury calx is exactly equal to the weight of the
calx ; and that the increase in the weight of the mercury in the formation of the
calx is equal to the weight of the oxygen taken from the air. In his Reflexions
sur le phlogistique (Paris, 1783) A. L. Lavoisier said that during the combustion of
phosphorus in oxygen gas (vital air) :
There is a total absorption of vital air, or rather of oxygen, in the combustion of phos-
phorus, and the weight of the phosphoric acid obtained is found to be rigoroasly equal
to the weight of the phosphorus, added to that of the vital air employed in the combustion.
The same agreement of weights is observed in the combustion of inflammable air, in the
combustion of charcoal, etc.
Hence, the mechanism of combustion according to A. L. Lavoisier is : Metal -|- Oxygen
= Metal calx, and not, as G. E. Stahl supposed to be the case : Metal - Phlogiston
= Metal calx. The phenomenon which occurs when oxygen unites with a metal
to form a calx is called oxidation, and the resulting calx is called an oxide.
A. L. Lavoisier thus showed that it is not the calces that are simple and the metals
compound, but just the reverse; so that the phlogistians have therefore been
said to have perceived the relations between these two classes of bodies upside
down. In all reductions with charcoal, said A. L. Lavoisier, fixed air is obtained
owing to the union of the charcoal with the pure portion of the air — oxygen
which was fixed in the calx during the oxidation of the metal.
If I take a metallic calx and heat it with carbon in a closed vessel, at the moment the
calx is reduced to the metallic state— at the moment, for example, when litharge, the calx
of lead, is changed into metallic lead, there reappears the air, which had become fixed
when the metallic lead had been made into a calx, and an aerial product — fixed air— -can
be collected at least a thousand times more bulky than the solid litharge employed, inia
experiment appears to be one of the most interesting which has been made smce the time
of Stahl.
Assuming that this interpretation of the experiments is correct, A. L. If vo^^ier
inferred that by mixing azote and oxygen in the right proportions, it ought to be
possible to reproduce atmospheric air. This A. L. Lavoisier did, and the mixture
was found to behave with respect to '' combustion, respiration, and the calcination
of metals similar in every respect to atmospheric air." Lavoisier similarly showed
that if the calcination of the metal is attended by the union of the vital air--oxygen
—of the atmosphere with the metal, then, when the calcination is ettected in an
inverted glass vessel containing vital air, the whole should be absorbed, inis
deduction was " proved by weight and measure." . . , ,
According to the phlogistians, a lighted candle burns because it is a compound oi
candle-matter and phlogiston. The compound is decomposed little by iittie, as
70 INORGANIC AND THEORETICAL CHEMISTRY
the candle burns from tip to base, and the phlogiston passes into the surrounding
atmosphere. A. L. Lavoisier inverted Vancienne hypothese. He supposed the
hydrogen and carbon of the candle, during the burning, to unite with the oxygen
of the air to form the oxides of carbon and of hydrogen ; and generally, when a
substance is burned, it does not give out an imaginary levitative phlogiston, but
rather takes in real gravitative oxygen.
In his Mhnoire sur la combustion en general (Paris, 1777), A. L. Lavoisier alto-
gether rejected the principle of combustion advocated by G. E. Stahl, and argued
that his own hypothesis " seemed to be more probable, more conformable with the
laws of nature, and to involve less strained explanations and fewer contradictions "
than the doctrine of G. E. Stahl. About ten years later, A. L. Lavoisier collected
and organized such an array of facts in defence of his proposition that he was able
to write with much greater confidence in his Reflexions sur le fhlogistique (Paris,
1783), and he claimed the phlogistic doctrine to be an error fatal to the progress of
chemistry :
If in chemistry everything can be satisfactorily explained without the aid of phlogiston,
it thereby becomes eminently probable that phlogiston does not exist, that it is a hypo-
thetical being, a gratuitoiis assumption.
It is easier to make new discoveries than to eliminate old prejudices. Chemists
were painfully slow to recognize the part played by air in combustion and calcina-
tion. In his Reflexions sur le pJilogistique (Paris, 1783), A. L. Lavoisier said :
Chemists have turned phlogiston into a vague principle, one not rigorously defined,
and which consequently adapts itself to all the explanations for which it might be required.
Sometimes this principle has weight, sometimes not ; sometimes it is free fire, sometimes it
is fire combined with the earthy element ; sometimes it passes through the pores of vessels ;
sometimes the vessels are impervious to it ; it explains both causticity and non-causticity,
transparency and opacity, colours and their absence ; it is a veritable Protean, changing
in form each instant.
A. F. de Fourcroy began to teach A. L. Lavoisier's theory in 1787 ; C. L. Ber-
thollet joined the new cause about the same time. Then followed L. B. Guy ton de
Morveau, and nearly all the French and British chemists. The Berlin Academy
abandoned phlogiston in 1792, and the controversy which had waged for some
years between the phlogistians was virtually at an end.i The downfall of phlogiston,
a relic of Egyptian and Chaldean lore, was celebrated by Madame Lavoisier,
habited as a Greek priestess, burning the writings of G. E. Stahl upon an altar
dedicated to the new positive science. At the beginning of the new century a few-
petrified spirits, unable to march to the music of the new chemistry, still lingered
behind. Robert Boyle's admonition in his Considerations touching experimental
essays in general (1661), may have been forgotten :
It ought to be esteemed much less disgraceful to quit an error for a truth than to be
guilty of the vanity and perverseness of believing a thing still because we once believed
it. . . . Until a man is sure he is infallible it is not fit for him to be unalterable.
The observed facts were sterile and barren before they were vivified by the
fire of Lavoisier's genius. Indeed, enthusiasts have said that chemistry as a science
was not born until A. L. Lavoisier's theory of burning had been demonstrated.
Many writers — e.g. S. Brown (1858) 2— have emphasized that tradition and
prejudice were all against A. L. Lavoisier, and however much he owed to his pre-
decessors and contemporaries — J. Rey, P. Bayen, J. Priestley, H. Cavendish, and
C. W. Scheele — he scarcely owed them one glimmering ray of thought — rather
the reverse. The legacies of fact inherited by A. L. Lavoisier were beclouded and
distorted by the false hypotheses through which their discoverers saw them, and
it required a master mind to co-ordinate the facts accumulated by many workers
into one system. We can feel with A. Wurtz when — following A. F. de Fourcroy
(1797) — he opened his Histoire des doctrines chimiques (Paris, 1869) thus : La chimie
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 71
est la science francaise, ellefut constituee par Lavoisier d' immortelle tnemoire; other-
wise expressed, chemistry is a French science, it was founded by Lavoisier of
immortal memory. This statement seems to have needlessly irritated some of
our own historical writers. Can we wonder that Frenchmen are proud of their
Lavoisier ^ Surely '' we can amiably pass without protest this ardent hero-
worship."
At first sight, it does seem curious that such a long period of time should have
been required to work from P. E. de Sultzbach's note in 1489 to the effect that
metals increase in weight when calcined in air, to A. L. Lavoisier's proof in 1774
that the increase in weight is due to the absorption of oxygen from the air. This
will occasion no surprise when we remember the difference between the properties
of air which cannot be seen, and the properties of solids and liquids which can be
readily seen and handled. As G. F. Rodwell has emphasized, the most obvious
property of matter is its visibility, and the conception of matter divested of this
quality is no small effort to a mind untutored in invisible bodies, which exercise no
apparent effect on surrounding objects, and it belongs to an advanced order of experi-
mental philosophy. There were no means of recognizing even the more salient
properties of air at the disposal of the chemists until a comparatively late period,
and the earlier chemists, accordingly, believed air to be intrinsically different in its
essence from more familiar visible substances. To illustrate the ideas about air
which prevailed at the end of the eighteenth century, the opening words of A. L.
Lavoisier's Memoire sur la nature du principe qui se combine avec les metaux pendant
leur calcination et qui en augmente le poids (Paris, 1775) may be quoted :
Do different kinds of air exist ? Is it enough that a body should be permanently ex-
panded for it to be considered a particular kind of air ? Are the different airs found in
nature or formed by us specific substances, or are they modifications of atmospheric air ?
Again, altogether apart from the skill required in the manipulation of gases,
it is not at all surprising that writers on chemistry in the Middle Ages failed to
interpret the mechanism of the burning of a candle in air when the knowledge
required to explain the chemical side of the phenomenon is recalled :
(i) Air is composed of two gases both sparingly soluble in water ; (ii) During combiistion
one of the gases unites and the other does not imite with the burning body ; (iii) Air
contains four volumes of the inert gas, and one volume of the gas which unites with the
burning body ; (iv) A gas soluble in water is produced during the combustion ; and (v)
The increase in weight of the combustible body during the burning is equal to the decrease
in the weight of the air.
The phlogiston hypothesis is sometimes held up to ridicule. It must be borne
in mind that the hypothesis was adopted by nearly all the leading chemists in the
earlier part of the eighteenth century when it appeared to be as firmly fixed among
the root principles of chemistry as the kinetic theory does to-day. Thus, the ardent
and devoted J. Priestley could say :
If any opinion in all the modern doctrine concerning air be weU-founded, it is certamly
this, that nitrous air is highly charged with phlogiston. If I have completely ascertamea
anything at all relating to air, it is this ;
and the diplomatic P. J. Macquer, in his Mernents de chymie pratique (Paris, 1751), that
We cannot say how phlogiston is fixed by substances ; but without pretending to gueM
the cause of the phenomenon, let us rest contented with the certamty of the tact.
The phlogiston theory represented the most perfect generalization known to the
best intellects of its day, and J. J. Becher and G. E. Stahl were the prophets of a
new mode of viewing chemical mutations. The doctrine served to give coherence
to the thoughts and work of a race of chemists extendmg from J.J. l^echer ana
G. E. Stahl down to H. Cavendish, J. Priestley, and C. W. Scheele.
The phlogistic hypothesis enabled chemistry to escape m part from mystic and
72 INORGANIC AND THEORETICAL CHEMISTRY
mediaeval empiricism,^ for it introduced a certain amount of order among a chaotic
mass of facts. Like phenomena were grouped together, and chemistry thrived and
multiplied its proportions while under its sway. Phlogiston prepared the way
for A. L. Lavoisier's balance, just as the balance heralded J. Dalton's arithmetic.
There is what A. Comte * called la hi de succession running through history. The
early struggles of man in quest of knowledge and truth were not in vain. The
sun-worshipper and the phlogistian, each in his own way, had been working to a
common end. All generations — past and future — thus seem to be linked in one
common service.
It is inconceivable that men like T. Bergmann, H. Cavendish, J. Priestley, and
C. W. Scheele would counsel what they considered to be an inconsistent doctrine.
Phlogiston was regarded by them, not as a temporary hypothesis, but as a permanent
acquisition, an enduring conquest of truth. To-day the word is but an empty
symbol.
It mxist be added that H. St. C. Deville (1860), C. Brown (1866), and W. Odling (1871) ^
have pointed out that phlogiston occupied a similar position in the chemistry of the eighteenth
century that potential energy does to-day. Said Deville : On arrive a admeUre que Vajjinite
(en intensite) n'est pas autre chose que la quantite de chaleur latente ou phlogistique enfermee
dans les corps. Even A. L. Lavoisier, in his Traite eUmentaire de chimie (Paris, 1. 60, 1793),
considered oxygen to be made up of caloric and the matter of oxygen. Lavoisier's caloric
• — a veritable ghost of phlogiston- — was supposed to be the matter of heat possessing no
weight whatever. Ordinary oxygen thus contained the principle of oxygen plus caloric.
The latter has also been identified with potential energy. Here then the old revives in the
new. The chemistry of to-day is not materialistic, for it is concerned with both energy
and matter.
Theories perish, facts remain. — Much of what we think best in the theories of
to-day may to-morrow be rejected, with phlogiston, worthless. There is no reason
to suppose that fewer errors are believed to-day than in the days when phlogiston
reigned supreme ; and it is not at all improbable that posterity will smile at our in-
explicable ignorance in some departments of science. This need cause the student
no embarrassment. A fallacious theory may be a valuable guide to experiment.
Experiment and labour applied to the explication of the most extravagant hypo-
thesis are not always lost. Guided by wrong hypotheses, men haves ought one thing
and found another ; Columbus sought the Indies and found America. W. Whewell
has pointed out that when a theory, which has been received on good evidence,
appears to fail, the really essential and vital part of it survives the fall, that which
has been discovered continues to be true. It is necessary, however, to follow
Rene Descartes advice : Give unqualified assent to no proposition which is not
presented to the mind so clearly that there is no room for doubt. As Aristotle
would have said, we do not need to cultivate the art of doubting, but rather the art
•of doubting well ; for the art of doubting well is the necessary antecedent of
progress. Doubt is the parent of inquiry.
It is not always expedient to follow the history of each hypothesis and each
conquest of truth, step by step, as in the case of air. That of course would be an
ideal plan of work ; but it is not always a waste of time to study the exploded fallacies
once cherished by the potentates of old. The right attitude of mind towards an
hypothesis or law can be developed only when history has taught how man has
had to climb with slow faltering steps until he obtained a clear view of each new
principle of chemical science. J. W. Goethe was quite right : The history of a science
is the science itself : The past is key to the present ; although, as A. Comte (1839)
expressed it : On ne connait pas complctement une science tant qu'on n'en sail que
Vhistoire. Unfortunately, time cannot always be spared to wander with the original
investigators into the byways of knowledge, and a more economical plan must
usually be followed. If every one had to pass through all the stages traversed by
all who have gone before, it would be impossible to reach the vantage ground gained
by the labours of his predecessors.
THE EVOLUTION AND METHODOLOGY OF CHEMISTRY 73
References.
1 J. Priestley, Considerations on the doctrine of phlogiston and the decomposition of loater,
London, 1796 ; R. Kirwan, An essay on phlogiston and the constitution of acids, London, 1789.
2 S. Brown, Essays, Scientific and Literary, Edinburgh, 1. 186, 1858 ; P. Duhem, La chimie est-
elk une science fran^aise ? Paris, 1916 ; R. Jagnaux, Histoire de la chimie, Paris, 1891.
* R. Lote, Les origines mystiques de la science allemande, Paris, 91, 1913 ; F. le Dantec, Rev.
Scient., 51. 740. 1913.
* A. Comte, Cotirs de philosophic positive, Paris, 1839 ; J. H. Bridges, Essays, Ix)ndon, 1907.
6 W. Odling, Proc. Boy. Inst., 6. 323, 1871 ; A. C. Brown, Proc. Roy. Soc. Edin., 5. 328, 1806 ;
H. St. C. Deville, Conipt. Rend., 50. 534, 584, 1860 ; Lemons sur la dissociation, Paris, 1860.
CHAPTER II
COMBINATION BY WEIGHT
§ 1. What is an Element ?
The elements count as simple substances not because we know that they are so, but
because we do not know that they are not. — J. von Liebig (1857).
A. L. Lavoisier showed that atmospheric air is no more an elementary principle
than the water of the ocean, for it can be resolved into two simpler gases — oxygen
and nitrogen. It is further possible to resolve all known substances — air, water,
etc. — into about eighty distinct elemental or primitive forms of matter. The
present-day concept of an element is one of those ideas which has gradually grown
into chemistry. Epicurus, about 300 B.C., held that corporeal things are either
composite, or else they are the constituent parts of which the composite things are
compounded ; and that the continued division of the composite must at last
furnish ultimate, indivisible, unchangeable particles of the elements. Aristotle, in
his De coelo (3. 3), also defined an element. He said :
Everything is either an element or composed of elements. . . . An element is that into
which other bodies can be resolved, and which exists in them either potentially or actually,
but which cannot itself be resolved into anything simpler, or different in kind.
This precise and accurate concept was soon beclouded with the idea that all the
different varieties of matter observed in nature are composed of a primitive element
with varying proportions of wetness or dryness, or of coldness or hotness. This
quaternary of attributes gradually materialized into earth, water, air, and fire.
All the different forms of matter were vaguely supposed to have been compounded
in some inscrutable manner from varying proportions of this quartet.
In 1661, Robert Boyle's attention was arrested by the loose way in which the
term element was employed, and in his The Sceptical Chymist (Oxford, 1661),
he gave a clear concept for an element. He said :
I mean by elements, as those chy mists that speak plainest do by their principles, certain
primitive and simple, or perfectly unmingled bodies ; which not being made of any other
bodies, or of one another, are the ingredients of which all those called perfectly mixt bodies
are immediately compounded, and into which they are ultimately resolved. ... I must
not look upon any body as a true principle or element, which is not perfectly homogeneous,
but is further resolvable into any niimber of distinct substances.
N. le Febvre,! whom J. B. Dumas called riiomme d' imagination, flourished in
the seventeenth century about the time of Robert Boyle. N. le Febvre showed
that Empedocles' analysis of wood into four elements — flame or fire, smoke or air,
moisture or water, and ashes or earth— does not include all the principles of which
this form of matter is compounded. By the destructive distillation of wood,
he found that water charged with acetic acid and an oily inflammable liquid
condensed in the receiver ; a gas escaped ; and charcoal remained. The charcoal
burnt in air, giving fire and ashes ; and the ashes were resolved by water into a
soluble salt, and an insoluble earth. N. le Febvre thus resolved wood into six
ingredients, and he got very near to recognizing that the only proof of an elementary
principle is the fact of its yielding nothing else to analysis. He maintained that
chemistry is not the doctrine of the four elements, an art of transmutation, or a
• 74
COMBINATION BY WEIGHT 75
science of mixts ; but is rather the art of analysis with a view to discover la con-
nmssance de toutes les choses que Dieu a tirees du chaos par la creation—a knowledge
of all the ingredients of all the various kinds of matter which God has created out
of chaos.
Even as late as the latter part of the eighteenth century, fire, air, water, and
earth were regarded as elemental. Thus, P. J. Macquer, in his Dictionnaire de
chymie (Pans, 2. 4, 1778), gave a juste definition of an element, and added :
Although fire, air, water, and earth are reputed to be simple, it is possible that they
are not so ; they may be very complex, and may result from the union of several other more
simple substances . . . but as experience teaches us absolutely nothing on this subject, we
may consider without inconvenience, and indeed in chemistry we ought to consider fire,
air, water, and earth as le8 corps simples, because they really act as such in all chemicaJ
operations.
We are also indebted to A. L. Lavoisier (1789) for further clarifying the concept
of an element. A. L. Lavoisier, quite logically, considered lime, magnesia, baryta,
and alumina to be elements. We now know that these elements of A. L. Lavoisier
are compounds of oxygen with calcium, magnesium, barium, and aluminium
respectively. This was not known to Lavoisier, and he rightly said : '* We are
certainly authorized to consider them simple bodies until, by new discoveries, their
constituent elements have been ascertained." Again, in 1811, the question whether
chlorine^ — ^then called oxymuriatic gas — was really an element or a compound of
oxygen with some other element was raised by Humphry Davy. H. Davy claimed
that chlorine is an element because, although oxygen was believed to be present,
none could be found. " Hence," added H. Davy, " we have no more right to say
that oxymuriatic gas (i.e. chlorine) contains oxygen than to say that tin contains
hydrogen. . . . Until a body is decomposed, it should be considered simple."
It is not possible to improve upon Lavoisier's conception of a chemical element,
and I feel compelled to quote his words, although written before 1789 : 2
When we apply the term elements or principles to bodies to express our idea of the leist
point which analysis is capable of reaching, we must admit, as elements, all substances
into which we are able to reduce bodies by decomposition. Not that we are entitled to
affirm that these substances which we consider as simple, may not themselves be compoimded
of two, or even of a greater number of more simple principles ; but since these principles
cannot be separated, or rather, since we have not hitherto discovered the means of separat-
ing them, they are, with regard to us, as simple substances, and we ought never to suppose
them compo\uided until experiment and observation have proved them to be so.
The definition of an element is not founded upon any intrinsic property of the
elements, but rather upon the limited resources of the chemist. To find if a given
substance is an element or a compound, it is usual to assume that it is a compound
and then to apply all known methods for resolving compounds into simple sub-
stances. If the methods fail to effect a decomposition, the substance is said to be
an element. Hence, the statement that any given substance is an element has
been said to be a confession of the impotence of human powers. In fine, element
is a conventional term employed to represent the limit of present-day methods of
analysis or decomposition. ^ We may, therefore, summarize these ideas in the
definition : An element is a substance which, so far as we know, contains only
one kind o£ matter. To say the substances we call elements caw/<of be decomposed
may be regarded as an unwarranted reflection on the powers of our successors.
The moment Auer von Welsbach (1885) proved that didymium was a mixture of
praseodymium and neodymium, one element ceased to exist, and two elements
were born. If it were found to-morrow that the element chlorine is really a com-
pound of two new elements previously unknown, the fact would be important
and it would change the face of chemistry, but it would not render useless any facts
we know about chlorine. . . ,
The old alchemists sought to transform some of the common metals mto go d.
Whenever the attempt has been made with materials known to be free from gold,
76 INORGANIC AND THEORETICAL CHEMISTRY
no transmutation has been observed. There is nothing intrinsically absurd in the
notion, but at present, no authentic transmutation has been deliberately, or rather
intentionally, accomplished. Works like P. J. von Lewinheim Sachs' Aurum
chymicum (Genevae, 1702) and K. C. Schmeider's Geschichte der Alchemie (Halle,
1832) professed to examine critically the authenticity of the legendary reports by
the alchemists of the reality of the transmutation of the metals, and concluded that
in some cases the legends are above suspicion, and this in spite of the fact that
H. von Osten, in his Eine grosse Herzstdrkung fiir die Chymisten (Berlin, 1771), had
exposed some forty-five tricks and deceptions practised by alchemical knavery.
All the reports now stand discredited. K. C. Schmeider criticized the legends
imperfectly, and failed to recognize that fictions may be plausible as well as
extravagant. When the evidence has permitted a critical examination, every re-
corded instance has been traced to a mal-observation ; and evidence which cannot
be tested is outside the range of scientific methods. In the words of J. M. Wilson
(1917), in science, there is no statement of fundamental importance that depends
on history or on any testimony which cannot be verified.
The next inquiry arises from the question : What relations subsist between
the weights and volumes of the different elements which make up the different
kinds of matter known to man ?
References.
1 N. le Febvre, Traicte de la chymie, Paris, 1660 ; London, 1664 ; J, B. A. Dumas, LeQons sur la
philosophie chimique, Paris, 51, 1837 ; S. Brown, Lectures on the Atomic Theory, Edinburgh, 9,
1858.
2 A. L. Lavoisier, Traite elementaire de chimie, Paris, 1789 ; H. Davy, Phil. Trans., 98. 39,
1808 ; 99. 450, 1809 ; 100. 231, 1810 ; 101. 1, 1811.
* H. Spencer, Essays, London, 3. 234, 1891 ; Justus Liebicfs und Friedrich Wohler's Brief-
wechsel in dem Jahren 1829-1873, Braunschweig, 2. 43, 1888.
§ 2. The Law of Constant Composition— Proust's Law
Nature in her inscrutable wisdom has set limits which she never oversteps. — Jean Rey.
The proportion in which one element can unite with another is fixed by nature, and the
power of augmenting or diminishing this pondua naturce is not given to man.' — J. L. Proust
(1801).
Attention must now be directed to the singular observation made by Jean Eey
(1630) that during the calcination of a metal in air, " the weight of the metal
increased from the beginning to the end, but when the metal is saturated, it can
take up no more air. Do not continue the calcination in this hope : you' would
lose your labour." The examples previously quoted — Cap. I, Table I — have shown
that one gram, and only one gram, of air is absorbed by definite amounts of the given
metals under the conditions of the experiment, and Lavoisier's work proves that
the oxygen of the air is alone absorbed. Accordingly, one part by weight of oxygen
is equivalent to 1-52 parts magnesium ; 4-06 of zinc ; 1*12 of aluminium ; 3-97 of
copper ; and 3" 72 of tin. Instead of taking the weight of oxygen unity, it will be
more in accord with general usage to make oxygen 8. Hence, multiply the
preceding numbers by 8 :
Oxygen. Magnesium. Zinc. Aluminium. Copper. Tin.
8 1216 32-48 8-96 3196 29*76
When magnesium is calcined in the presence of oxygen, or air, the metal always
unites with the oxygen in the proportion of one part of oxygen per 1*52 parts of
magnesium, or 8 parts by weight of oxygen per 12-16 parts by weight of magnesium.
The same principle obtains when magnesium oxide is made in several dilierent ways ;
and likewise with the other metallic oxides. The proportions in which two elements
unite together do not vary in a fortuitous manner, but in fixed and definite
COMBINATION BY WEIGHT 77
proportions. Hence, as P. J. Hartog 1 puts it : two like portions of matter
have the same composition. The converse of this statement is not necessarily
true, for two portions of matter compounded from the same proportions of the
same elements are not necessarily alike.
The exact work of J. S. Stas 2 and T. W. Richards and many others has firmly
established the constancy of the composition of the regular type of chemical com-
pounds. J. S. Stas, in his famous Recherches sur les lots des proportions chimiqites
(1860-65), for example, studied among other things, the composition of silver
chloride prepared by four different processes at different temperatures. He found
that 100 parts of silver furnished 132-8425, 132'8475, 132-8480 parts of silver
chloride ; and that neither the temperature nor the method of preparation had any
influence on the composition of the chloride. The difference between the two
extremes is less than 0-006 part per 100 parts of silver. This shows that the errors,
incidental to all experimental work, are here remarkably small. J. S. Stas likewise
proved that ammonium chloride prepared from quite different sources, and purified
in different ways, always contains exactly the same proportion of chlorine. Still
further, he proved that the combining weight of an element is not affected in the slightest
degree hy the various elements with which it might combine. For example, silver com-
bines with iodine to form the iodide, and with iodine and oxygen to form the iodate.
The ratio of silver to iodine in both compounds is the same, and is not modified
by the large quantity of oxygen present in the iodate. Hence, J. S. Stas stated :
" If the recognized constancy of stable chemical compounds needed further de-
monstration, I consider the almost absolute identity of my results has now com-
pletely proved it."
The law o£ constant proportions, however, can never be proved with mathe-
matical exactness. However skilful a chemist may be, it is impossible to make an
exact measurement without committing an error of observation or an error of
experiment. It is assumed that the small difference O'OOS per cent, between the
two extreme results of J. S. Stas (1) is wholly due to the unavoidable errors of
experiment, for we cannot expect an exact solution of the problem ; and (2) is
not due to a very slight inexactitude in the law of constant proportions. In 1860,
J. C. G. de Marignac considered that the experiment did not prove definitely
that the composition of compounds might not vary within very minute
limits :
I do not consider that it has been absolutely demonstrated that chemical compounds
do not normally have an excess of one of the constituents. It is true that this excess is
very minute, but it is still appreciable in very delicate measurements.
The composition of a definite compound appears to be independent of its mode
of formation, and therefore it is inferred that substances always combine in definite
proportions. If an excess of one substance be present, the amount in excess remains
uncombined as extraneous matter. This deduction from the observed facts is
called the law of definite proportions, or the law of constant composition. The law
is sometimes enunciated : a particular chemical compound always contains the same
elements united together in the same proportions — hy mass. This statement, if inter-
preted literally, holds good for a particular mixture, as well as for a particular com-
pound ; and it has nothing to say as to the distinction between a mixture and a
compound. Probably no generalization in chemistry is more firmly established than
that like compounds possess the same quantitative composition. So great is the faith
of chemists in the truth of this generalization that a few accurate and careful
experiments are considered sufficient to settle, once for all, the composition of a sub
stance. For instance, if a substance possessing all the properties of magnesium
oxide be given to a chemist, without taking any more trouble, he knows that it
will contain 12-16 parts of magnesium for every eight parts of oxygen. The law
of constant composition furnishes a kind of a priori control over quantitative
analysis. Constancy in composition is regarded as a proof of purity, and purity
78 INORGANIC AND THEORETICAL CHEMISTRY
is attended by constancy in composition. Hence arose the concept of a chemical
compound as distinct from a mixture. ^
References.
1 P. J. Hartog, Nature'bO. 149, 1894 ; B. A. Rep., 618, 1894.
* J. S. Stas, (Euvres completes^ Bruxellos, 1894 ; T. W. Richards, Experimenlelh Unler-
auchungen uher Atomgewichte, Hamburg, 1909.
» E. J. Mills, Phil. Mag., (4), 40. 259, 1870.
§ 3. History o! the Law of Constant Composition
Ce n*est que du conflit des opinions contraires que jaillit la veritc'-. — ¥. Hoefer (1843).
The law of constant composition was not discovered by any particular man,
but it gradually grew among the doctrines of chemistry. The law was tacitly
accepted by many before it was overtly enunciated. This is shown by J. Rey's
views (1630), previously stated. In 1699, G. Homberg,^ in his Observations sur la
quantite d'acides absorbes par les alcalis terreux, described measurements of the
amounts of different acids (vinegar, spirits of salt, aqua fortis, and vitriolic acid)
required to saturate a given amount of potassium carbonate (salt of tartar) ; he
evaporated the saturated liquid to dryness and weighed the resulting solid. His
results were compiled in the form of a table which has been regarded as embodying
the first hint of the law of definite proportions. G. Homberg considered that the
quantity of ^n acid which unites with an alkali is la mesure de la force passive de eel
alcali, and further added that by la force des acides he means the solvent action of
the acid, and similarly for the alkalies. F. Hoefer (1843) suggests translating
G. Homberg's " solvent action " by " neutralizing," and " solubility " by " neu-
tralizable." Isaac Newton referred to the saturation capacity of acids for different
metals in the thirty-first query of his Opticks (London, 1704). G. E. Stahl also in
his Fundamenta chymim (Norimbergse, 1723) spoke of the pondus naturcp. as the
proportions which ought to exist between the masses of the different ingredients
in order that a determinate compound be produced. In 1717, E. F. Geoffrey
analyzed saltpetre and stated its quantitative composition. A. S. Marggraf
(1749) ; H. T. Scheffer (1750) ; T. Bergmann (1775-84) ; R. Kirwan (1790-1800) ;
J. Black (1794) ; M. H. Klaproth (1795) ; V. Rose (1803-5) ; C. F. Bucholz
(1799-1802) ; and L. N. Vauquelin (1812) all based analyses of chemical com-
pounds on the tacit assumption that this law is valid ; and W. Higgins' theory of
atoms (1789) implies that chemical compounds must have a constant composition.
A. L. Lavoisier appears to have had no doubts on the subject. In every oxide, said
he, the relation of oxygen to the metal is constant.
In 1767, H. Cavendish said that those quantities of bases — e.g. potash and
lime — are equivalent which neutralize the same amount of acid ; and, in 1788,
he showed that this equivalency is independent of the nature of the acid. C. F.
Wenzel (1777) had a fairly clear idea that a definite weight of a base neutralized
a definite amount of a given acid, and in his Lehre von der Venvandschafi der Korper
(Dresden, 1777), he gave measurements of the weights of over twenty metals and
bases which were required to saturate about a dozen acids ; and he also examined
quantitatively the products of some reactions — e.g. copper sulphate and lead acetate ;
mercuric sulphide and silver chloride ; etc. Shortly after C. F. Wenzel's book had
appeared, J. B. Richter, in an important study of this subject, published evidence
in his Ueber die neueren Gegenstdnde der Chemie (Breslau, 1791-1802), and in his
Anfangsgrunde der Stocky ometrie oder Messkunst chymischer Elemente (Breslau, 1792-
4), which demonstrated conclusively that the weights of the various acids which
neutralize certain fixed weights of the bases are the same ; and the same
numbers hold good for the neutralization of all acids by the bases ; otherwise
expressed : Acids and alkalies unite in constant proportions to form salts- this
COMBINATION BY WEIGHT 79
is Richter's law of proportionality, or Richter's law of equivalent ratios. Conse-
quently, it is possible to assign equivalent numbers to the acids and bases. For
instance, using modern data and terms :
Acids.
Equivalent
weight.
Bases.
Equivalent
weight.
. 3505
. 37-06
. 40-01
. 56-00
Hydrofluoric acid
Hydrochloric acid
Sulphuric acid
Nitric acid .
. 2001
. 36-47
. 49-04
. 63-02
Ammonium hydroxide .
Calcium hydroxide
Sodium hydroxide.
Potassium hydroxide
J. B. Kichter gave separate tables for the neutralization equivalents of each acid
and each base ; but Gr. E. Fischer, in an appendix to his German translation of
C. L. Berthollet's Recherches sur les his de Vaffinite, showed that J. B. Richter's
data could be reduced to a single table containing twenty-one numbers divided into
two columns as just indicated. These tables can be regarded as the first tables of
equivalent weights published. The weights of the acids in one column represent
the amounts required to neutralize the quantity of any of the bases indicated in
the other column ; and conversely, the weights of the bases in the second colunm
represent the amounts required to neutralize the quantity of any one of the acids
indicated in the first column. Thus 56 grams of potassium hydroxide will neutralize
20"01 grams of hydrofluoric acid, 36'47 grams of hydrochloric acid, 49*04 grams of
sulphuric acid, 63*02 grams of nitric acid, etc., and 63"02 grams of nitric acid will
neutralize 35'05 grams of ammonium hydroxide, 37*06 grams of calcium hydroxide,
etc. Richter claimed that the rule he gave is a true touchstone — Probierstein — for
the proportions wherewith the acids and bases neutralize one another, because if
the observed numbers do not agree with those demanded by the rule, they may
be regarded as erroneous.
J. B. Richter mixed up much valuable work with several fantastic hypotheses ;
he supposed that the weights of the bases required to neutralize a constant weight
of acid are in arithmetical progression ; and the weights of the acids required to
neutralize a constant weight of any base are in geometrical progression. Richter
appears to have cooked some of his results to make them fit his erroneous hypo-
thesis so that the numbers represent what he thinks he ought to have obtained
rather than what he actually observed. Such a procedure is quite antagonistic to
the spirit of science, and made chemists reasonably sceptical about the accuracy
of the whole of Richter's work. It was thought, wrongly as it happens, falsus in
uno, falsus in omnibus (false in one, false in all). Consequently, the above
generalization did not attract the attention it deserved. On reading J. B. Richter's
work on chemical ratios, said J. J. Berzelius (1827), we are astonished that the
further study of the subject could ever have been neglected.
It must be added that the validity of the law of definite composition was the
subject of an interesting controversy during the years between 1800 and 1808.
J. L. Proust 2 maintained that constant composition is the invariable rule ; C. L.
Berthollet did not assert that cases of constant composition are non-existent, but
he argued that these instances were due to special circumstances, and maintained
that constant composition is the exception, variable composition the rule. J. L.
Proust's words are worth quoting :
According to my view, a compound is a privileged product to which nature has assigned
a fixed composition. Nature never produces a compound even through the agency ot
man, other than balance in hand, pondere et messura. Between pole and pole compounds
are identical in composition. Their appearance may vary owing to their manner ot aggre-
gation, but their properties never. No differences have yet been observed between tne
oxides of iron from the South, and those from the North ; the cmnabar of Japan has tne
same composition as the cinnabar of Spain ; silver chloride is identically tlie same whetner
obtained from Peru or from Siberia ; in all the world there is but one sodnini chloride ;
one saltpetre ; one calcium sulphate ; and one barium sulphate. Analysis confirms tnese
facts at every step,
It might be thought that positive assertions of this kind, backed by accurate
80 INORGANIC AND THEORETICAL CHEMISTRY
experimental work, would leave no subject for disputation ; but, surveying the
battlefield in the light of the present-day knowledge, it seems that another quite
different phenomena was confused with the law of constant composition ; and the
methods of analysis were not very precise. Some, probably from the unfounded
belief that " Proust deservedly annihilated Berthollet," call the generalization
discussed in this chapter, Proust's law. The arguments against the law of constant
composition was silenced not by J. L. Proust, but by the work which developed
from J. Dalton's atomic theory. 3 J. L. Proust did not satisfactorily answer all
C. L. Berthollet's objections.
According to C. Daubeny (1850) it has been stated that something hke the theory of
constant composition can be found among the dogmas of the old sage Pythagoras (c. 520).
This philosopher is sometimes supposed to have derived what is the most valuable part of
his philosophy from the Egyptian priests during his sojourn in Egypt. Pythagoras taught
that number — whatever was meant by that term — is a bond sustaining by its power the
permanent existence of everything on the earth. The influence of Pythagoras has been
traced in the doctrine laid down by Philo the Jew— or who ever wrote the apocryphal
book of wisdom— God ordained all things by measure, number, and weight. It is, however,
certain that European chemistry is in no way indebted to the Egyptian priesthood or to
the Pythagorean philosophy for the concept of the law of constant composition. It would
indeed require the exercise of a good deal of ingenuity to disentangle the law of chemical
combination from the conflicting statements which have been made as to the meaning to
be attached to Pythagoras' doctrine of numbers.
F. Wald (1895-9) ^ argues that the composition of chemical compounds is
variable, and that the observed constancy in the composition of chemical com-
pounds must be attributed to the selection by chemists of special preparations.
Hence, says F. Wald, the statement of the law of constant composition is quite
empirical, and the assumption that these selected preparations are alone true com-
poimds is quite arbitrary.
References.
1 G. Homberg, Mim. Acad., 64, 1700.
2 J. L. Proust, Ann. Chim. Phys., { 1), 32. 26, 1799 ; Journ. Phys., 53. 89, 1801 ; 55. 325, 1802 ;
59. 260. 265, 321, 350, 1 804 ; 60. 347, 1805 ; 63. 421, 1806 ; C. L. Berthollet, Pecherches sur les lois de
Vaffinite., Paris, 1801 ; Essai de statique chimique, Paris, 1803 ; Journ. Phys., 60. 284, 347, 1805 ;
61. 352, 1805.
3 A. N. Meldrum, Mem. Proc. Manchester, Lit. Phil. Soc.,5^. 7, 1910 ; C. Daubeny, An Intro-
duction to the Atomic Theory, London, 1850.
4 F. Wald, Zeit. phys. Chem., 17. 337, 1895 ; 19. 607, 1896 ; 22. 253, 1897 ; 23. 78, 1897 ; 24.
315, 634, 1897 ; 25. 525, 1898 ; 26. 77, 1898; 28. 13, 1899 ; Chem. Ztg., 30. 963, 978, 1906 ; 31. 756,
769, 1907 ; 0. Kuhn, ib., 31, 688, 1907 ; 32. 55, 1908 ; E. Bauer, Zeil. anorg. Chem., 50. 199,
1906 ; C. Benedicks, ih., 49. 284, 1906 ; L. Henry, Bvll. Acad. Belgique, 975, 1904 ; S. Cannizzaro,
Rend. Soc. Chim. Roma, 2. 128, 1904; R. Nasini, Rend. Accad. Lincei, (5), 5. 119, 1905;
L. Duhem, Le mixte et la comhinaison chimique, Paris, 1902 ; W. Ostwald, The Fundamental
Principles of Chemistry, London, 1909 ; Journ. Chem Soc., 85. 506, 1904.
§ 4. Pure Substances
Pure water is never found in nature. One may oven say that no man has over seen or
handled absolutely pure water. It is an ideal substance, to which some specimens of
highly purified water have nearly approached. — M. M. P. Mum.
It is only in " tall talk " or in advertisements that any human preparation, elementary
or not, can be spoken of as chemisch rein. — P. G. Tait (1881).
The substance we call water has its own properties, but sea-water, spring-water,
rain-water, and distilled water show certain differences in their properties. The
differences, however, are not due to the water, but to the substances — impurities —
which the water has dissolved from its surroundings. If sea-water be distilled, the
'* impurities " — sodium chloride, magnesium chloride, etc. — remain behind. Sea-
water is therefore a homogeneous substance, but, rightly or wrongly, it is often
stated to be a mixture, because water and various salts can be separated by simple
COMBINATION BY WEIGHT 81
evaporation or by freezing. Table salt is more or less impure sodium chloride
The presence of a little magnesium chloride in table salt makes the salt more
hygroscopic, so that the contaminated table salt deliquesces more readily than if
magnesium chloride were absent.
The term deliquescence— from deliqnescere, to melt or dissolve —refers to the process
of absorbmg moisture from the air so that a salt becomes moist, or even dissolves in the
moisture it has absorbed from the air ; e.g. when potassium carbonate is exposed to the
atmosphere it rapidly gains in weight. The term hygroscopic— from Sypos, wel^refers
to the absorption or adsorption of moisture from the atmosphere. Most substances
particularly when powdered, are hygroscopic, even if they do not deliquesce. The term
efflorescence from efflorescere, to blossom, refers to the formation of a crust generally
white — on the surface of a body. The phenomenon is very often due to the loss of water
from the surface of certain crystalline salts ; e.g. when crj^stals of washing soda are exposed
to a dry atmosphere, they gradually lose weight.
Air is a mixture of oxygen and nitrogen, with a little carbon dioxide, and it is
habitually moist owing to the presence of a varying proportion of water vapour.
In a chapter contained in J. B. Porta's Magice naturaUs (Naples, 1589), on the
extraction of water from air, it is shown that if a large glass flask be filled with a
mixture of ice and nitre, water condenses from the air to the outer walls of the
vessel, and trickles down into a basin below as receiver. Isaac Newton i said that
potassium carbonate deliquesces in air because of an attraction between the salt
and the particles of moisture in the atmosphere, and asked : Why does not common
salt or nitre deliquesce in the same way except for want of such an attraction ?
In H. B. de Saussure's Essais sur Vhydrometrie (Neuchatel, 1783) there is an
excellent study of the moisture which is normally present in atmospheric air. He
exposed " equal quantities of salt of tartar, quicklime, wood, lime, etc., all dried
as perfectly as possible," to the same air, and found that they " imbibed water
and increased in weight in unequal quantities." The salt of tartar took more than
the lime, and the lime more than the wood. H. B. de Saussure said that " these
differences can only proceed from the different degrees of the affinity of these bodies
for water," and he called this affinity, the hygroscopic affinity of the bodies for the
vapour, so that the amount of vapour imbibed by different substances from the air
" is proportional to their affinity for water vapour." H. B. de Saussure also showed
that the thirst or the attractive force of the body for aqueous vapour diminishes
from moment to moment " in proportion as it drinks the vapour," otherwise
expressed, the hygroscopic activity of the body diminishes in proportion as it
approaches the point of saturation.
Lavoisier's experiments on the transformation of water into earth.— A com-
pound may be contaminated with impurities in many ways — from the raw materials
used in preparing the compound ; from the vessels in which it was prepared or
stored ; by exposure to the atmosphere ; by the partial decomposition of the
substance when exposed to light, etc. It was once believed that air can be condensed
to water, as was thought to be proved by the falling dew ; and that water can be
changed into an earth, as is evidenced by the residue obtained when rain-water or
distilled water is evaporated to dryness in glass vessels. Thus, 0. Borrichius in
his Hernietis, Mgyjptorum et cheynicorum sapientia (Hafnia?, 1674), said that " when
100 pounds of snow, hail, or of rain-water, are evaporated, the water is transformed
into a dusty earth which contains some common salt ; " R. Boyle ^ found on
distilling and re-distilling pure rain-water, time and again, in glass vessels, a white
powdery substance was obtained each time the water was evaporated ; and the
more the water distilled from a given glass vessel, the larger the amount of whit€
powder. He added that a friend, of unsuspected credit, had distilled water two
hundred times " without finding the liquor grow weary of affording the white earth."
It seemed to him as if water " might be very nigh totally brought into earth, since
out of an ounce of distilled rain-water he had already obtained nearly three-quarters
of an ounce, if not more, of the often-mentioned earth." A. L. Lavoisier 8 first
VOL. 1. ^
82 INORGANIC AND THEORETICAL CHEMISTRY
traced the true source of this earth. In his paper Sur la nature de Veau et sur les
experiences par lesquelles on a pretendu prouver la possihilite de son changement en
terre (1770), A. L. Lavoisier described experiments with the object of " settling by
decisive experiments whether water can be changed into earth as was thought by
the old philosophers, and still is thought by some chemists of the day." By heating
water in hermetically sealed glass vessels, after some days, the water became turbid
and little white specks separated from the water and floated about. The hermeti-
cally sealed glass vessels were weighed before and after the experiment ; it was
proved : (1) The earth does not come from outside the vessel, because the weight
of the vessel and its contents does not alter. This is against Boyle's hypothesis
that fine igneous particles are able to pass through the glass, and are precipitated
in the form of a white powder when they come in contact with water.
Consequently, I conclude that nothing can pass through the pores of the glass, and these
little white particles, be they caused by what they may, are not caused by igneous particles
passing through the glass.
Still further, it was shown (2) The earth does not come from the water, because
the weight of the water remains the same before and after the experiment ; (3) The
earth comes from the vessel, because the vessel loses in weight ; and (4) The earth
comes wholly from the vessel, because the loss in weight of the vessel is virtually
equal to the weight of the earth formed. Hence, adds Lavoisier, " it follows from
these experiments that the greater part, possibly the whole of the earth separated
from rain-water by evaporation, is due to the solution of the vessels in which the
water has been collected and evaporated." C. W. Scheele (1777) ^ deduced a similar
conclusion from other experiments. He analyzed tlie earth produced during the
evaporation of water in glass vessels and showed that it has a similar composition
to the stuff of which the vessel was made.
K. F. von Walther (1915) has an interesting experiment to demonstrate the solubility
of glass in water, 500 c.c. of water are placed in a common litre flask with sufficient alizarine
to produce a pale yellow colour, the colour changes to a reddish -violet owing to the dis-
solution of alkali from the glass. By adding dilute sulphuric acid from a burette, the colour
changes back to pale yellow when the alkali is neutralized. He found that after an hour's
boiling, alkali equivalent to 18*3 c.c. of centinormal sulphuric acid had been dissolved from
a glass vessel.
The purity of commercial compounds. — The term pure or cJiemically pure, is
unfortunately used when it is desired to emphasize the fact that the substance has
not sufficient impurity to influence appreciably the most exact work for which it
is to be employed. There cannot be degrees of purity. A thing is either pure or
impure. It may be convenient to use terms like highly pure, all but pure, very
impure, etc., but the term, chemically pure, in the sense of nearly pure, is objection-
able. This recalls Basil Valentine's statement that water exists in three degrees
of excellence — ^the pure, the purer, and the purest ! The labels on commercial
reagents with their pare, purissimum, and chemically pure, are almost equivalent.
F. Mylius * proposed distinguishing degrees of purity as of the first, second, and
sixth grades according as they contain one part of total impurity in 10, 10^, . . .
10* parts.
The terms reagents and chemicals are applied to the substances used in chemistry
for producing special reactions with other substances. The former term is more
particularly used in analytical work. Chemically pure substances, paradoxical as
it may seem, are sold with a statement on the labels indicating what impurities are
present as well as how much of each. Commercial reagents, on the other hand,
have not been specially purified, and hence are sold at a cheaper rate than the
chemically pure substances. Purification is an expensive operation, and the cheaper
commercial reagents are used whenever specially purified materials are not required.
Some hold that " perfectly pure substances are unknown." This is possible, but to
establish the proposition, we should be involved in a metaphysical discussion, and
\^e might be led to say with A. Laurent : " Chemistry is the science of substances
COMBINATION BY WEIGHT 83
which do not exist," or perhaps with G. W. F. Hegel : " Pure being is pure
nothing."
Positive and negative evidence. — One positive proof may demonstrate an
indefinite number of negatives. Thus, if a test proves that a given substance is
silver chloride, it at the same time proves that the substance is not metallic copper,
arsenic oxide, etc. On the other hand, inability to prove a direct negative is not
to be regarded as equivalent to a positive proof. Thus, let it be asserted that a
third substance, say moisture, must be present when two substances interact chemi-
cally. Against this, it can be shown that substances like mercury and chlorine do
react when most carefully purified and dried ; but it could be, and has been argued
that this circumstance is due to the presence of an unrecognized impurity. Similarly,
some argue that if the elements could be obtained absolutely free from unknown
impurities their atomic weights would be whole numbers. Negative arguments of
this type are invulnerable in controversies because they cannot be controverted by
proofs to the contrary. True, the most skilful workers with the most refined
instruments cannot find an impurity, but still, it can be asserted that better equipped
searchers might be more successful. This might, however, does not prove the thesis
in question. Nevertheless, the argument is often used. For instance, H. Davy's
quest for oxygen in chlorine ; T. Bergmann's proof of the individuality of nickel ;
W. Ostwald's statement that catalytic agents can change only the velocity of
existing reactions ; etc.
The effect of traces of impurity on the properties of a compound. — It may be
well to emphasize, just here, that sometimes a minute trace of impurity is of vital
importance. Some reactions proceed quite differently in the presence, and in the
absence of traces of moisture or maybe other impurities. The properties of many
substances, too, are modified in a remarkable manner by small traces of impurity.
H. Vivian says that f^th part of antimony will convert the best selected copper
into the worst conceivable ; Lord Kelvin, that the presence of ^^^th part of
bismuth in copper would reduce its electrical conductivity so as to be fatal to the
success of the submarine cable ; H. le Chatelier, that the absorption of a quite
imperceptible weight of gas changes the melting-point of highly purified silver nearly
30° ; G. le Bon (1900) that the presence of ,-Woo*^ part of mercury in magnesium
makes the metal decompose water at ordinary temperatures ; and W. C. Roberts
Austin, that ^t]i part of bismuth in gold will render gold useless from the point
of view of coinage, because the metal would crumble under pressure in the die.
J. F. W. Herschel (1851) considered the fact that such minute proportions of
extraneous matter should be found capable of communicating sensible properties
of a definite character to the bodies with which they are mixed, to be perhaps
one of the most extraordinary facts that has appeared in chemistry.
References.
^ I. Newton, Opticks, London, 1704. . . ,
2 R. Boyle, The origin of forms and qualities, Oxford, 1666 ; A. L. Lavoisier, Mem. Acad^
73, 107, 1770. ^^ , ___
3 C. W. Scheele, Chemische Abhandlungen von der Luft uvd dem Feuer, Upsala, 1777.
* F. Mylius, Zeit. Elektrochem., 23. 152, 1917; T. S. Hunt, Amer. Journ Science (i), Jb. i n>,
226, 1853 ; (2), 16. 203, 1854; R. F. von Walther, Journ. prakt. Ckem., (2), 91, 33A l^i^-
§ 5. Physical and Chemical Changes.
Most of the substances belonging to our globe are constantly undergoing aJtor«^»^" jj^
sensible quantities, and one variety becomes as it were transmuted mto another &uc
changes, whether natural or artificial, whether slowly or r^P^^i^y Pf ^°"^*';*; ,*rj„Xa
chemical. Thus, the gradual and almost imperceptible decay of the leaves and branches
of a fallen tree exposed to the atmosphere, and the rapid combustion of v^ood in our hres,
are both chemical operations.— H. Davy. ,
The early chemists did not clearly distinguish between uniform mixtures and
homogeneous compounds so that many substances now known to be mechanical
84 INORGANIC AND THEORETICAL CHEMISTRY
mixtures were classed with substances known to be homogeneous compounds ;
again, owing to the fact that they were seldom able to prepare compounds of a high
degree of purity, the properties of compounds seemed more or less variable. Even
so late as the end of the seventeenth century, chemists were not all clear that
substances could be obtained with fixed and invariable properties. The properties
of a substance are those qualities, or attributes, by which its nature is manifested.
About 1730, H. Boerhaave distilled mercury five hundred times with the idea of
finding if its properties thereby suffered any change. About this time, it was
recognized that a homogeneous pure substance always has the same properties and
behaves in the same way, when the conditions are the same ; and generally, that
one element or compound is distinguished from all other elements or compounds
in possessing certain specific and characteristic properties ; or, in the words of an
old alchemist : " God hath sealed each substance with a particular idea."
First and foremost, a chemical compound has a fixed and definite composition ;
then again, a compound or element usually melts and boils at definite temperatures ;
its specific gravity, specific heat, crystalline form, colour, odour, behaviour when
in contact with other substances, etc., are characteristic of one particular chemical
compound. When the melting-point of, say, pure silver chloride has been once
accurately determined, it follows that all other samples of pure silver chloride will
melt at the same temperature under the same conditions. Many changes in the
properties of matter are not immediately perceptible to- the senses ; and in the
majority of cases, the processes for the identification and difierentiation of the
different forms of matter are based upon their behaviour towards certain reagents.
The more salient characteristic properties of an element or compound are employed
for its identification — that is, for distinguishing it from other known elements or
compounds. Thus, a student would be probably correct in stating that a solution
contained a silver compound if it gave a white precipitate when acidified with
hydrochloric acid, and the precipitate was insoluble in hot water, and soluble in
aqueous ammonia ; and if the spectrum of a burning body has a yellow line in a
particular part, the presence of sodium would be inferred. About 1661, Robert
Boyle 1 noticed many examples of the use of chemical reagents for the detection
or identification of certain substances, and in 1780, T. Bergmann collected together
a number of reagents useful for detecting the commoner elements or acids, and
described the effects produced. M. H. Klaproth, L. N. Vauquelin, J. J. Berzelius,
F. Wohler, H. Rose, C. R. Fresenius, and others built upon this foundation the
present system of qualitative analysis. ^
Physical changes. — -When liquid water becomes ice or steam there is no change
in the chemical nature of the substance, for the matter which makes steam and ice
is the same in kind as that of liquid water. A substance can generally change its
state, as when liquid water becomes steam or ice. The idea is further emphasized
by the fact that in most cases a substance is called by the same name, whether it
be in a solid, liquid, or gaseous state of aggregation. For instance, we speak of
liquid oxygen, liquid air, tnolten silver chloride, etc. Again, matter may change
its volwne by expansion or contraction ; it may change its texture, as when a porous
solid is fused to a vitreous mass ; it may change its magnetic qualities, as when a
piece of soft iron in contact with a magnet attracts other pieces of iron, etc. It is
conventionally agreed to say that in none of these cases of physical change is there
any evidence of the formation of a new substance ; and that the matter does not
lose or change those properties which distinguish it from other forms of matter.
A physical change involves an alteration in the properties of a substance without
the formation of a new substance.
Chemical changes.— When magnesium metal is heated in air, a white powder
is formed, and when mercuric oxide is similarly treated, mercury and oxygen are
obtained. The action of heat in both cases furnishes forms of matter with very
different specific properties from those forms of matter employed at the start. A
chemical change involves the formation of a fresh substance with different
COMBINATION BY WEIGHT 86
specific properties from the original substance or substances. In both chemical
and physical changes the total weight of matter before and after the change remains
constant, but in chemical changes alone the kind of matter alters.
It is not always easy to distinguish between physical and chemical changes,
because the only real distinction between the two turns on the question : is there
any evidence of the formation of a new substance during the change ? The evidence,
as we shall soon see, is not always conclusive. When red mercuric iodide is heated
above 126° it turns yellow, and the red colour is resumed on cooling. Two chemical
changes are involved, because the new substance produced on heating the iodide
re-forms the original compound on cooling. So, when water is heated, complex
aggregates of simple particles are riven asunder to again coalesce or associate to-
gether on cooling. To the physicist, with his attention fixed on the temperature,
or volume, the heating of water is a physical process ; to the chemist, with his
attention on the nature of the constituent particles, it is a chemical process, because
when heated the particles of water become less and less complex as the temperature
rises. What we call a body, said E. Mach, is a complex of properties which affects
the senses in different ways. . . . One or more properties of the complex are altered
in a physical change, while in a chemical change, the whole complex is affected.3
The distinction between chemical and physical changes is a subject for the end,
not the beginning of chemistry. It is remarkable that the first principles of a
science are really the most difficult to grasp, because, said J. F. Ferrier, that
which is first in the order of nature, is last in the order of knowledge :
The apotheosis and final triumph of the human reason will be, when, having traversed
the whole cycle of thought, she returns- — ^enriched only with a deeper insight and clearer
consciousness — to be merged in the glorious innocence of her primitive and inspired
incunabula.
References.
1 Robert Boyle, The Sceptical Chymist, Oxford, 1661 ; Experiments and Observations on
Colours, London, 1663 ; T. Bergmann, De minerarum docimasia humida, Holmise, 1780.
2 M. H. Klaproth, Beitrdge zur chemischen Kenntniss der Mineralkorper, Freiberg, 1795 ; L. N.
Vauquelin, Scherer's Journ., 3. 410, 1799 ; J. J. Berzelius, De Vanalyse des corps inorganiques,
Paris, 1827 ; H, Rose, Handbuch der analytischen Chemie, Berlin, 1829 ; F. Wohler, Praktische
Uebungen in der chemischen Analyse, Gottingen, 1862 ; C. R. Fresenius, A nleitung zur qualitativen
chemischen Analyse, Bonn, 1841.
^ P. V. Wells, Journ. Washington Acad., 9. 361, 1919; I. La,ngmmr, Journ. Anier Chem.Soc.,
39, 1848, 1917 ; L. Gurwitsch, Zeit. phys. Chem., 87. 323, 1914; E. J. Mills, Phil. Mag., (5), 1.
1, 1876; J. F. Ferrier, Institutes of Metaphysics, London, 12, 1854.
§ 6. Compounds and Mixtures
The common operations of chemistry give rise in almost every instance to producta
which bear no resemblance to the material employed. Nothing can be so false as to expect
that the qualities of the elements shall be discoverable, in an altered form, in the com-
pound.'— W. Whewell (1840).
In his De generatione et corruptione, Aristotle regarded the difference between
what we call to-day physical and chemical mixtures, as dependent on the distinction
between what is potential and what is actual. Aristotle recognized a form of com-
bination—now called physical mixture— in which the elements were supposed to
exist actually ; and another— chemical combination— in which the elements were
supposed to exist potentially— e.^. the elements oxygen and hydrogen exist actually
as such in a free state, but "'in water they exist potentially, for they can be educed
and become actual onlv by the destruction of the water or of that special form
which in water they actually possessed.i Consequently it may be said what is
actually one substance may be potentially another. In a mere mixture said Aris-
totle, you have only mixture, juxtaposition or o-uVl^co-i? ; but m chemical combina-
tion you have a mingling or fxC^a where the elements disappear as such, but they
86 INORGANIC AND THEORETICAL CHEMISTRY
still remain potentially. This kind of combination — chemical combination — is
defined very well by Aristotle as " the unification of mingled elements that have
changed their nature as elements."
1. The constituents of a compound are combined in definite proportions. —
The law of constant proportions is of fundamental importance in forming a con-
ception of the meaning of the term " chemical compound." If a substance produced
in different ways be not constant in composition, it is not considered to be a chemical
compound, but rather a mixture. R. Bunsen (1846), for example, showed that the
proportion of oxygen to nitrogen in atmospheric air is not constant, because the
ox}^gen varies from 20'97 to 20'84 per cent, by volume, by methods of measurement
with an error not exceeding 003 per cent. Hence, the oxygen and nitrogen in
atmospheric air are said to be simply mixed together, and not combined chemically.
The so-called eutectic mixtures and cryohydrates show that substances with a definite
composition are not always chemical compounds.
2. Compounds are homogeneous, mixtures are usually heterogeneous. — It is
comparatively easy to detect particles of sugar and sand in a mixture of the two ;
and a simple inspection of a piece of Cornish granite will show it is a mixture of at
least four constituents — silvery flakes of mica ; black patches of schorl ; whitish
crystals of felspar ; and clear glassy crystals of quartz. Although the particles of
felspar, mica, schorl, and quartz differ from one another in size and shape, no
essential difference can be detected in the composition and properties of different
samples of pure quartz, felspar, mica, and schorl. Hence, it is inferred that the
sample of granite is a mixture of schorl, felspar, quartz, and mica ; and that each
of these minerals is a true chemical compound. Very frequently the constituents
of a mixture are too small to be distinguished by simple inspection, and the body
appears homogeneous. A microscopic examination may reveal the heterogeneous
character of the substance. Blood and milk, for instance, appear to be homogeneous
fluids, but under the microscope the former appears as a colourless fluid with red
corpuscles in suspension ; and milk appears as a transparent liquid containing
innumerable white globules (fat). Naturally, too, the stronger the magnification,
the greater the probability of detecting whether the body is homogeneous or not.
Sometimes the microscope fails to detect non-homogeneity under conditions where
other tests indicate heterogeneity.
Before constant composition can be accepted as a proof of chemical combination,
it must also be shown that the substance is homogeneous. Chemical individuals
are homogeneous. A homogeneous substance is one in which every part has
exactly the same composition and properties as every other part. A substance
may have a fixed and constant composition and yet not be homogeneous — e.g.
cryohydrates and eutectic mixtures to be described later. A substance may be
homogeneous, for all we can tell to the contrary, and yet not have a constant
composition — e.g. atmospheric air ; a solution of sugar in water, etc. This simply
means that all chemical compounds are homogeneous, but all homogeneous sub-
stances are not chemical compounds. Indeed, it is sometimes quite impossible
to tell by any single test whether a given substance is a mixture or a true chemical
compound. It is therefore not satisfactory to classify matter into (i) homogeneous
bodies (meaning elements and chemical compounds), and (ii) mixtures, because some
mixtures would have to be included with homogeneous bodies. It might also be
added that the term substance is used in chemistry in two ways : It is employed as
a synonym for body or matter, and also for a specific form of matter which is
chemically homogeneous. 2
3. The constituents of a mixture can usually be separated by mechanical
processes. — The properties of a mixture of finely powdered iron and sulphur have
been used in chemical text-books from the beginning of the nineteenth century in
order to illustrate the difference between mixtures and compounds. It would be
difficult to find a better example. If a mixture containing, say, 6 grams of iron
and 4 grams of sulphur be rubbed in a mortar, (1) the colour of the mixture is
COMBINATION BY WEIGHT 87
intermediate between the colour of the iron and of the sulphur ; (2) the particles of
iron and sulphur can be readily distinguished under the microscope ; (3) most of
the iron can be removed without dijSiculty by means of a magnet ; and (4) the two
constituents can be separated quite readily by washing the mixture on a dry filter
paper by means of carbon disulphide. The sulphur dissolves in the carbon disulphide ;
and the former can be recovered by evaporating the carbon disulphide from the
filtered solution. Sulphur remains behind as a crystalline residue. The metallic
iron remains on the filter paper. Here then the constituents of the mixture have
been separated by the mechanical processes— (1) magnetting, and (2) the action of
solvents.
In 1826, J. J. Berzelius published analyses of the precipitate obtained when hydrogen
sulphide is passed into a slightly acid solution of a salt of tellurous acid, and these showed
that the proportions of sulphur and telluriiun satisfied the law of constant composition,
and hence J. J. Berzelius inferred that a true chemical compound — tellurium sulphide —
was formed. Accordingly, tellurium sulphide — with its method of formation and a de-
scription of its chemical and physical properties — was regularly described in chemical
literature. This sulphide is now considered to be a myth, because half a century lat«r,
F. Becker (1876) discovered that when the material was digested with carbon disulphide,
the sulphur dissolved and tellurium remained imdissolved. Hence it was inferred that
Berzelius' sulphide is not a chemical individual, but a mixture of siilphur and tellurium
in constant proportions. The assumption is of course made that the carbon disulphide
does not decompose the precipitate.
It is generally stated that " a solution of sugar or of salt in water is a mechanical
mixture because, though homogeneous, the salt or sugar can be recovered unchanged
from the water by the mechanical process of evaporation." This is an unwarranted
assumption. The salt and water may have combined, and the product of the
chemical combination may be decomposed into salt and water during the process
of evaporation. The intervention of a solvent sometimes decomposes a compound
into its constituents, or conversely, causes the constituents of a mixture to
combine in such a manner as to produce compounds which previously did not
exist.
The so-called mechanical processes of separation usually include: (1) Magnetting, hand-
picking, sieving, etc. (2) Elutriation, or treatment with water flowing at different speeds
such that the lighter particles are carried off by the slower streams, and the heavier particles
by the faster streams. Settling and lixiviation are modifications of this type of separation.
(3) Flotation, or fractional levigation. If some mixtures be placed in liquids of the right
specific gravity, the lighter constituents will float and the heavier constituents will sink ;
and if some mixtures be treated with oils, etc., the oil so affects the particles of some
substances that they are buoyed up in liquids where otherwise they would sink — such
substances can be separated in this way from other substances not so affected by the oil.
(4) Fractional solution, or crystallization, depend on differences in the solubility of the
constituents in suitable solvents. (5) Distillation, evaporation, freezing, liquation, melting,
diffusion, cupellation, etc.
4. A mixture usually possesses the common specific properties of its consti-
tuents ; the properties of a compound are usually characteristic of itself alone. —
The properties of a mixture are nearly always additive, i.e. the resultant of the
properties of the constituents of the mixture. For instance, a mixture of equal
parts of a white and black powder will be grey, whereas sodium metal and greenish-
yellow chlorine gas give a white pulverulent compound — common salt.
Specific gravity is a number which expresses how much heavier a given substance
is than an equal volume of a standard substance (say water at 4°) taken at a standard
temperature and pressure. In the case of gases, either air=unity, oxygen = 16, hydrogen
= 1, or hydrogen = 2 is taken as standard ; and in the case of liquids and solids, water at
+ 4°, or at 0°, is taken as unity. The great value of specific gravity data lies in the fact
that specific gravity is a number which enables volume meastirements to be convertea into
weights, and weight measurements to be converted into volumes, for weight = specific gravity
X volume. Specific gravity may thus be regarded as the weight of unit volume if water
=unity be taken as a standard, and the weights are reckoned in grams, and volumes in
cubic centimetres. There is no need here to elaborate distinctions between density ana
88 INORGANIC AND THEORETICAL CHEMISTRY
specific gravity. The density is the mass of unit volume, so that if D, m, andt; respectively
denote the density, mass, and volxime of a substance, D =mjv.
The specific gravity of a mixture of equal volumes of two substances of specific
gravity 3 and 5 will be 4, because if one c.c. of water weighs one gram, there will
be a mixture of 05 c.c. weighing I'S gram of one substance ; 0*5 c.c. of the other
substance weighing 2*5 grams ; and l-54-2'5=4 grams per c.c. It must be added
that the specific gravities of compounds are not necessarily a mean of the specific
gravities of their components ; indeed, if elements mix without change in volume
that fact alone is strong presumptive evidence that a compound has not been formed.
It must be added, too, that a small contraction would not be considered a sufficient
proof of chemical action because liquid chlorine and bromine contract a little when
mixed together, and this reaches a maximum — 2 per cent. — when the mixture
corresponds approximately with the atomic proportions Br -f- CI. The specific
gravity of compounds may be greater or less than the average specific gravity of
their constituents. This shows that the force which causes compounds to
unite chemically is not an attractive force independent of the nature of the
combining sub.stances. Hence, although this force is sometimes called chemical
attraction, the term is used metaphorically. Some properties of compounds
— like weight — are additive, for they are the sum of the properties of their
constituents.
Examples. — (1) What is the specific gravity of air containing a mixture of one volume
of nitrogen when the specific gravity of oxygen is 16, and the specific gravity of nitrogen
14-01 ? One-fifth volume of oxygen weighs 3*2 units, and four-fifths volume of nitrogen
weighs 11*2 luiits. Hence, one volume of the mixture will weigh 14*4 units.
(2) Ozonized air- — ^a mixture of air and ozone — has a specific gravity 1-3698, and it
contains 13-84 per cent, by weight of air, specific gravity unity, and 86*16 per cent, of
ozone. What is the specific gravity of ozone ? Here 13-84 grams of air occupy 13-84 -^ 1
volumes ; and 86-16 grams of ozone occupy 86-16-i-ic volumes, where x denotes the specific
gravity of ozone. Hence, 100 grams of ozonized air occupy 100-^1*3698 = 73 volumes;
and 73-00 = (86-16-i-a;)+ 13-84; ora;=l-46.
The law of mixtures may be stated in symbolic form. If a mixture of two
substances contains x fractional parts of a substance of specific gravity ^j, it will
contain 1 — x fractional parts of the other substance of specific gravity Sg- Then
if S be the specific gravity of the mixture, xsi-[-{\—x)s2=8.
Example. — Lord Rayleigh and W. Ramsay (1895) found that a mixture of argon and
nitrogen had a specific gravity 2-3102 (air unity), and the specific gravity of nitrogen alone
is 2-2990 ; what is the specific gravity of argon if the mixture contained 1-04 per cent, of
argon? Here a; = 0-0 104; 1— a;=0-9896; «2=2-2990; >S' = 2-3102. By substituting these data
in the above expression, 2-2990 + (2-3102— 2-2990) -hO-0104=Si, or the specific gravity of
argon (air unity), is Si= 3-376.
If a portion of the mixture of finely divided sulphur and iron be placed in a
hard glass test-tube and warmed over Bunsen's flame, the contents of the tube
begin to glow and a kind of combustion spreads throughout the whole mass. When
cold, break the test-tube, and note that (1) the porous black mass formed during
the action is quite different from the original mixture ; (2) the microscope shows
that the powdered mass is homogeneous ; (3) it is not magnetic like iron (provided
the iron was not in excess); and (4) it gives up no sulphur when digested with carbon
disulphide (provided the sulphur was not in excess). These facts lead to the assump •
tion that there has been a chemical reaction between the sulphur and the iron.
When chemical combination occurs, the reacting constituents appear to lose their
individuality or identity more or less completely, and each neiv substance which is
formed has its own distinctive j)roperlies.
5. Thermal, actinic (light), or electrical phenomena usually occur during
chemical changes. Attention must be directed to the fact that a great deal of heat
was developed during the combustion of the iron and sulphur. The heat required
to start the reaction does not account for the amount of heat developed during the
COMBINATION BY WEIGHT 89
reaction. This point is perhaps better emphasized by placing an intimate mixture
of powdered sulphur and zinc on a stone slab. After the flame of a Bunsen's burner
has been allowed to play on a portion of the mixture for a short time to start the
reaction, the zinc and sulphur combine with almost explosive violence. Large
amounts of heat and light are developed during the reaction.
If a plate of commercial zinc be placed in dilute sulphuric acid, bubbles of gas
are copiously evolved, and if a thermometer be placed in the vessel, the rise of
temperature shows that heat is generated during the chemical action. If the zinc
be pure, very little, if any, gas is developed. It makes no difference if a plate of
platinum be dipped in the same vessel as the zinc, provided the plates are not
allowed to come into contact with one another. If the two plates are connected
by a piece of copper wire, a rapid stream of gas bubbles arise from the surface of
the platinum plate, and some gas also comes from the zinc plate. The platinum
is not attacked by the acid in any way, but the zinc is rapidly dissolved. If a
voltmeter and shunt or an electric bell be interposed in the circuit between the two
plates, the deflection of the needle or the ringing of the bell will show that an electric
current passes from the platinum to the zinc. The electric current is generated by
the chemical reaction between the zinc and the acid, which results in the formation
of zinc sulphate and a gas. The action will continue until all the acid or the zinc
is used up.
For convenience, the zinc plate of the cell B is conventionally called the positive plate
and is often represented by a short thick line, and the platinum plate is likewise called the
negative plate and is represented by a longer thinner line as illustrated by the plan, Fig. 1.
Here G represents the voltmeter or galvanometer and shunt. The vessel of acid with ita
two plates is called a voltaic cell, and this particular combination can be symbolized :
. Platinum | Dilute sulphuric acid | Zinc
The voltaic cell originally used by A. Volta (1800) had copper in place of platinum.
The chemical reaction just indicated is far from being the most economical mode
of generating electricity, but all the different forms of voltaic cell on the market
agree in this : Electricity is generated during chemical action.
The development of heat, light, or electrification are the usual concomitants of
chemical action. The absence of such phenomena when substances are simply
mixed together is usually taken as one sign that chemical action has not taken
place. When nitrogen and oxygen are mixed together in suitable proportions to
make atmospheric air, there is no sign of chemical action, and this fact is sometimes
cited among the proofs that air is a mixture. The argument is not conclusive
because the condensation of steam and the freezing of water are usually cited as
physical changes although heat is evolved during both transformations.
The tests for distinguishing chemical compounds from mixtures involve answers
to the following questions : (1) Is the substance homogeneous ? (2) Are the
different constituents united in definite and constant proportions ? (3) Are the
properties of the substance additive 1 (4) Were thermal, actinic, or electrical
phenomena developed when the substance was compounded 1 (5) Can the con-
stituents be separated by mechanical processes 1 The list does not necessarily
exhaust the available tests, but in spite of what we know, there is sometimes a
fingering doubt whether a particular substance is a mixture or a true chemical
compound. This arises from the fact that some of the tests are impracticable,
others are indecisive. Owing to our ignorance, it is not always easy to state " the
truth and nothing but the truth." As P. J. Hartog 3 has emphasized, C. L. Berthol-
let repeatedly asked J. L. Proust to furnish an experimental distinction between
chemical compounds and mixtures, but without success. Even to-day, there is
no experimental method of generally distinguishing the two. The usual definition
is a theoretical distinction based on molecules, but one can also be adapted from
the phase rule (q.v.).
90 INORGANIC AND THEORETICAL CHEMISTRY
References.
» F. W. Bain, On the Realization of the Possible, London, 171, 1899.
* W. Ostwald, The Fundamental Principles of Chemistry, London, 1909 ; Natural Philosophy,
London, 1911.
» P. J. Hartog, Nature, 50. 149, 1894 ; B. A. Rep., 618, 1894.
§ 7. Circumstantial and Cumulative Evidence
To find the truth is a matter of luck, the full value of which is only realized when we
can prove that what we have found is true. Unfortunately, the certainty of our knowledge
is at so low a level that all we can do is to follow al(^ng the lines of greatest probability. —
J. J. Berzelius.
Suppose a substance is suspected to be a chemical compound because it appears
to be homogeneous ; on investigation, we find that it has a fixed definite com-
position. This verifies our first suspicion, and the joint testimony gives a very
much more probable conclusion than either alone. By piling up the evidence in
this manner, for or against our suspicion, we can make a chain of circumstantial
evidence which enables a highly probable conclusion to be drawn. Each bit of
evidence by itself is not of much value, but all the evidence taken collectively has
tremendous weight. A successful hypothesis is strengthened by the testimony
furnished by diverse facts, and the more numerous and significant the particular
instances embraced by the hypothesis the more nearly will their joint testimony
mount to the altitude of proof, and plausible hypotheses neatly dovetailed may fit
together so well as to apparently strengthen rather than weaken one another. How-
ever, it is easy to see that the probability of an hypothesis being valid becomes less as
the number of unproved assumptions on which it is based becomes greater. We can
even get a numerical illustration. // the definite- compound test be right nine
times out of ten, the probability that a given substance of definite composition
is a true compound is ~ ; similarly, if the homogeneity test be right three times out
of four, the probability that the given homogeneous substance is a chemical com-
pound is I ; and the probability that the given homogeneous substance of definite
composition is a true compound is ||. Every bit of additional evidence in favour
of a conclusion multiplies the probability of its being correct in an emphatic
manner; and evidence against a conclusion acts similarly in the converse
way. Thomas Huxley has stated that one of the tragedies in science is
the slaughter of a beautiful hypothesis by one incongruent fact : a conclusion
based solely upon circumstantial evidence is always in danger of this Damoclean
sword.
A writer has said : " When two facts seem to be in conflict, we may be driven
to decide which is the more credible of the two." This statement may give rise to
a misunderstanding. We cannot admit the possibility of two contradictory facts.
Facts can, and often do, contradict hypotheses. Again, a fact is a fact and cannot
be disputed ; all facts are equally true. Scientific knowledge cannot be arranged
in two compartments, one for truth and one for error. The degree of confidence to
be placed in a statement can be made onlv after the evidence has been sifted and
weighed. If there be any doubt about the truth of an alleged fact, something is
wrong. The laboratory, not the study, is the place to decide if the alleged fact is
the result of an incomplete or of a mal-observation. Facts qua facts cannot be
graded in degrees of probability or credibility, since the difference between
probability and certainty does not represent any quality of the objective fact,
it merely describes a state or attitude of the mind which ranges from ignorance to
knowledge.
I
COMBINATION BY WEIGHT 91
§ 8. Analysis and Synthesis
The earliest chemists were familiar with changes due to the union of distinct
forms of matter to produce a different substance with new properties of its own ;
and also with the separation of two or more definite substances from another quite
different substance. The term spagyric art {(nrdv, to separate ; dyeipetv, to
assemble), applied to chemistry about the sixteenth century, emphasized the fact
that chemical changes were regarded as involving either combinations or decom-
positions ; and as the balance came into more and more extended use, it was
gradually recognized that when elements or compounds have suffered a chemical
change, the original substances can be recovered, qualitatively and quantitatively
the same, by reversing the chemical operation.
The term synthesis — from avv, with ; nOio), I place — is employed for the
operations involved in the formation of a particular compound from its constituents.
The term analysis — from dvd, back ; Avw, I loosen — is employed for the process
of separating the constituents of a compound or mixture. Thus mercuric oxide is
broken down into its constituents when heated. The object of the analysis may be
to answer the question : What are the constituents of the mixture or compound ?
The analysis is then said to be qualitative. If the relative quantities of the different
constituents are to be determined, the analysis is said to be quantitative.
There is one period in the history of chemistry when the discovery or synthesis
of new substances was considered to be the main aim of the chemist ; new sub-
stances were made unmeasured and unclothed with properties, which now re-
quire to be critically scrutinized all over again. The style of some old text-books
on chemistry was not far removed from that of cookery recipe books, for they gave
a long dreary list of modes of preparing different substances which led E. J. Mills
(1876) to say : Chemistry has become an art of breeding (new compounds). The
pioneer workhasbeenuseful, for it has furnished modern chemistry with raw empirical
material to be worked up into science ; indeed a great many more empirical data are
now available than chemists have been able to co-ordinate and assimilate into their
science. Consequently, we are beginning to recognize the truth of the inspired
words of M. W. Lomonossoff, cited above, though written in 1751 ; and the growing
use of tables of measurements and of squared paper in chemical text-books is a
sign of the times. In the words of K. Fittig :
We are now forced to increase the number of compounds, not merely in order to prepare
new substances, but in order to discover natural laws.
The solution which remains when the dilute sulphuric acid can dissolve no more
zinc, may be filtered and evaporated over a hot plate until a drop of the hot solution
crystallizes when placed on a cold glass plate. Crystals of zinc sulphate will separate
as the solution cools. By evaporating a large volume of the solution very slowly,
crystals over a foot long have been obtained. This experiment illustrates the
synthesis of zinc sulphate from metallic zinc and dilute sulphuric acid. The earlier
alchemists assumed that when a metal dissolves in acid, the metal is destroyed,
J. B. van Helmont 1 showed that this assumption is ill-founded because just as
when a certain amount of common salt is dissolved in water, the same amount of
salt can be recovered from the solvent, so, when silver is dissolved in aqua fortis,
the metal passes into solution, but is not essentially altered. In the present case,
the zinc dissolved by the acid can be recovered as zinc sulphate, and if need be as
metallic zinc.
The analysis of aqueous solutions of zinc sulphate by the electric current.--
An electric current is developed during the reaction between dilute sulphuric acid
92
INORGANIC AND THEORETICAL CHEMISTRY
and metallic zinc which results in the formation of zinc sulphate and the evolution
of a gas.
Place two platinum plates, E, Fig. 1, and pure distilled water in the clean glass jar,
which will now be called the " electrolytic cell." Connect the two platinum plates with
an accumulator or secondary battery, and a voltmeter and shunt as indicated in Fig. 1.
The object of the accumulator is to generate an electric current. If the water is pure the
needle of the voltmeter moves very little, if at all. Add a concentrated solution of zinc
sulphate to the water in the glass jar. The jump of the needle of the voltmeter shows
that a current of electricity is flowing through the circuit and hence also through the
solution of zinc sulphate. If chloroform, benzene, or an aqueous solution of cane sugar
had been used in place of the solution of zinc sulphate in the electrolytic cell, no current
would p€tss through the circuit. Hence, liquids may be either conductors or non-conductors
of electricity.
An electric current passing through an aqueous solution of zinc sulphate produces
some remarkable changes : (1) a spongy mass of metallic zinc accumulates about one
of the platinum plates ; (2) if the solution be tested, particularly in the neighbour-
hood of the other platinum plate, sulphuric acid will be found to be accumulating
in the solution during the process of electrolysis ; and (3) bubbles of oxygen gas,
easily tested by collecting some in a test-tube, rise from the same platinum plate
Eleclrodes
Cathode Anode
r>j 7 ^ Platinum
Platinum & ^/P/^fp
^^EleclroW fie Cell
Accumulator ^v_x_v w i/v; iv vyiyufjc^ uNi>-^^ ^
Fig. 1. — Chemical Action induced by Electric Current — Electrolysis.
about which the acid accumulates. If the experiment be continued long enough,
metallic zinc and sulphuric acid will be produced in appreciable quantities. If
the accumulator be disconnected, and the connecting wires be joined together, the
zinc will redissolve in the acid, re-producing zinc sulphate ; and an electric current
will be generated during the dissolution of the zinc.
The process of decomposition or analysis by the aid of the electric current is
called electrolysis. The liquid which is decomposed is called the electrolsrte. The
passing of the electric current through the conducting copper wires, and through
the conducting platinum plates, produces no change in these metals. Hence, we
recognize two kinds of conductivity — in one the conducting medium is decomposed
by the current — electrolyte ; and in the other the conducting medium is not
decomposed by the current — non-electrolyte. The plate at which the zinc collects
is called the cathode— from Kara, down ; 0805, a path— and the other plate, about
which the acid collects, is called the anode— from dm, up ; 080s a path. The
anode and cathode together are called the electrodes. These terms were suggested
to M. Faraday by W. Whewell.2 With the conventions already indicated as to
direction, the electric current is said to enter the electrolytic cell via the anode,
and to leave the cell ma the cathode. The two electrodes are thus " the doors
or ways by which the current passes into or out of the decomposing body." It seems
as if the electric current first splits the decomposing liquid into two parts which
pass to the electrodes. The term anion — from avidv^ that which goes up — is applied
COMBINATION BY WEIGHT 93
to those parts of the decomposing fluid which go to the anode ; those passing to
the cathode are called cations — from KartoV, that which goes down ; and when
reference is made to both the anions and the cations, the term ions- from, Ton',
traveller— is employed. Ion is thus a general term for those bodies which pass
to the electrodes during electrolysis ; or for the two parts, no matter how
complex, into which the electrolyte is primarily divided during electrolysis. This
notation was proposed by M. Faraday in 1834.
The experiments indicated above illustrate an important principle — the principle
of reversibility : If an antecedent event A produces an effect B, then an antecedent
event B will reproduce the effect A. Thus, chemical action can produce an electric
current, and conversely, an electric current can produce chemical action, Fig. 1.
The one can undo the work of the other. Many other examples of the principle will
be recalled — for example, heat causes gases to expand ; conversely, if a gas expands
by its own elastic force, the gas will be cooled ; a crystal of tourmaline is electrified
by uniformly raising its temperature, and Lord Kelvin (1877) showed that the reverse
effect can be induced, for a change of temperature occurs when the electrical state
of the crystal is changed ; etc.
References.
1 J. B. van Helmont, Ortua medicince, Lugduni Batavorum, 1656.
2 I. Todhunter, William Wheivell,D.D.,London, 2, 178, 1876; M. Faraday, Phil. Trans., 124.
77, 1834.
§ 9. Dalton's Law of Multiple Proportions
If Dalton's hypothesis of multiple proportions be found correct, we shall have to regard
it as the greatest advance chemistry has yet made towards its development into a science.
— J. J. Bebzelius (1811).
The formation of chemical compounds is not a capricious and fortuitous process,
but it proceeds in an orderly fashion. Chemical combination is restricted to certain
fixed proportions of matter. These limitations appear to have been prescribed by
nature as part of her scheme in building the material universe. This fact arrested
the attention of J. Rey in 1630. J. Key's conclusion that in the calcination of the
metals " nature has set limits which she does not overstep," agrees with many
facts ; but there are certain limitations. If one gram of lead be calcined for a long
time at 500°, never more than 1'103 gram of a red powder — red lead — is obtained.
Here, 64 grams of oxygen correspond with 621 grams of lead. If the lead be
calcined at about 750°, one gram of lead will not take up more than 0-078 gram of
oxygen to form a yellow powder — litharge; otherwise expressed, 64 grams of oxygen
correspond with 828 grams of lead. Here then nature has set two limits; lead
forms at least two definite oxides— a red oxide stable at a dull red heat, and a
yellow oxide stable at a bright red heat. A puce oxide can also be obtained by
treating the red oxide with nitric acid, and the puce oxide contains 414 grams of
lead for 64 grams of oxygen. The relative proportions of lead and oxygen in the
three oxides are as follows :
Oxygen. Lead.
Puce oxide (lead peroxide) . . 64 414 = 207x2
Red oxide (red lead) ... 64 621=207x3
Yellow oxide (litharge) ... 64 828 = 207x4
This means that for a given weight of oxygen, the yellow oxide has four-thirds as
much lead as the red oxide, and twice as much as the puce oxide. Smularly, carbon
forms two well-defined oxides, called respectively carbon monoxide, and carbon
dioxide. In these we have : ^ ,
Oxygen. Carbon.
Carbon dioxide 8 l-t^l
Carbon monoxide . • • • ^ H-JXZ
94 INORGANIC AND THEORETICAL CHEMISTRY
Perhaps the oxides of nitrogen furnish the most convenient illustration of the
principles ; at least six have been reported (the real existence of the hexoxide has
not been established satisfactorily). In these, the relative proportions of nitrogen
and oxygen are as follows :
Nitrogen.
Oxygeu.
Nitrogen monoxide
14
8 = 8X1
Nitrogen dioxide
14
16 = 8X2
Nitrogen trioxide
14
24 = 8X3
Nitrogen tetroxide
14
32=8X4
Nitrogen pentoxide
14
40 = 8X5
(Nitrogen hexoxide
14
48 = 8X6)
These six compounds of the same elements united in different proportions form
a series of substances so well marked and contra-distinguished that it is questionable
if the most acute human intellect would ever have guessed a priori that they contained
the same constituents. Starting from the compound with the least oxygen, we see
that for every 14 grams of nitrogen, the amount of oxygen increases by steps of 8
grams. Accordingly, in all six compounds of nitrogen and oxygen the masses of
nitrogen and oxygen are to one another as mxl4 : wx8, where m and n are whole
numbers.
If an aqueous solution of sodium hydroxide be mixed with successive small
quantities of hydrochloric acid, the relative proportions of the two substances can
be varied at pleasure, but there is not an infinite variety of compounds of soda and
acid. The one sole product of the reaction is sodium chloride, and this has always
one fixed and definite composition. If an excess of either acid or soda be present,
it is assumed that the excess remains uncombined, because, when the solution is
concentrated by evaporation, crystals of sodium chloride are obtained along with
the excess of soda or of acid if such be present. If sulphuric acid be substituted
for hydrochloric acid, crystals of two distinct and definite products can be separated
— the one is called sodium hisulphate, and the other normal sodium sulphate —
according as the acid or alkali is in excess. Here then is an apparent exception
to the old saw, natura non facit saltum, for nature does make jumps. The leaps
are shown in the relations by weights between the soda and acid in the two products :
Soda.
Acid.
Soda.
Acid
Sodiiun bisulphate
. 52 ,
160
or
52
160
Normal sodium sulphate
. 52
80
or
104
160
Hundreds of cases equally simple might be cited. Similar facts helped to
establish an idea deduced by J. Dalton (1802-4) from the atomic theory, and now
called the law of multiple proportions : when one substance unites with another
in more than one proportion, these different proportions bear a simple ratio to one
another.
There is no difficulty in tracing the simple ratio m: n in the cases which
precede, but it is not always easy to detect the simplicity of this ratio in perhaps
the larger number of cases. Eor instance, the ratio w : ?? for compounds of carbon
and hydrogen passes from 1 : 4 in methane, up to 60 : 122 in dimyrcyl, and still
more complex cases are not uncommon ; the methods of analysis are scarcely
sensitive enough to distinguish the comparatively simple triacontane where carbon :
hydrogen is as 30 : 62, from hentriacontane where this ratio is 31 : 64. Again, the
masses of carbon which unite with one of hydrogen, in methane, ethylene, and acety-
lene are 3, 6, and 12 respectively, but in methane, ethane, propane, hexane, eico-
sane, and anthracene, j;hey are 3, 4, 4-5, 5143, 5-714, and 168 respectively. Several
attempts have been made to get around the difficulty, by rewording the statement
of the law. Thus, B. D. Balaref[ i recommends : " The masses of the different
elements in a compound are directly proportional to their equivalent weights or
to simple multiples of their equivalents," but E. Puxeddu has discussed these
various forms and shown that they are intrinsically different in meaning from the
original Daltonian law.
COMBINATION BY WEIGHT 95
Still the Daltonian law is considered to be so well founded that it can be applied
to predict the composition of compounds which have never been prepared. Thus,
if an oxide of nitrogen containing rather more oxygen than nitrogen hexoxide be
made, it may be predicted that it wilJ contain 7x8=56 parts of oxygen for
every 14 parts of nitrogen by weight. Again, if a substance be found to contain
oxygen and nitrogen, not in the proportion 14 : 8 or a multiple of 8, it is in all
probability a mixture, not a true compound. Again, air contains oxygen and
nitrogen, but the proportions of nitrogen to oxygen is as 14 : 4-29. This is
usually given along with other circumstantial evidence to show the probability that
air is a mixture and not a chemical compound.
Are solutions chemical compounds or mixtures ?— Our definitions say mixtures,
because the composition of solutions follows neither the constant nor the multiple
proportion law. We might easily be led to reason in a vicious circle — in circulo
prohando — by a rigid application of the so-called constant and multiple proportion
laws. Salts dissolve in water in all proportions up to a certain limiting value.
The process of solution, in some cases, seems to be otherwise indistinguishable
from chemical combination, and C. L. Berthollet (1803) 2 considered that " solution
is a true combination " produced by " a feeble combination which does not cause
the characteristic properties of the dissolved body to disappear." It is sometimes
said that the process of solution cannot be a case of chemical combination because
there are no signs of abrupt per saltum changes characteristic of combination in
multiple proportions. The composition of homogeneous solutions can vary con-
tinuously within certain limits while a chemical compound has one fixed and
definite composition ; accordingly, we refuse to call substances compounds which
do not conform with this definition. Hence, in virtue of arbitrarily compiled
definitions, solutions are said to be mixtures, not chemical compounds, and this in
spite of the fact that the dissolution of salts may be accompanied by those very
phenomena which are usually recognized as characterizing chemical combination
— changes in volume, specific heat, temperature, etc. — so that the product of the
reaction (solution) has different properties from the average of its components.
One writer has said : " Efforts have been made to find compounds which do
not conform to the laws of chemical combination, but all attempts have resulted in
failure ; " another writer says, " The law of multiple proportions has been tested by
the analysis of thousands of compounds, and, like the law of constant proportions,
it is one of the perfect laws from which no deviation has been discovered." From
what has been said, if exceptions to the laws of chemical combination were
discovered, chemists would refuse to call them compounds, and the quest for ex-
ceptions must therefore end in failure. For the same reason, the .appeal to ex-
perience is useless, it can neither establish nor refute the laws of constant and
multiple proportions. More bluntly expressed : a prejudice in favour of the defini-
tions in question may warp the judgment to such an extent as to lead to a denial
of the possibility of contradictory phenomena. Such a perversion of the judgment
must be detrimental to the progress of science. Hence the danger of cherishing
a blind faith in our so-called laws of nature, which, at the present day, are little
more than conventional definitions. With such definitions one can easily be deluded
with the belief that he worships in the temple of certainty as indicated in the above
two quotations.
References.
1 D. B. Balareff, Journ. prakt. Chem., (2), 95. 397, 1911 ; E. Puxeddu, f/azz. Chim. ItaL, 49.
i, 203, 1919 ; P. Duhem, Le mixte et la comhinaison chimique, Paris, 1902.
* C. L. Berthollet, E-^sai de statique chimique^ Paris, 1803.
96 INORGANIC AND THEORETICAL CHEMISTRY
§ 10. The History of the Law of Multiple Proportions
Communities of atoms are called clieraical combinations, and they possess every degree
of stability. The existence of some is so precarious that the chemist in his laboratory can
barely retain them for a moment ; others are so stubborn that he can barely break them up.
The more persistent or stable combinations succeed in the struggle for life and are found in
vast quantities as in the cases of common salt and of the combinations of silicon. Stability
is a property of relationship to siu-rounding conditions ; it denotes adaptation to environ-
ment. Thus, salt is adapted to the struggle for existence on earth, but it cannot withstand
the severer conditions which exist on the sun.- — G. H. Darwin (1905).
William Higgius, in his book A comparative view of the phlogistic and
antiphlogistic theories with inductions (London, 1789), stated that one particle of
sulphur and one of oxygen constitute sulphurous acid, while a* particle of sulphur
and two particles of oxygen constitute sulphuric acid ; he also stated that in the
compounds of nitrogen and oxygen, the particles of the two ingredients are to each
other respectively in the ratio 1 : 1 or 2, 3, 4, or 5. According to C. Daubeny (1850),
owing to imperfections in the available chemical analyses, W. Higgins could not
have estabHshed the proposition as a general rule ; and judging from the cursory
manner in which Higgins refers to the relation between the proportions in which
the constituents of these compounds unite to form compounds, he did not attach
much importance to the principle. W. Higgins here appears to have followed
Isaac Newton, who, in his Opticks (London, 1704), said :
The smallest particles of matter may cohere by the strongest attractions and compose
bigger particles of weaker virtue ; and many of these may cohere and compose bigger
particles whose virtue is still weaker, and so on for divers successions, until the progression
ends in the biggest particles on which the operations in chymistry depend.
It has been suggested that Newton's idea of chemical affinity, dependent on the
successive addition of atoms, may have given W. Higgins and J. Dalton the hint
which they needed for producing the law of multiple proportions.
Even before John Dalton enunciated the law of multiple proportions, many
observations had shown that compounds unite together in more than one proportion.
Indeed, it now seems strange that chemists should have failed to notice the law of
multiple proportions when numerous analyses were available. E. von Meyer ^
attributes this to the results being calculated in such a way as to hide the law,
but A. N. Meldrum has shown that the data were frequently stated in precisely
the way required. J. B. Richter (1792) noticed that certain metals have the
power of combining with oxygen to form oxides with two different proportions of
oxygen ; J. L. Proust (1799) obtained a similar result in connection with copper,
but partly owing to inaccurate analyses, and partly owing to the fact that he
had no guiding principle, he failed to recognize the law of multiple proportions.
A. L. Lavoisier (1789) knew that certain substances united with oxygen in several
different proportions each of which corresponded with a fixed and constant relation
between the weights of the combining elements. F. Clement and J. B. Desormes
(1801) also analyzed carbon monoxide and found that it contained just half the
amount of oxygen contained in carbon dioxide, and it afterwards struck J. Dalton as
curious that the two French chemists did not take more notice of this remarkable
result. J. Bostock's analyses of the lead acetates in 1805 were shown by J. Dalton
to be in good agreement with the law. Between 1802 and 1807, J. Dalton gave a
number of examples of the law of multiple proportions from his own analyses and
those of others.
In 1808, in a memoir On oxalic acid, T. Thomson ^ showed that, in the formation
of the two potassium salts of oxalic acid, the quantity of potash which reacts with
a given amount of oxalic acid is in one case j ust double the proportion in the other ;
similar results were obtained with the two strontium oxalates — one of which is
obtained by saturating oxalic acid with strontia water, and the other by mixing
solutions of ammonium oxalate and strontium chloride. It is remarkable, said
COMBINATION BY WEIGHT 97
T. Thomson, that thefirst contains just double the proportion of base contained in
the second. In a paper On swper-acid and sub-acid salts (1808), W. H. Wollaston
^also found that the amounts of carbonic acid relative to a given amount of potash
in the two potassium carbonates are related as 1 : 1 and 1:2. These two papers
are of historical interest, and they attracted some attention because, at that time,
so few facts were known which could be employed to test the law of multiple pro-
portions. In 1810, J. J. Berzelius began to pubHsh a series of investigations designed
" to find the fixed and simple ratios in which the constituents of inorganic nature
are combined ; " he gave a number of accurate analyses which enabled him to say
that if two substances A and B unite in more thap one ratio, the various masses
of A which unite with a fixed mass of B bear a simple ratio to one another. These
experiments played so important a part in establishing the law of multiple proportions
that the law itself has been called Berzelius' law. Some years later, in reviewing
J. Dalton's hypothesis, J. J. Berzelius said :
It may be doubted if J. Dalton was sufl&ciently cautious in applying the new hypothesis
to the system of chemistry. It appeared to me that the paucity of analyses given in support
of the generalization indicated a desire on the part of the experimenter to obtain a certain
result ; but this is just the attitude which must be avoided when proofs for or against a
preconceived theory are sought. Notwithstanding all this, to Dalton belongs the honour
of discovery that part of the doctrine of chemical composition termed the law of multiple
ratios, which no one had previously observed.
In the celebrated Proust v. Berthollet controversy, C. L. Berthollet showed that
some elements unite in more than one proportion, and therefore he argued that
compounds do not necessarily have a fixed and definite composition ; but J. L.
Proust demonstrated that when a metal unites with, say, oxygen in more than one
proportion, the proportion in which the two elements combine do not vary in a
continuous manner, but they proceed in jumps, per saltum, and each of the compounds
has then a fixed and definite composition. J. L. Proust, however, failed to recognize
the law of multiple proportions subsequently developed by J. Dalton.
References.
^ E. von Meyer, History of Chemistry, London, 195, 1906 ; J. L. Proust, Ann. Chim. Phys.,
(1), 28. 214, 1798 ; Journ. Phys., 54. 92, 1802 ; 55. 330, 1802 ; 59. 324, 352, 1804 ; 62. 138,
1806 ; 63. 431, 1806 ; A. N. Meldrum, Mem. Manchester Lit. Phil. Soc., 55. 6, 1911 ; F. Clement
and J. B. Desormes, Gilbert's Ann., 9. 409, 1801 ; J. Bostock, Nicholson's Journ., 11. 75, 1805;
29. 150, 1811 ; A. L. Lavoisier, Traite elementaire de c^imie, Paris, 1789; J. B. Richter, Ueber
die neueren Gegenstdnde der Chymie, Breslau, 1791-1802.
2 T. Thomson, Phil. Trans., 98. 63, 1808 ; W. H. Wollaston, t&., 98. 96, 1808 ; J. J. Berzelius,
GiWerfs Ann., 40. 320, 1812 ; 42. 274, 1812.
§ 11. Richter's Law of Reciprocal Proportions
After long centuries of painful and continuous effort, chemistry has discovered that the
elements combine with one another in definite and unchanging ratios of quantity ; and
that, when their compounds are decomposed, they yield up those identical ratios.— S.
Brown (1843).
Between 1810 to 1812, J. J. BerzeUus i pubHshed the results of a careful study
of the quantitative relations of some of the elements— T'erswc^ die hestimmten wid
einfachen Verhdltnisse aufzufinden nach welcken die Bestandtheile der unorganischen
Natur mit einander verhunden sind. He found that 100 parts of iron, 230 parts of
copper, and 381 parts of lead are equivalent, for they unite with 296 parts of ox>'gen
forming oxides, and with 58-73 parts of sulphur, forming sulphides. Hence, smce
58-73 parts of sulphur and 29-6 parts of oxygen unite respectively with 138 parts
of lead, then, if sulphur and oxygen unite chemically, 58-73 parts of sulphur will
unite with 29-6 parts of oxygen, or, taking the law of multiple proportions into
consideration, with some simple multiple or submultiple of 29' 6 parts of oxygen.
In confirmation, J. J. Berzelius found that in sulphur dioxide, 5873 parts of sulphur
VOL. I. "
98 INORGANIC AND THEORETICAL CHEMISTRY
are united with 57*45 parts of oxygen. The difierence between 2 X 296 = 59*2 and
57 '45 is rather great, but some of the methods of analysis were crude in the time of
J. J. Berzelius, and very much closer approximations — very nearly 1 in 50,000—"
have been obtained in recent years,
J. B. Richter, some twenty years before J. J. Berzelius' work, proved that a similar
relation held good for the combination of acids and alkalies. J. J. Berzelius extended
J. B. Richter's law to combinations between the elements. The above relations are
included in the generalization sometimes called the law of reciprocal proportions,
or the law of equivalent weights. The weights — multiple or submultiple — of the
various elements which react with certain fixed weight of some other element
taken abitrarily as a standard, also react with one another. If each of two sub-
stances, A and B, combines with a third substance C, then A and B can combine
with each other only in those proportions in which they combine with C, or in some
multiple of those proportions. This law does not mean thatif each of the elements
A and B combines with C, then the elements A and B will combine with one another.
A. L. Lavoisier, in his Traite elementaire de chimie (Paris, 1. 116, 1789), argued that
if two elements have une grande appetence for a third element, they should have an
affinity for one another : qucB sunt eadem uni tertio sunt eadem inter se ; and he
added : c'est ce qu'on observe en effet. Further knowledge has shown that the direct
converse is more nearly in accord with facts.
The law of reciprocal proportions may be regarded as a corollary of the law of
multiple proportions on the further assumption that A, B, and C can form binary
compounds — AB, BC, CA — with one another. Consequently it follows that if a
compound be formed by the union of two elements A and B, it is only necessary
to find the proportions in which a third element C unites with one of the two elements,
say A, to be able to predict the proportions in which C will unite with B ; if the law
of reciprocal proportions did not hold, this prediction would be impossible. These
numerical relations come out very clearly by comparing the proportions ii;i which
the difierent members of a series of elements, selected at random, combine with
a constant weight of several other elements. Suppose the analysis of a substance
shows that its ingredients are not in those proportions which we should expect
from the known combinations of each of its components with another substance,
we might safely infer that the substance analyzed is a mixture, and not a single
compound. At ordinary temperatures, alcohol mixes in all proportions with ether
and with water, but ether and water cannot be mixed in all proportions.
Example. — If one gram of hydrogen unites with eight grams of oxygen to form water,
and if one gram of hydrogen iinites with 35"5 grams of chlorine to form hydrogen chloride,
in what proportion will oxygen and chlorine be likely to combine ? Ansr.' — If oxygen
and chlorine unite at all, they will be likely to do so in the proportion of 8 grms. of oxygen
to 35*5 grms. of chlorine, or some multiple or submultiple of this ratio. As a matter of
fact, 8 grms. of oxygen do unite with 35' 5 grms. of chlorine to produce chlorine monoxide.
The laws of constant, multiple, and reciprocal proportions are wonderful examples
of the beauty and harmony of nature ; and yet, we have glimmering hints that these
are but symbols of a sublimer generalization which, when discovered,
Will make one music as before
But vaster.
References.
1 J. J. Berzelius, Gilbert's Ann., 37. 249, 415, 1811 ; 38. 161, 227, 1811 ; 40. 162, 235, 1812 ;
42. 276, 1812 ; Essai sur la theorie des proportions chimiques et sur Vinfluence chimique de V electricity,
Paris, 1819 ; J. B. Richter, Ueber die n^ueren Oegenstdnde der Chymie, Breslau, 1791-1802.
COMBINATION BY WEIGHT 99
§ 12. Combining, Reacting, or Equivalent Weights
Since it is already settled for us by custom that quantities of different substances are
to be called equal when or because they are equivalent gravimetrically, we have no choice
but also, from the chemical point of view, to call those quantities of substance equal which
mteract in single chemical changes. — E. Divers (1902).
The following numbers represent the results obtained by the chemical analysis
of a number of substances selected at random :
Per cent. Per cent.
Silicon dioxide . . . Silicon 46-93 ; Oxygen 63-07
. Hydrogen 2-76 ; Chlorine 97-23
. Magnesium 25-53 ; Chlorine 74*47
. Hydrogen IMS; Oxygen 88-81
. Silver 75-26; Chlorine 24*74
. Silver 70*05; Fluorine 29-95
Hydrogen chloride
Magnesium chloride
Water
Silver chloride
Silver fluoride
Analyses are generally calculated so that the sum of all the constituents is 100
(per cent.) within the limits of experimental error. This is simply a convention of
the analyst, for the results could be just as intelligibly summed to any other number.
Taking any one of the elements as a standard, let us calculate what amount of each
of the other elements will combine with a given quantity of the selected element.
To save time, take oxygen = 8 as the standard. Starting with silicon, 53"07 parts
of oxygen are combined with 46"93 parts of siHcon. Consequently, we have the
proportion 53'07 : 8 = 46*93 : a? ; or, a; = 7'07, for siUcon when the unit oxygen ia 8.
Similarly, for water, hydrogen is 1-008 when oxygen is 8. Again, in hydrogen chloride
when hydrogen is r008, chlorine is 35*4:5 ; in silver chloride, silver is 107*88 when
chlorine is 35-45 ; when silver is 107*88, fluorine is 19*0 ; and when chlorine is 35*45,
magnesium is 12*16. Collecting together the results of these calculations, we get
Oxygen.
Silicon.
Hydrogen.
Chlorine.
Silver.
Fluorine.
Magnesium
8
7-07
1-008
34*45
107-88
19
12*16
We have previously obtained a number of results for some metals for the standard
oxygen 8 by a different process, and the number for magnesium obtained by an
indirect process : Oxygen -> Hydrogen (water) -> chlorine (hydrogen chloride) -»
magnesium (magnesium chloride) gives the same results within the Hmits of experi-
mental error as was obtained by a totally different process. Similar results are
obtained in all cases, subject, of course, to the greater risk of experimental error
when a long chain of compounds is involved. As a rule, there is no need to follow
such an extended series as we have done, for fluorine and for magnesium. Most
of the elements unite directly with oxygen ; and with the other elements, one
intermediate step usually suffices.
We are therefore able to deduce an important generalization : The combining
weights of the elements are specific constants, i.e. they change from element to
element, but for each element, the combining weight is fixed and invariable. Other-
wise expressed: A number can be assigned to each element ; this number — called
the combining, reacting, or equivalent weight— represents the number of parts by
weight of the given element which can enter into combination with 8 parts by
weight of oxygen, or one part by weight of hydrogen. All combining weights are
relative numbers, and they are conventionally referred to oxygen 8, or hydrogen
unity. When an element unites with another element in more than one proportion,
the higher proportions will always be simple multiples of the combining weights— one
for each element. This is the so-called law of combining or reacting weights :
when a substance enters into chemical combination it always does so in quanti-
ties which are proportional to its combining weight ; and the law of multiple
proportions becomes : i The quantities of the different elements in a compound
are simple multiples of their equivalent weights. The term equivalent weight
is generally attributed to W. H. Wollaston (1814), and combining weight to
T. Young (1813).2
100 INORGANIC AND THEORETICAL CHEMISTRY
If the combining weights of the elemeijts are fixed, as they undoubtedly are,
and since the elements can combine to form compounds which, in turn, can form
compounds with other elements and with one another, jt follows that the com-
pounds themselves also have combining weights if they also can enter into chemical
combination. Hence the so-called law of compound proportion — the combining
weight of a compound body is the sum of the combining weights of its components.
This deduction from the Jaw of combining weights is as firmly established experi-
mentally as the law of combining weights itself. The neutralization of acids by
bases, and numerous other chemical reactions, can be cited in illustration.
The experimental results, indicated in § 2, raise the suspicion that there is a
difference between chemical and gravimetric equahty. E. Divers (1902) has
pointed out that in the latter, equal quantities of the different forms of matter are
represented by equal weights ; whereas, in a chemical sense, equal quantities of
matter are the weights or masses of different forms of matter which unite with one
another chemically. Consequently, chemical union may be regarded as a measure
of the amounts of the different forms of matter which are chemically equivalent.
Chemical equality is thus as clearly defined as gravimetric equaUty. The former is
a measure of chemical and the latter a measure of physical phenomena ; the latter
is wholly independent of, and the former mainly dej)endent upon the nature of the
substances compared.
References.
1 D. B. Balareff, Journ. prakL Ghem., (2), 95, 397, 1911.
* W. H. WoUaston, Phil. Trans., 104. 1, 1814 ; T. Young, Introdiiction to Medical Literature,
London, 1813 ; E. Divers, B. A. Rep., 557, 1902.
§ 13. The Perdurability of Matter
The annihilation of matter is unthinkable for the same reason that the creation of matter
is unthinkable, the reason namely that nothing cannot be an object of thought. — H.
Spenckr (1851).
I cannot see what warrant there is for assuming that when a weight A of one substance
combines with another whose weight is B, the weight of the resulting compound is uni-
versally and necessarily A-\-B. — A. D. Risteen (1895).
In 1774, A. L. Lavoisier heated tin with air in a closed vessel and found that
the weight of the whole system, before and after the calcination of the tin, was the
same, thus showing that the whole system neither gained nor lost in weight during
the oxidation of the metal. H. Follinus also noticed, in 1613, that mercur}^ could
be transformed into the sulphide and the product transformed back to the metal
without a change in the weight of the mercury, and Jean Rey was very emphatic,
for he said in 1630 :
I now give a flat denial to the erroneous maxim which has been current since the birth
of philosophy — that the elements mutually undergoing change, one into the other, lose
or gain weight according as in changing they become rarefied or condensed. With the arms
of reason I boldly enter the lists to combat this error, and to sustain that weight is so closely
united to the primary matter of the element that they can never be deprived of it. The
weight with which each portion of matter was endowed at the cradle will be carried by it
to the grave.
I
J. R. Glauber, in his Furni novi philosophici (Amsterdam, 1648), described ti
reaction between a solution of gold in aqua regia and a solution of siUca in potai
lye, by stating :
The potash paralyses the action of the acid with the result that the gold and silica are
respectively deprived of their solvents, and are accordingly precipitated. The weight of
the precipitate so obtained is the sum of the weights of the silica and gold originally taken.
These experiments are here mentioned because they emphasize very well the
fact that, in spite of the most painstaking care, every time all the substances taking
COMBINATION BY WEIGHT 101
part in a chemical reaction are weighed before and after the change, there is no
sign of any alteration in the quantity of matter. The need for assuming the per-
durability or constancy of matter emphasized in the so-called Imv of the indestructi-
bility of matter has been recognized from the very beginning of the Ionian physics ;
for example, Democritus said twenty-four centuries ago : Nothing can ever become
something, nor can something become nothing— eic niUlo nihil fit, et in nihilum nihil
potest reverti. J. B. van Helmont's experiment on the transformation of water into
vegetable substances, and the analytical work indicated in connection with the law
of constant composition, all tacitly assume the principle of the indestructibility of
matter. A. L. Lavoisier is generally supposed to have first demonstrated the law
in 1774 by experiments like that cited above, but the law is very much older ; it was
definitely enunciated in 1756 by M. W. LomonossofE; and the law must have been
at the back of J. Black's mind when he worked on the alkaline earths in 1755.
The alleged demonstrations that " in all changes of a corporeal nature, the total
quantity of matter remains the same, being neither created nor destroyed," illustrate
but do not prove the proposition, and they assume that no new substance can
possibly come into or go out of existence.
The chemist's law of " indestructibility of matter " really means that, in all
cases which have been examined, the total iveight of the elements in any reacting
system remains constant through all the physical and chemical changes it is made
to undergo ; although the observed facts are better generalized as the law of
persistence of weight : no measurable change in the total weight of all the
substances taking part in any chemical process has ever been observed. If A and
B represent respectively the weights of two compounds which form two other
compounds M and N ; and if the symbol = be employed in place of " produces,"
and + for " together with," the law of persistence of weights can be symbolized
algebraically A + B = M+N. If the weight of one of these four compounds be
unknown, it can be computed by solving the equation. Chemists constantly use
this principle in their work, for, as A. L. Lavoisier said in 1774 :
Experiments can be rectified by calculations, and calculations by experiments. I
have often taken advantage of this method in order to correct the first results of my ex-
periments, and to direct me in repeating them with all proper precautions.
When faith in magic was more prevalent than it is to-day, many believed
that by some potent incantation or charm, matter could be called out of nothingness,
or could be made non-existent. i^ Superficial observation might lead to the belief
that a growing tree, the evaporation of water, and the burning of a candle prove
the creation and the destruction of matter, but a careful study of these and in-
numerable other phenomena, has shown that the apparent destruction of matter
is an illusion. Matter may change its state as when liquid water is vaporized, and
when a candle is burnt. In the case of a growing tree, the nutrition the tree receives
from the soil and from the air (carbon dioxide) is overlooked. There is an old
demonstration experiment commonly used to illustrate the fact that the apparent
destruction of matter in the burning of a candle is illusory :
A candle is fixed on one pan of a balance below a cylinder fitted with wire gauze, quick-
lime, soda lime, and glass wool. Weights are added to the right scale-pan until the beam
of the balance is horizontal. The candle is lighted. The gases rising from the flame pass
through the cylinder where the products of combustion are absorbed by the soda lime.
In 3 or 4 minutes the pan carrying the candle is depressed. The increase in weight is due
to the fixation of the products of combustion by the soda lime. The products of com-
bustion are formed by the combination of the carbon and hydrogen of the candle with
oxygen from the air ; this oxygen was not included in the first weighing. The fact illus-
trated by this experiment is undoubtedly true, but the experiment, though popular, is
inconclusive because quicklime and soda lime both absorb moisture and carbon dioxide
from the air. Hence, to make the experiment conclusive, it would be necessary to remove
these compounds from the air used in the burning of the candle, or else to make due allow-
ance for them. This would involve complicated operations ; the test has been made, and
the result is qualitatively the same as with the simpler experiment.
102 INORGANIC AND THEORETICAL CHEMISTRY
Every time a chemical reaction takes place in a closed vessel, which permits
neither the egress nor the ingress of matter, the total weight remains unchanged
within the limits of experimental error. The more carefully the experiments are
made, the more nearly do the values approach identity. Both A, Heydweiller
(1901) and J. J. Manley (1912) have tried to find if a loss in weight occurs during
chemical action, taking the most extreme precautions known to man in order to
secure the utmost accuracy.
The experiment may be illustrated by introducing a solution of silver nitrate into one
limb of the ^-shaped tube by means of a suitable funnel and a solution of potassium
chromate in the other limb. The opening of the tube is then sealed, the tube is weighed
and tilted so as to mix the solutions and start the reaction. The tube is again weighed.
When the reaction is over and the conditions of temperature, etc., are the same as when
the first weighing was made (for illustrative work on the lecture table, the opening of the
tube may be corked and the solutions mixed). Other pairs of solutions are : a solution
of potassiima iodate, slightly acidulated with hydrochloric acid, and potassium iodide ;
lead acetate and sodium sulphide ; acidulated potassium chromate and sodium sulphite ;
etc.
No diiference has been detailed in the weights of the initial and final products
of the reaction within the limits of experimental error — 0*000006 grm. After an
examination of fifteen different reactions, H. Landolt (1909) ^ again failed to detect
a variation in weight ; and added, " since there seems no prospect of pushing the
precision of the experiments further than the degree of exactness attained, the
experimental proof of the law may be regarded as established."
The law of the persistence of weight or the so-called law of the indestructibility
of matter means that a variation in the total weight of the substance taking part in
chemical reactions, greater than the limits of experimental error, has never been
detected. Hence it is inferred that in chemical reactions, substance persists while
matter changes its form. It might also be added that the many and varied deter-
minations of the atomic weights of the elements furnish valuable illustrations of
the law in question. The law of persistence of weight is quite empirical like the law
of excluded perpetual motion. It is shown later, that if a real difference of weight
in the substances taking part in a reaction could be detected, perpetual motion
would be possible.
If immeasurably small and trifling differences be taken into consideration, as
is sometimes done in theoretical speculations, objection might be made to the state-
ment that the weight of a compound must be equal to the weight of the separate
constituent elements, for, as I. Todhunter ^ pointed out in 1876, the converse is the
strict truth. The weight of a body depends upon the positions of the component
particles, and, in general, by altering the positions of the particles, the resultant
effect which we call weight is altered, though it may be to but an inappreciable extent.
Moreover, even the time at which the weighing is performed is theoretically important,
for the weight must change to a trifling extent with the changing position of the sun
and moon in the sky. It is quite conceivable, too, that the weight of the iron in,
say, magnetic oxide of iron might appear to be greater than the same amount of
iron in, say, potassium ferrocyanide because of the effect of the earth's magnetic
field upon the former. But if such an effect were observed, it would not interfere
with our faith in the law as soon as the disturbing effect was recognized.
H. Spencer considers that all the so-called experimental proofs by weighing
tacitly assume the object being proved, since weighing impUes that the matter forming
the weights remains unchanged in quantity ; or as H. S. Redgrove puts it, " weight
measures matter because matter is indestructible, and matter is indestructible
because weight measures matter."
Refeeences.
^ H. Spencer, First Principles, London, 1884.
2 H. Landolt, Zeit. phys. Chem., 55. 589, 1906 ; Ueher die Erhaltung der Masse hei chemischen
Umsetzungen, Halle a. «., 1909 ; A. Heydweiller, Ann. Physik, (4), 5. 394, 1901 ; Lord Rayleigh,
COMBINATION BY WEIGHT 103
Nature, 64. 181, 1901 ; P. Joly, Trans. Roy. Soc. DvJblin, 8. 23, 1903 ; A. W. Surdo, N\U)vo Cimenio,
(5), 8. 45, 1904 ; (5), 12. 299, 1906.
' I. Todhunter, William Whewell, D.D., London, 1876 ; H. S. Redgrove, AkJtemy Ancient and
Modern, London, 1910 ; H. Spencer, First Principles, London, 1884.
§ 14. The Atomic Theory of John Dalton
It seems probable to me, that God in the beginning formed matter in solid, massy, hard,
impenetrable, movable particles, of such sizes and figures, and with such other properties,
and in such proportion to space, as must conduce to the end for which He formed them ;
and that these primitive particles, being solids, are incomparably harder than any porous
body compounded of them, even so hard as never to wear or break in pieces ; no ordinary
power being able to divide what God Himself made one in the first creation. . . . The
changes of corporeal things are to be placed only in the various separations and new associa-
tions and motions of these permanent particles. . . . These principles I consider not as
occult qualities, but as general laws of nature by which the things themselves are formed ;
their truth appearing to us by phenomena, though their causes be not yet discovered. —
Isaac Newton.
The three laws of chemical combinatioD : (1) the law of constant composition ;
(2) the law of multiple proportions ; (3) the law of reciprocal proportions ; and
the law of the persistence of weight, summarize observed facts. They exist quite
independently of any hypothesis we might devise about their inner meaning ; but
we have an intuitive feeling that there must be some peculiarity in the constitution
of matter which will account for the facts.
An atom is the unit of chemical exchange. — Chemists in imagination have
invested matter with a granular structure. Matter is supposed to be discrete,
and built up of corporeal atoms. The imagination can subdivide matter inde-
finitely ; the chemist says that however true this may be, nothing less than an
atom ever takes part in a chemical reaction. The atom is the limiting size so
far as chemical combination is concerned. An atom cannot be subdivided by
any known chemical process. What A. Kekule wrote in 1867 appUes equally
well to-day, in spite of some interesting though abortive attempts to eliminate
atoms from chemistry. Should the progress of chemistry lead to a different view
of the constitution of matter, it will make little alteration to the chemist's atom.
The chemical atom will always remain the chemist's unit. As a chemist, con-
tinued A. Kekule,! the assumption of atoms appears to be not only advisable but
absolutely necessary provided that the term be understood to denote those particles
of matter which undergo no further division in chemical transformations.
Compare this hypothesis with observation. Fix the attention on the facts:
Elements combine with one another either in amounts which correspond with their
combining weights (law of constant composition), or with multiples of their combining
weights (law of multiple proportions). Otherwise expressed, definite amounts of
matter — the atoms — corresponding with the combining weights, act as chemical
units. Keactions between different elements are reactions between these uoits.
Atoms of the same element all have the same constant weight, and atoms of different
elements have different weights. All this is in agreement with the law of constant
combining weights. It is not the mass per se but the constituent particles of the
elements which combine each to each.
Fractions of an atom do not take part in chemical changes.— The proportions
in which one element combines with another can alter only by steps one atom at
a time ; 1, 2, 3, , . . atoms of one element can combine with 1, 2, 3, . . . atoms of
another element. This is but one way of stating the laws of multiple and reciprocal
proportions. The weight of an atom of each element is a constant quantity, and
therefore elements can only combine with each other in certain constant proportions
or in multiples thereof. The atoms of the elements are the units from which nature
has fashioned all the different varieties of matter in the universe. One atom of
mercury unites with one atom of oxygen to form mercuric oxide. If two atoms ot
104 INORGANIC AND THEORETICAL CHEMISTRY
mercury united with one atom of oxygen, the result would not be mercuric oxide,
but some other oxide of mercury — if otherwise, the law of constant composition
would be false. As a matter of fact, such a compound is known, but it is mercurous
oxide. Mercurous oxide has its own specific properties which are different from those
of mercuric oxide.
The analyses of C. F. Wenzel (1777), J. B. Richter (1791) J. L. Proust (1800),
J. Dalton (1801), J. J. Berzelius (1810), and a host of followers are summarized
in the laws of chemical change, and these laws, in turn, are rendered luminous
and coherent by the hypothesis which assumes that all the different forms of matter
in the universe are aggregates of insensibly small homoeomeric particles which all
the powers of chemistry cannot further subdivide. We thus adopt the view of
J. B. Dumas and of M. Faraday that whether matter be atomic or not, this much
is certain, granting it be atomic, it would behave in chemical transformations as
it does now ; A. Kekule expressed similar views in 1867 :
The question whether atoms exist or not has but little significance from a chemical point
of view ; its discussion belongs rather to philosophy. In chemistry we have only to decide
whether the assumption of atoms is a hypothesis adapted to the explanation of chemical
phenomena, . . . and to advance our knowledge of the mechanism of chemical phenomena.
It remains to find the canons by which chemists have been able to fix the relations
between the weights of the atoms of different elements.
Atomic weights are relative. — The combining weights of the atoms can be
expressed in any desired units ; it is quite immaterial whether a grain or a ton be
imagined. In dealing with combining or atomic weights, the conception of abso-
lute quantity is irrelevant. Given sufficient oxygen, 100 tons, kilograms, pounds,
grams, or grains of mercury will give respectively 108 tons, kilograms, pounds,
grams, or grains of mercuric oxide — no more, no less. Several different lines of argu-
ment, given by 0. E. Meyer in his The Kinetic Theory of Gases (London, 1899),
indicate that there are about 1280,000000,000000,000000 or 12*8 X lO^o hydrogen
atoms in a milligram, so that the weight of an atom of hydrogen is not far from
i28o,oooooo.oooooo.ooooooth or 12-8 X 10-20 of a milligram. This estimate may not be
exact, and it is not here emphasized as a fact, although it is probably not far
out. Suppose for the sake of illustration it is true, then, with the evidence so
far adduced, an atom of mercury will weigh 100 X 12"8 X 10" -^tb milligram, and
an atom of oxygen 8 X 12*8 X lO-^o mgrm. We do not know the absolute weights
with any degree of precision, but the relative weights are known with a fair degree
of accuracy. Given the relative weights, and the weight of an atom of one of the
elements, the absolute weights of the atoms of all the other elements can be com-
puted, for the masses of the other elements bear the same ratios to one another
that are assigned to them in the table of atomic weights. The ratio of the weights
of the different kinds of elements in a compound represents the relation between
the weights of the several different kinds of atoms (or aggregates of atoms) which
make up the compound.
J. Dalton's atomic hypothesis. — It is impossible to say who invented the
atomic theory, because it has grown up with chemistry itself. It certainly did not
arise by one effort of modern science, as W. Nernst supposes, " like a phoenix from
the ashes of the old Greek philosophy." In the work of William Higgins and his
predecessors, the hypothesis was little more than an inanimate doctrine. It
remained for Dalton to quicken the dead dogma into a living hypothesis. John
Dalton's atomic hypothesis explains the structure of matter and of chemical com-
bination upon the following postulates, which may be regarded as a very brief
statement of what is called Dalton's atomic theory :
1. Atoms are real discrete particles of matter which cannot be subdivided
by any known chemical process. 2. Atoms of the same element are similar
to one another, and equal in weight. 3. Atoms of different elements have
different properties— weight, afi&nity, etc. 4. Compounds are formed by the
COMBINATION BY WEIGHT 105
union of atoms of different elements in simple numerical proportions —
1:1; 1:2; 2:1; 2:3; etc. This led Dalton to deduce the law of
multiple proportions which was later confirmed by experiments. 5. The com-
bining weights of the elements represent the combining weights of the atoms.
J. Dalton seems to have assumed that the atoms are in perfect repose, unless
disturbed by mechanical or chemical forces.^
Some defects in Dalton's atomic theory.— The hypothesis of Dalton's respecting
atoms, and more particularly atomic weights, is not quite that which prevails in
modern chemistry. According to the atomic theory : an atom is the smallest
particle of an element which can enter into or be expelled from chemical com-
bination. The assumption that the combining weights of the elements represent
the combining weights of the atoms has caused some difficulty. How is the smallest
combining weight of an atom to be fixed 1 In carbon monoxide, for example, we
have oxygen and carbon in the following proportions by weight : Oxygen : carbon
8 : 6, and in carbon dioxide : Oxygen : carbon 8:3 or as 16 : 6. What is the
atomic weight of carbon if the atomic weight of oxygen is 8 ^ Obviously, the
evidence now before us would be consistent with many, different views. Carbon
monoxide may be a compound of one oxygen atom with two carbon atoms each
with a combining weight of 3 ; or a compound of one oxygen atom with one
carbon atom with a combining weight of 6. In the latter case, carbon dioxide
is a compound of one carbon atom of combining weight 6 with two oxygen atoms,
and the same combining weights would have been obtained if any number n of
carbon atoms were combined with 2n oxygen atoms. Again in, order to ascertain
the complexity of a combination of atoms, J. Dalton ^ stated that
If only one combination of two elements exist, it must be presumed to be binary ; if
two combinations exist, one will be a binary compound and the other a ternary compound.
This hypothesis was also adopted by J. J. Berzelius,^ but in the case of the
so-called carbon dioxide or carbon monoxide, there is at present nothing to show
which is the binary and which the ternary compound. Similar difficulties arise
when the idea of atoms so far developed is applied to other combinations of the
elements. There is therefore some confusion. The concept of the atom becomes
more or less indistinct and vague when the attempt is made to develop a
consistent system on the basis of the atomic hypothesis as propounded by
Dalton. Dalton's theory is defective because it lacks a standard for fixing the
atomic weights of the different elements. The difficulty was removed only when
chemists had learned the value of Avogadro's hypothesis in fixing a definite
standard" for evaluating atomic w'eights. Chemists then conventionally came
to an understanding as to the relation between the composition and specific gravity
of a vapour or gas.
References.
1 J. B. A. Dumas, Lt>^on.s sur la philosophie chimique, Paris, 1836 ; A. W. Williamson,
Journ. Chem. Soc, 22. 328, 1869; A. Kekule, Zeit. Chem., (2), 3. 216, 1867; M. M. P. Muir, A
History of Chemical Theories and Laws, New York, 1907 ; 1. Freund, The Study of Chemical
Composition, Cambridge, 1904; M. Faraday, Phil. Mag., (3), 24. 136, 1844.
2 J. Dalton, A Neiv System of Chemical Philosophy, London, 1. 135, 136, 147, 189, 190, 180».
3 J. Dalton, A New System of Chemical Philosophy, London, 1. 214, 1808.
* J. J. Berzelius, Essai sur la tMorie des proportions chimiques et aur Vtnjluencc chimtque de
Velectricite, Paris, 117, 1819.
§ 16. The Evolution of the Atomic Theory up to the time of Dalton
II est regrettable que les traitees modemes negligent I'histoire et pr^sentent comme des
monuments acheves des sciences en perpetuelle Evolution. — F. OsMONr> (190b).
The atomic theory seems to have been born in the twiUght of liistory. The
earhest philosophers of the Eastern fore-world made many quamt guesses at the
106 INOKGANIC AND THEORETICAL CHEMISTRY
constitution of matter. Among these guesses, there is one which appears to have
been promulgated by Kanada as a doctrine among the ancient Hindus i long prior
to the rise of Grecian philosophy. This doctrine assumed that the world of sensible
matter is produced or constituted by the concourse of substantial or concrete
monads or atoms moving more or less freely about one another. A similar guess
was propounded by Leucippus about 450 B.C., and advocated as a doctrine about
thirty years later — 420 B.C. — by his disciple Democritus.2 About 300 B.C. the
same guess was elaborated by Epicurus into a definite system, and the same guess
still lives, more or less modified, in modern chemistry.
From the imperfect fragments which have been transmitted to us, it is scarcely
possible to dissociate the ideas of Leucippus from those of Democritus. Epicurus
taught Democritus' views, which thence passed to Lucretius, and were summarized
in an immortal poem De rerum natura (written about 80 B.C.). According to C. Dau-
beny,3 a Phoenician named Mochus promulgated similar views before Leucippus
time ; and it has also been stated that the ideas of Pythagoras (c. 500 B.C.) about
corpuscular monads, mentioned by Aristotle, in his Metaphysics (12. 6), were derived
from the Egyptian priests. E. Zeller * has argued that the available evidence does
not justify the assumption that Leucippus derived his hypothesis from Mochus,
and he further considers that Democritus adopted nothing but mathematics from
Pythagorean sources, since there is no affinity between the two philosophies. De-
mocritus, however, travelled extensively on his own account ; and it is probable
that he visited the Egyptian priests, the Chaldeans, and the Persians. There are
traces of atomistic views in the writings of Empedocles (c. 500 B.C.), Anaxagoras
(c. 450 B.C.), and Heracleitus (c. 450 B.C.). P. Gomperz ^ has emphasized his belief
that the atomic theory of Leucippus and Democritus was a resultant of the labours
of their predecessors, and that it " was the ripe fruit on the tree of the old doctrine
of matter which has been tended by the Ionian philosophers."
In the fifth century before Christ, Anaxagoras' attempt to compress an inflated
bladder led him to recognize the impenetrability of matter. If matter be con-
tinuous it was not so easy to see how movement without appreciable hindrance
could be possible in air, and yet be impassably resisted by a rock. The atomic
theory of Leucippus provided a satisfactory explanation. Motion in a medium
is easy or difficult according to the disposition of the constituent atoms which makes
it easy or difficult for the atoms to be displaced. The Hellenic theory of atoms
seems also to have been opposed as a counter-proposition to the idea of Zeno (c. 460
B.C.) that matter is infinitely divisible. Zeno argued that whatever be the dimen-
sions of matter, it must be geometrically divisible, for however small a particle may
be, it can be supposed to be halved, quartered, or split into a thousand parts. The
atomicians, however, postulated that the monads or atoms could not be cut, bruised,
broken, or frayed ; otherwise they would wax old, crumble, and lose their shape.
Consequently, substances formed by the aggregation of wearable atoms would
gradually change their characteristics. Water and earth, said Isaac Newton,
composed of old worn particles would not be the same in nature and texture as water
and earth originally composed of unworn particles. There is no reason to suppose
that there has been any change in the character of water and earth in past ages,
and hence, in order that nature may be enduring and permanent, it was inferred
that the atoms must be adamantine and perdurable. Zeno's concept is quite
different from that of the atomicians'. % The latter could have readily admitted with
Zeno that atoms are capable of geometrical subdivision, but reserved the right to
hypothecate that further subdivision does not occur. Consequently, with those
apparently opposing tenets, said S. Brown, the disputants did not argue in answer
to one another at all. They crossed swords without touching one another. Each
fought his own shadow.
Among other names for atoms, Democritus employed &to/xo, but Lucretius does not use
this term. Lucretius' favourite expression is primordia or rerum primordia, which is trans-
lated " the first elements " or " the first beginnings of things." Lucretius also uaes figures,
COMBINATION BY WEIGHT 107
semina, or aetnina rerum — the seeds of things ; materia corpora genitalia or prima ; corpora
or corpora rerum or corpora materia ; elementa ; and corpuacula — but never atom. Cicero
used Democritus' term atomi for these primitive corpuscles. The derivation of the term
atom— a, not ; Te>i/a>, 1 cut — means something which cannot be subdivided. The present-
day definition of an atom says nothing about its ultimate nature. John Dalton certainly
considered the atom to be indivisible, and this is illustrated by his favom-ite aphorism :
*' Thou knowest no man can split an atom." Thomas Graham (1842) defined the atom,
not as a thing which cannot be divided, but as one which has not been divided. The term
atom was once used for a small interval of time — according to Ducange, the ^|th part of
a second— a moment. Thus, in the Greek text of Paul's First Epistle to the Corinthians
15. 52), there is an expression : iv arSfi^, iv piirr) otpdaXnov — in an atom or moment, in the
twinkling of an eye.
Concrete indivisible atoms.— The more characteristic features of the Hellenic
theory of the atomic constitution of matter, as expounded by Lucretius,^ are best
illustrated by quotations from II. A. J. Munro's translation of Lucretius' poem.
1. Matter is discrete, not a continuum.
However long you may hold out by iirging many objections, you must needs in the end
admit that there is a void in things. . . . Wherever there is empty space which we call
void, there body is not. ... If there were no empty void, the universe would be solid . . .
for without void, nothing seems to admit of being crushed in, broken up, or split in two.
2. All substances are formed of soUd atoms which are separated from one
another by void space. Each atom is a distinct individual.
First beginnings are of solid singleness, massed together and cohering closely by means
of least parts, not compounded out of a union of those parts, but, rather, strong in ever-
lasting singleness. . . . First beginnings are strong in solid singleness, and by a denser
combination of these all things can be closely packed and exhibit enduring strength.
3. The atoms are impenetrable, indivisible, and indestructible. They are as
perfect and fresh to-day as when the world was new.
There are therefore certain bodies which can neither be broken in pieces by the stroke
of blows from without, nor have their texture undone by aught piercing to their core, nor
give way to any other kind of assault. . . . Since by the laws of nature it stands decreed
what these things can do and what they cannot do, and since nothing is changed, but all
things are constant . . . they must sure enough have a body of unchangeable matter also.
Therefore, if first bodies are as I have shown solid and without void, they must be everlasting.
. . . For if the first begumings or things could in any way be vanquished and changed it
would then be uncertain too what could and what could not rise into being, in short on what
each thing has its powers defined, its deepest boundary mark. . . . From these parts nature
allows nothing to be torn, nothing further to be worn away, reserving them as the seeds of
things.
4. The atoms differ from one another in shape, size, and weight.
Next in order apprehended of what kind and how widely differing in form are the b^in-
ning of things, how varied by manifold diversities of shape. . . . The things which are able
to affect the senses pleasantly consist of smooth round elements ; while all those, on the
other hand, which are found to be bitter and harsh, are held in connexion by particles that
are more hooked and for this reason are wont to tear open passages in our senses, and on
entering in to break through the body. . . . And quickly as we see wines flow through a
strainer, sluggish oil on the other hand is slow to do so, because sure enough it consists of
elements either larger in size or more hooked and tangled in one another. . . . Again things
which look hard and dense must consist of particles more hooked together, and be held in
imion because compacted throughout with branch-like elements. . . . Those things which
are liquid and of fluid body ought to consist more of smooth and round elements.
5. There is a finite number of different kinds of atoms, but an infinite number
of homoeomeric atoms of each kind.
The first beginnings of things have different shapes, but the number of shapes is finite.
If this were not so, then once more it would f oUow that some seeds must be of mfinite bulk
of body. . . . Wherefore you cannot possibly believe that seeds have an infinite variety of
forms, lest you force some to be of monstrous hugeness. . . . The first begmnings of things
which have a like shape one with another, are infinite in number. For smce the difference
of forms is finite, those which are like must be infinite or the sum of matter will be finite
which I proved not to be the case. ... It is clear then that in any class you like the tirst
beginnings of things are infinite, out of which all supphes are furnished.
108 INORGANIC AND THEORETICAL CHEMISTRY
6. The properties of all substances depend upon the nature of the constituent
atoms, and the way the atoms are arranged. In his Metaphysics, Aristotle illustrated
the effect of shape, arrangement, and position by examples borrowed from the Greek
alphabet, and his illustration may be interpreted : The difference of sha^e is illus-
trated by the opposition of A and N ; the difference of arrangement or contact, by
AN and NA ; and that of 'position^ by the conversion of N to Z by turning the former
on its side.
It often makes a great difference with what things and in what position the same first
beginnings are held in union and what motions they mutually impart and receive ; for the
same make up heaven, sea, lands, rivers, sun ; the same make up com, trees, living beings ;
but they are mixed up with different things and in different ways as they move. Nay, you
see throughout even in these verses of ours many elements conunon to many words, though
you must needs admit that the lines and words differ one from the other both in meaning
and in the sound wherewith they sound. So much can elements effect by a mere change of
order, but those elements are the first beginnings of things can bring with them more
combinations out of which different things can severally be produced.
7. The atoms are in constant motion ; motion is an inherent property of atoms.
Solid bodies of matter fly for ever unvanquished through all time. . , . The first
beginnings of things move of themselves, . . . No rest is given the first bodies through the
imfathomable void, but driven on rather in ceaseless and varied motion they partly, after
they have been pressed together, rebound leaving great spaces between, while in part they
are so dashed away after the stroke as to leave but small spaces between. . . . Herein you
need not wonder at this, that though the first beginnings of things are all in motion, yet
the sun is seen to rest in supreme repose, unless where a thing exhibits motions with its
individual body. For all the nature of first things lies far away from our senses beneath our
ken ; and therefore since they are themselves beyond what you can see, they must with-
draw their motions from sight also ; and the more so that the things which you can see, do
yet often conceal their motions when a great distance off. For often the woolly flocks as
they crop the glad pastures on a hill, creep on whither the grass jewelled with fresh dew
sununons and invites each, and the lambs, fed to the full, gambol and playfully butt ; all
which objects appear to us from a distance to be together and to rest like a white spot on
a green hill.
8. Combination or aggregation is due to the coalescence of moving particles.
Democritus supposed the particles to move in straight lines, and the collisions to
be accidental. In order to better the account for the coalescence, Epicurus supposed
that the atoms moved in paths which deviated sUghtly from the rectilineal.
When bodies are borne downwards sheer through void, at quite uncertain times and
uncertain points of space they swerve a little from their equal poise ; you just and only
just can call it a change of inclination. If they did not swerve, they all would fall
down, like drops of rain, through the deep void, and no clashing would have been begotten,
nor blow produced among the first beginnings ; thus nature never would have produced
aught.
A. A. Cournot ^ believes that none of the ideas bequeathed to us by the ancients
has had a greater or even a similar success to the atomic doctrine of Leucippus and
Democritus. So far as the experimental evidence available to the Grecian philo-
sophers in support of this particular doctrine is concerned, its long life, in the form
of the chemist's atomic theory, can be attributed to chance, for if a sufficient
number of thinkers speculate about the structure of matter, without checking their
conclusions with facts, it is but in accord with the laws of probability that some of
them will approximate to the truth. As C. Daubeny has said :
The earliest philosophers appear to have often lighted upon the most sublime truths,
astonishing us with an intermixture of the noblest views of nature with the most crude and
vulgar conceits, and often leaving to their successors little more than the task of selecting
from the mass of error, the grains of truth which are disguised by and confounded with it.
The modern theory, unlike the older speculation, is based upon the observed laws
of chemical change, and can scarcely stand apart from them.
There is almost an historical continuity in the treatment of the doctrine from
Leucippus to John Dalton (1801) — with a break during the dark ages. The atomism
COMBINATION BY WEIGHT 109
of Deinocritus and Epicurus grew into the corpuscular mechanics of the seventeenth
century, and into the atomic theory of the nineteenth century. Francis Bacon ®
was one of the first of the Renaissance philosophers of the seventeenth century to
recognize the importance of Democritus' doctrine of atoms ; but he later regarded
the study as unprofitable :
Men do not cease from dissecting nature until they arrive at the atom ; a thing which
if true, can do but little for the welfare of mankind.
The atomic hypothesis was accepted with minor modifications by Robert
Boyle,® who said in his Sceptical Chymist (Oxford, 1661) :
There are clusters wherein the particles do not stick so close together, but they may
meet with corpuscles of another denomination, disposed to be more closely united with them
than they were among themselves ; and in such case, two corpuscles thus combining, losing
that shape, size, or motion upon whose account they exhibited such a determinate quality,
each of them really ceases to be a corpuscle of the same denomination as it was before ;
and from the coalition of these, there may result a new body, as really one as either of the
corpuscles before they were confounded.
If this were paraphrased into the language of to-day it would be taken to embody
the idea of a chemical affinity uniting atoms into compounds. Robert Hooke (1665),
John Mayow (1669), Nicolas Lemery (1675), and most of the philosophers of
the Renaissance — R. Descartes (1644), Pierre Gassend (1647), C. Huygens (1690),
G. Amontons (1702), N. de Malebranche (1712), and M. N. LomonossofE (1741)
— were atomicians.^® Rene Descartes seems to have beheved in the existence of
atoms, but he substituted in place of an interatomic void, a subtle imponderable
atomic fluid, the materia coelestis, which occupied the space between the atoms of
matter. Therefore, while a given space could be freed from ponderable matter,
the materia coelestis still remained. This is equivalent to the more modern state-
ment that an aether- vacuum is impossible. N. de Malebranche (1712) dogmatically
asserted that
The matiere subtile or ether ee is necessarily composed of petUs tourhillona- — small vortices
— which are the natural cause of all material changes, and of the most general phenomena
— e.g. hardness, fluidity, weight, buoyancy, the refraction and reflection of light, etc.
Isaac Newton (1675) ^ assumed that the atoms of a compound were held together
by attractive forces so long as they did not approach within a certain limiting
distance ; within this limit repulsive forces were supposed to come into play which
prevented absolute contact and gave rise to the resilience of the particles during
impact. Newton also tried to explain Boyle's law on the assumption that gases
were made up of mutually repulsive particles, which recede from one another as far
as the pressure of the superincumbent atmosphere will let them ; and he referred
chemical changes to different associations of atoms.
R. Kirwan (1783), Bryan Higgins (1776), and William Higgins (1789),i2 with
more or less confidence, explained the constant composition of salts in terms of
atoms. Bryan Higgins recognized seven elements composed of " atoms homo-
geneal, impenetrable, immutable, in figure inconvertible, and globular ; " and he
appears to have held the view that two different atoms combine in the proportions
of 1 : 1, and in that proportion only. William Higgins imagined a combination in
multiple proportions, but believed that the binary combination 1 : 1 was the most
stable. Thus, he said :
In volatile vitriolic acid, a smgle ultimate particle of sulphur is united only to a single
particle of dephlogisticated air ; and in perfect vitriohc acid, eveiy smgle particle of sulphur
is imited to two of dephlogisticated air, being the quantity necessary to saturation.
This idea appears to have arisen in Higgins' mind because it was assumed that
atoms of the same kind are mutually repulsive and that a combmation contaimng
two similar atoms would have a greater tendency to disruption on account of the
110 INORGANIC AND THEORETICAL CHEMISTRY
assumed mutual tendency of similar atoms to break apart. About this time,
W. Nicholson i^ defined chemistry as a science of atoms, for he said :
Chemistry, as a science, teaches the methods of accounting for the changes produced
in bodies by the motions of their parts amongst each other which are too minute to affect
the senses individually ; and, as an art, it consists in the application of bodies to each other
in such situations as are best calculated to produce those changes.
Then followed John Dalton's announcement of the atomic theory and the law
of multiple proportions at a lecture delivered at the Royal Institution, London, in
1803-4 ; the theory was described in T. Thomson's System of Chemistry (Edinburgh,
1807), and by Dalton himself in the following year, in the first part of his remarkable
book, A New System of Chemical Philosophy (Manchester, 1808-10), where he sa} b :
It is one great object of this work to show the importance and advantage of ascertaining
the relative weights of the ultimate particles, both of simple and compound bodies, the
number of simple elementary particles which constitute one compound particle, and the
number of less compound particles which enter into the formation of one or more compoimd
particles.
Quite a number of different suggestions have been made to explain how Dalton
came to give to the atomic hypothesis he had no doubt imbibed from Isaac Newton,
the distinguishing features which led to its being called Dalton's atomic theory.
Dalton's own accounts of the genesis of the hypothesis are not always consistent,
so much so, that H. Debus i* has breathed an improbable suspicion that J. Dalton
dehberately made a mystery of the evolution of the theory. In his System of
Chemistry^ 1807, T. Thomson stated that the theory was suggested to J. Dalton by
a comparison of the analyses of marsh gas and olefiant gas ; but J. Dalton's note-
books show that the experiments on these gases were made in the summer of 1804,
nearly a year after the first table of atomic weights had been compiled. H. E.
Roscoe and A. Harden,i^ in opposition to H. Debus,i6 attempted to prove that Dalton
was influenced in the development of the theory by experiments on the diffusion
and solubiHty of gases, which led him to try to find the relative sizes of the particles
of different gases. This involved a determination of the relative weights of the
particles of each gas, which, in turn, necessitated a determination of the chemical
composition of the gas. The results so obtained led J. Dalton to deduce the atomic
theory. In a series of important papers on The Development of the Atomic Theory
(1909-11), A. N. Meldrum i7 showed that the facts admit of a somewhat different
interpretation.
At the beginning of the nineteenth century, the diffusion of gases was supposed
to be the work of chemical affinity, and the oxygen and nitrogen in the atmosphere
were supposed to be chemically combined. In 1801, J. Dalton ^^ argued that the
phenomenon is physical and that the mixture of oxygen and nitrogen gases in
atmospheric air is mechanical because the " nitric acid gas " formed by the union of
these two elements is "an elastic fluid totally distinct in its properties from either
of the ingredients." Dalton frequently quoted Newton's views on the attraction
and repulsion of atoms, and, in a lecture in 1810, Dalton explained that he did not
at first consider a possible difference in the sizes of the particles of the two elastic
fluids, but he said that in 1805, he considered that the sizes must be different because,
no equilibrium can be established by particles of different sizes pressing against
each other. In Dalton's notebooks, these views are dated Sept. 14th, 1804.
According to H. E. Roscoe and A. Harden, these dates are wrong, for they assume
that, having established a difference in the sizes of the particles of the elastic fluids,
Dalt-on thence proceeded to determine the relative sizes and tveights, together with the
relative numbers of atoms in a given volume. This led the way to the combination of gases.
. . . Thus a train of investigation was laid for determining the number and weight of all
chemical elementary principles which enter into any sort of combination with one another.
Otherwise expressed, it is assumed that Dalton first satisfied himself that the
atoms of different gases have different sizes, and then devised the chemical theory.
COMBINATION BY WEIGHT
111
A. N. Meldrum (1911), however, has shown that J. Dalton did not conclude that the
atoms of different gases were different in size until after the chemical theory had
been formed. In J. Dalton's notebook, dated Sept. 6th, 1803, the first table of
atomic weights appears in the annexed form :
Ult. at. hydrogen
oxygen
,, azote
„ carbon
„ water
„ ammonia.
„ nitrous gas
1
5-66
4
4-5
6-66
5
9-66
Ult. at.
nitrous oxide .
. 13-66
nitric acid
. 15-32
sulphur
. 17
sulphuroTis acid
. 22-66
sulphuric acid .
28-32
carbonic acid .
15-8
oxide of carbon
. 10-2
A. N. Meldrum has also indicated that John Dalton probably arrived at the
law of multiple proportions as a result of experiments on the combination of nitric
oxide and oxygen whereby he was able to write in his notebook, Aug. 4th, 1803,
that 100 measures of air could take 36 or 72 of nitric oxide. J. Dalton then probably
framed the rule that atoms combine in the proportion 1:1, and on considering the
more complex cases, he tested the possibility of combination in other proportions
by the available analytical data, so that, in the following month, Sept. 6th, he was
able to draw up the table of atomic weights.
Punctual atoms or centres of force. — The Lucretian school has never receded
from the primary assumption that matter is composed of ultimate, solid particles —
potentially divisible, but physically incapable of further subdivision ; but another
school of atomicians has assumed that there is no limit to the divisibiUty of the
particles of matter, and that the smallest conceivable particle still consists of an
infinitude of smaller particles. Bene Descartes has said : i^
It is very easy to recognize that there can be in substance no atoms, that is to say parts
of bodies or matter which are by nature indivisible, as some philosophers have imagined ;
in as much as however small we may suppose these parts to be, yet, since they must be
extended, we see there is not one of them that cannot be further divided into two or more
others, of smaller size, and is accordingly divisible ;
and I. Kant (1781), in his Observations on the Second Antinomy ^^^ argued that those
who object to the infinite sub-divisibility of matter do not recognize the clearest
mathematical proofs as propositions relating to the constitution of space. Zeno
(460 B.C.) previously argued that matter must be made up of indivisible and un-
extended points. Some such particles as these — foints de substance — were imagined
by G. W. von Leibniz (1695), and called monads — /xova?, a unit— a term which
was employed by Pythagoras, and which is said to have been suggested to Leibniz
by G. Bruno's De ynonade (Frankfurt, 1591), or during his intercourse with F. M.
van Helmont. Leibniz's ideas were described in his posthumous La monadohgie
(Berlin, 1840) : 21
Material atoms, still composed of parts, are contrary to reason, for the inviolable attach-
ment of one part to another — if we could reasonably conceive or suppose such a thing —
would not destroy their diversity.
Newton himself seems to have had some misgivings about the indivisibility of atoms,
for he said in his Philosophice natiiralis principia mathematica (Londoni, 1687) :
Whether these parts, distinct, and as yet imdivided by material forces, are able to be
divided and sundered in their turn is uncertain.
The main difficulty with Leibniz's animated points is to understand how a body
can possess extension in space if it be made up of components which have no spatial
dimensions, for, as J. C. Maxwell (1877) observes, that which has neither figure nor
extent can have no existence. The Democritians— Newton, etc.— assumed that
it is necessary to suppose that the ultimate particles must possess some bulk, other-
wise they could not produce bulk by aggregation ; on the contrary, Zeno, \V olf ,
Schelling, etc., do not consider this assumption necessary, for a number of self-
repulsive points in limited space can also communicate bulk to the body they
112 INORGANIC AND THEORETICAL CHEMISTRY
compose. For instance, if a point were endowed with the irresistible power of
repelling the hand from a radius of one inch, the result would be the same as if
the hand were to grasp a 2-inch ball of adamant.
R. J. Boscovich,22 in 1763, attempted to improve Leibniz's ideas by assuming
that matter is made up of unextended points which mutually attract one another,
but which never come into contact because, as soon as they approach within a certain
limiting distance, they mutually re-pel one another ; the repulsive forces increase
more and more in intensity as the points approach closer and closer together, so that
they never come into absolute contact. Extension in space is an effect of this
repulsion, and the aggregation of matter is an effect of the attractive forces. He
said:
Matter is not mutually penetrable, but each atom-centre extends, so to say, throughout
the whole of the solar system, yet always retaining its own centre of force.
R. J. Boscovich assumed that when attraction predominates, the body is a solid,
and a gas when repulsion predominates, while if the two forces are more equally
balanced, a liquid results.
A great deal has been written in favour of both hypotheses — Newton's that
an atom is a solid nucleus surrounded by spheres of repulsive and attractive forces ;
and Boscovich' s that an atom is a mathematical point with a sphere of a repulsive
force surrounded by a sphere of an attractive force. In 1844, in A speculation
touching electrical conduction and the nature of matter, M. Faraday 23 points out that
in the ordinary atomic theory it is assumed that solids, liquids, and gases are com-
posed of material atoms occupying a definite space, and are held together by cohesive
forces ; and further, in order to explain the contraction in volume which occurs
on cooling or compressing solids, liquids, or gases, it is assumed that atoms cannot
be in actual contact, but must be separated by an intervening space. These
assumptions involve the following dilemma : If space is a non-conductor of elec-
tricity in non-conducting bodies, and a conductor in conducting bodies, we are
compelled to assume that space possesses opposite and contradictory qualities,
for if space be an insulator, it cannot exist in conducting bodies, and if it be a
conductor, it cannot exist in insulating bodies. Hence, M. Faraday wrote :
I feel a great difficulty in the conception of atoms of matter with intervening spaces not
occupied by the atoms. . . . The atoms of Boscovich appear to me to have a great advantage
over the more usual notion. His atoms are mere centres of forces or powers, not particles
of matter in which the powers themselves reside.
There is a similar dilemma involved in connection with the transmission of light,
and the physicists, A. M. Ampere (1835), A. L. Cauchy (1836), and M. Seguin (1853),
have accordingly regarded atoms as centres of force infinitely small, without
extension in space. Cauchy's punctual atoms were supposed to vibrate differently
in different directions so that the elasticity varied accordingly. J. F. Redten-
bacher (1857) 24 regarded this as an impossible assumption. The atoms, said Ampere,
regarded as les centres d' actions moleciilaires, ne doivent pas etre considerees seulement
comme tres petites relativement aux distances qui les separent, mais conime rigoureuse-
ment nulles.
The difference in the two sets of hypotheses turns on whether cohesive or other
forces emanate from imniaterial points of zero volume, or from material particles
each occupying a definite volume. Which hypothesis is to be accepted ? It must
be remembered that we can persuade ourselves that matter itself can be spirited
away by trying to conceive the residuum which remains when each property known
to be a manifestation of energy is subtracted from matter. An extended nothing,
said G. W. von Leibniz, is meaningless, an extended something must have quality,
and to call that quality extension is to cover up the difficulty with a name. J.
Locke (1690), and G. Berkeley (1713), M. Faraday (1844), W. Ostwald (1892), as well
as earlier and later philosophers, have emphasized how impossible it is to conceive
or imagine the existence of matter independent of energy ; we have evidence of
COMBINATION BY WEIGHT 113
the existence of energy, and therefore, the supposition that a material world reallv
exists apart from energy is undemonstrable and false. The chemist, however,
progresses with his work on the assumption that he lives in a material world which
it is his business to investigate.
The atomic theory is the only satisfactory hypothesis which has correlated the
numerous facts relating to the transformations of matter. It may be perfectly
true. Lord Kelvin (1874) has pointed out, that the assumption of atoms can explain
no property of a body which has not previously been attributed to the atoms
themselves. This, added H. von Helmholtz, is not evidence against the existence
of atoms, but is rather against efforts to derive the foundations of theoretical physics
from purely hypothetical assumptions as to the atomic structure of natural bodies.
The assumption of atoms has none the less proved an invaluable aid in forming
vivid mental pictures of the different phases of a chemical reaction ; it has served
as a wonderful stimulus to the chemical explorer, for it has enabled chemists to
anticipate successfully the results of experimental research. The vitality of this
time-honoured theory is remarkable ; it is ever assimilating new facts, and ever
enticing the chemist to fresh fields and pastures new. Innumerable prophecies
based on the atomic hypothesis have been completely verified so that the atomic
theory is now regarded as a pyramid of truth. Consequently, although no one has
ever seen an atom, A. R. A. Smith (1884) could say : We believe in atoms because,
so far as we can see, nature uses them. The greater the number of facts con-
sistently explained by one and the same theory, the greater the probability of its
being true. The overwhelming mass of circumstantial evidence, direct and in-
direct, which modern chemistry and physics offer, has justified the faith of Dalton ;
and almost, but not quite, demonstrated the real existence of tangible atoms.
References,
^ H. T. Colebrooke, Asiatic Researches of Calcutta, 5. 1, 1799.
2 J. Ferguson, Proc. Phil. Soc. Glasgow, 16. 36, 287, 1884.
^ C. Daubeny, An Introduction to the Atomic Theory, Oxford, 1831.
* E. Zeller, Die Philosophic der Griechen, Leipzig, 1876-82 ; The Pre-Socratic Philosophy,
London, 2. 207, 1881 ; F. A. Lange, Geschichte der Materialismus, Leipzig, 1. 3, 1908; 2. 181,
1908.
5 H. C. Bolton, Amer. Chemist, 3. 326, 1873 ; P. Gomperz, Greek Thinkers, London, 1. 323.
1901 ; S. Brown, Critical Lectures on. the Atomic Theory, Edinburgh, 1843 ; T. Graham, Elements
of Chemistry, London, 1842.
« H. A. J. Munro, T. Lucreti Cari de natura rerum, Cambridge, 1873 — ^the translations in the
text are mainly Munro's ; J. Masson, The Atomic Theory of Lucretius, London, 1884 ; A. Brieger,
Die Urbewegung der Atome und die Weltentstehung bei Leukipp und Demokrit, Halle, 1884 ; H. C.
Liepmann, Die Mechanik der leucipp-democrif schen Atome, Berlin, 1885 ; P. Gomperz, Greek
Thinkers, London, 1. 316, 1901 ; J. Burnet, Early Greek Philosophy, London, 380, 1908; J. C.
Maxwell, Encyc. Brit, 3. 36, 1877 ; 1. Freund, The Study of Chemical Composition, Cambridge, 1904 ;
M. Giua, Gazz. Chim. Ital., 49. ii, 1, 1919 ; J. Gregory, Science Progress, (2), 14. 479, 1920.
' A. A. Cournot, Traite de V enchainement des idees fondamentules datis les sciences et dans
Vhistoire, Paris, 1. 245, 1861 ; C. Daubeny, An Introduction to the Atomic Theory, Oxford, 1831.
8 F. Bacon, De principiis atque originibus, London, 1612 ; Novum Organum, London, 1620.
» R. Boyle, The Usefulness of Experimental Philosophy, Oxford, 1663 ; The Sceptical Chymist,
Oxford, 1661 ; R. Hooke, Micrographia, London, 1665 ; J. Mayow, De sal nitre et spiritu nitro-
aereo, Oxford 1, 669 ; N. Lemery, Cours de chimie, Paris, 1675.
*» P. Gassend, Opera omnia, Florentiae, 1727 ; F. Bemier, Abrege de la philosophic de Gassendi,
Lyons, 1684 ; C. Huygens, Discours de la cause de la pesanteur, Leiden, 1690 ; N. de Malebranchc,
Recherche de la verite, Paris, 1712 ; M. W. Lomonossoff, Elementa chymice mathematica, St. Peters-
burg, 1741 ; Ostwald's Klassiker, 178, 1910 ; R. Descartes, Principia philosophic, Amsterdam,
1644.
11 L Newton, Opticks, Jjon^on, 1704 ; Philosophicenaturalis principia mathematica, C&mhndge,
1687. .^ , J .
12 W. mggins. Comparative View of the Phlogistic and Antiphlogistic Theories with Inductions,
London, 1789 ; B. Higgins, Philosophical Essay concerning Light, London, 1776 ; A. N. Meldrum.
Mem. Proc. Manchester Lit. Phil. Soc., 55. 4, 1910.
i\W. Nicholson, A Dictionary of Chemistry, London, 1795.
1*; H. Debus, Zeit. phys. Chem., 29. 266, 1899.
VOL. T. ^
lU INORGANIC AND THEORETICAL CHEMISTRY
^^ H. E. Roscoe and A. Hai:den, A New View of the Origin of Dalton'a Atomic Theory, London,
1896.
** H. Debus, Ueber einige Fundamentcdsdtze der Chemie inshesondere das Dalton-Avogadrosche
Oesetz, Cassel, 1894; Phil. Mag., (5), 42. 350, 1896; Zeit.phys. Ghent., 20. 359, 1896; 24. 325,
1897 ; 29. 266, 1899 ; 30. 556, 1899 ; G. W. A. Kahlbaum, ih., 29. 700, 1899 ; H. E. Roscoe and
A. Harden, ih., 22. 241, 1897 ; Phil. Mag., (5), 43. 153, 1897.
1' A. N. Meldrum, Mem. Proc. Manchester Lit. Phil. Soc., 55. 5, 6, 1911.
18 J. Dalton, Mem. Proc. Manchester Lit. Phil. Soc., 5. 538, 1802.
i» R. Descartes, (Euvres, Paris, 3. 137, 1824.
20 I. Kant, Kritik der reinen Vernunft, Riga, 1781 ; London, 274, 1860.
21 G. W. von Leibniz, The Monadohgy, Oxford, 1898.
22 R. J. Boscovich, Theoria philosophice nnturalis reducta ad unicam legem virium in natura
existentium, Venetiis, 1763.
2» M. Faraday, Phil. Mag., (3). 24. 136, 1844 ; (3), 27. 345, 1845 ; E. J. Mills, ih., (4), 42. 112,
1871 ; R. Laming, ib., (3), 27. 420, 1845 ; H. Sloggett, ih., (3), 28. 443, 1846 ; W. H. Walenn, ih.,
(4), 39. 123, 1870 ; C. R. A. Wright, ih., (4), 43. 241, 503, 1872 ; R. W. Atkinson, ih., (4), 43. 428,
1872 ; (4), 44. 118, 1872 ; A. Tribe, ih., (4), 44. 121, 1872 ; A. W. Williamson, Jaarn. Chem. Soc, 22.
328, 1869.
2* A. L. Cauchy, Memoire sur la dispersion de la lumiere, Prag, 1836 ; J. F. Redtenbacher,
Das Dynamiden-system, Mannheim, 1857.
§ 16. The Language o£ Chemistry
However certain the facts of any science, however just the ideas derived from these
facts, we can only communicate false or imperfect impressions to others, if we want words
by which these may be properly expressed. — A. L. Lavoisier.
Words are the footsteps of reason. — Francis Bacon.
The nomenclature of a science, that is, the group of technical terms pecuUar
to that science, is of vital importance. It is virtually impossible to separate the
nomenclature from the science itself. Lavoisier emphasized the importance of this
in his classical Traite elementaire de cJiimie (Paris, 1789). Every science consists
of three things : (1) the facts which form the subject-matter ; (2) the ideas repre-
sented by those facts ; and (3) the words in which those ideas are expressed. Like
three impressions of the same seal, said Lavoisier, the word ought to produce the
idea ; and the idea ought to be a picture of the fact.
Special technical words have been invented to fix and describe the ideas and
principles of chemistry — as of all other sciences. Technical terms should be precise
and clear, and not tainted with ambiguity and vagueness. Such technical terms
form part of the current language of chemistry, and they are of international value.
Technical terms are obtained in two ways : (1) Owing to the poverty of language,
words in colloquial every-day use are commandeered, and are given, by a special
definition, a specific meaning. Such words are a proHfic source of error and con-
fusion, and they ofttimes lead to needless controversies because they have a variety
of difierent meanings — energy, force, atom, etc., are examples. (2) Terms are
specially invented for a specific purpose — electron, and telegraph, are examples.
These terms are much less liable to misapprehension than adaptations of every-day
words which possess several meanings. However strange the special terms may
appear at first, they soon grow familiar to the ear, and they can be used without
effort. W. Whewell has pointed out, very aptly, that " technical terms carry the
results of deep and laborious research. They convey the mental treasures of one
period to the generations that follow ; and laden with this, their precious freight,
they sail safely across the gulfs of time in which empires have suffered shipwreck,
and the language of common fife has sunk into oblivion." Witness : some of the
terms used in the chemistry of to-day ware coined by the early Arabian chemists
— e.g. alcohol, alkali, borax, elixir, lac, etc.
Naming the elements. — A great number of the elements have been endowed
with names which refer to some salient property or characteristic, e.g. iodine —
from its violet vapour ; chlorine — from its green colour ; chromium — from the
colour of its compounds ; rhodium — from the rose colour of its salts ; osmium —
COMBINATION BY WEIGHT 115
from its smell ; argorir-hom its indifference to chemical reagents ; similarly with
platinum which refers to the silvery appearance of the metal— from the Spanish
plata, silver. Likewise with the names phosphorus, radium, quicksilver, bromine,
nitrogen, oxygen, hydrogen, argon, glucinum, iridium, praseodymium,' thallium,'
mdium, caesium, and rubidium. Other elements have been named more or less
capriciously ; thus some elements are named after particular localities— sfrow^iwrn,
from Strontian (in Scotland) ; ruthenium, from Ruthenia (Russia) ; yttrium, ytter-
hium, erbium, and terbium are all derived from Ytterby (in Sweden). Some elements
have been named in honour of some country or from association with some other
event at the time of their discovery — e.g. helium, from its occurrence in the sun ;
gallium, from Gallia (Gaul) ; germanium, from Germany ; lutecium, from Leutece, an
old name for Paris ; 'palladium was named after the planetoid Pallas discovered
about the same time : uranium was likewise named in honour of the discovery of
the planet Uranus ; etc. Some names refer to the minerals in which they occur ;
beryllium is derived from the name of the mineral beryl ; zirconium, from the
mineral zircon : similarly with molybdenum, and many others. Some names
refer to renowned personages — e.g. victorium, from Queen Victoria ; similarly
with gadolinium, from J. Gadolin; and mosandrum, after G. Mosander.
Other names refer to mythological personages — e.g. thorium, from Thor,
the son of Odin, a god in Scandinavian mythology ; vanadium, from a Scan-
dinavian goddess, Vanadus ; tantalum, from Tantalus in Grecian mythology ;
niobium, from Niobe, daughter of Tantalus ; i and similarly with cerium, titanium,
palladium, and uranium. Some names are emblematic — e.g. selenium, cobalt,
and nickel.
Unfortunately some elements have not yet been christened with a name recog-
nized by all. Niobium — symbol Nb — and columbium — symbol Cb — are two different
names for one element : glucinum— symbol Gl — and beryllium — symbol Be — are
two different names for another element. There is at present a struggle for exist-
ence between these terms, no doubt the fittest will survive. The first terms here
employed were recommended by the International Association of Chemical
Societies, September, 1913 ; and F. W. Clarke wrote a strong protest, and claimed
columbium in place of niobium for historical reasons.
Symbols. — The old alchemists used to represent different substances by quaint
sometimes fantastic symbols — an example is given in Fig. 1, Cap. I. The hieroglyphs
of the Hermetic priests in Egypt, and the fantastic symbols of the alchemists of the
Middle Ages, were attempts to hide knowledge from the vulgar, and to surround
the study of nature with difficulties and mysteries. The symbols of the modern
chemist, on the contrary, are intended to facilitate the study of chemistr}^ by
abbreviating complicated expressions so that their meaning can be seen at a glance.
Some of the older symbols did come under this category ; for example, gold has
been represented by the picture of a king on his throne ; by the symbol O or Q,
for the sun, etc. ; silver, by (I, the moon ; etc. Fantastic symbols, like that
indicated in Fig. 1, Cap. I, could lead only to confusion. Symbols were employed
by Raymond fully somewhat frequently in the thirteenth century. Possibly the
alchemists intended the symbols to convey some idea of the peculiarities of the metals
they represented ; indeed, it has been suggested that the circle which appears in
certain of the symbols was intended to illustrate the perfection of the metalhc
state, and the half circle, an approximation thereto. In any case the alchemists
were very fond of symbols, and of obscuring their meaning by using mystic triangles
and special hieroglyphs so as to make their writings like cryptograms which required
a key before the meaning could be deciphered. 2 Thus, Raymond Lully in his
Testamentum, duobus libris universam artem chimicam complectens (Colon, 1568), used
the symbol yl to represent God the Creator, 5 stood for mercury,^, for saltpetre. . . .
These symbols were not in general use, and each writer devised his own. The
alchemists of the thirteenth century also represented Aristotle's four elements by
triangles : A, fire ; A, air ; V, water, and V, earth. Other symbols gradually
116 INORGANIC AND THEORETICAL CHEMISTRY
came into more or less general use ; thus, about the fourteenth century the symbol
A for sulphur was fairly common in the writings of the alchemists.
■^ At the beginning of the eighteenth century, symbols for chemical compounds
began to be used more frequently, not with the idea of making the literature
obscure and unintelUgible to the uninitiated, but rather for conciseness, brevity,
and clearness. St. F. Geoffroy, in his Table de differents rapports observes en chimie
entre differerUes substances (1718), used the ordinary alchemical symbols for the
metals and introduced a number of others, e.g. 0 for salt ; >0 for hydrochloric acid ;
>CD for nitric acid ; >0-< for sulphuric acid, etc. In his De attractionibus electiviis
(Upsala, 1775), T. Bergmann represented chemical reactions by symbols and signs.
The two subjoined diagrams illustrate T. Bergmann's method. The symbols to
the right and left, outside the brackets, represent the substances which react together ;
and those above and below, the products of the reaction, if any, which separate from
the system. The symbols within the brackets represent the reacting components ;
and the disposition of the brackets is intended to indicate whether the products of
the reaction are solid, or solution, or volatile. Thus,
Represents the action which occurs Represents the action which occurs
when an aqueous solution (V) of calcium when an alloy of gold and copper (0 $ ) is
sulphide (^4^) is treated with sulphuric a^id fused (A) with antimony sulphide (^);
((^). The lime {^) and sulphuric acid The copper ($) and gold ( 0 ) are separated ;
((^) unite together to form calcium sul- ,, , _ . , , , \x
r: /u^rTL X t,- I, • • -^ * J / X *li® copper ($ and sulphur (^) unite
phate (TUh,) which is precipitated (^-v— ) , ^^ ^ ^' ,. - v +
^ v+ I./ 11 V ; together to form a sohd (-v-), and the
and the sulphur ( A ) also remains as a gold ( 0 ) and antimony ( 5 ) also unite to
soUd. ( -^ ). form a solid ( —^ ).
A. F. de Fourcroy ^ employed a similar method in 1784. It must be added that,
about 1756, W. CuUen is said to have been the first to employ diagrams to illustrate
chemical reactions. A. L. Lavoisier used the symbol v for water ; i^ for oxygen ;
etc., and, like T. Bergmann (1775), he represented chemical reactions by combining
these symbols in various ways.
John Dalton, in his New System of Chemical Philosophy (Manchester, 1808),
made a step in advance by representing the atoms of the elements by symbols,
and combining these symbols so as to show the elements present in a compound.
Thu3, 0 represented hydrogen ; O oxygen ; # carbon, etc. Water was repre-
sented by O0 ; carbon monoxide by 0# ; carbon dioxide by 0#O ; etc. These
symbols have all been abandoned. They are too cumbrous. To-day, we follow
J. J. Berzelius' method, suggested in various editions of his Larbok i Kemien
(Upsala, 1811), and use one or two leading letters from the recognized name of
the element to represent any particular element. The first letter is always a
capital ; the second, if present, is always a small letter.
Thus, O representib oxygen ; H, hydrogen ; C, carbon ; N, nitrogen ; CI, chlorine ; etc.
The names of ten elements start with C, and to prevent the possibility of confusion, a second
leading letter is selected either from the name, or from the alternative Latin name of the
element. Thus C (carbon), Ca (calcium), Cb (columbium), Cd (cadmium), Ce (cerium), CI
(chlorine), Co (cobalt), Cr (chromium), Cs (caesium), and Cu {cuprum)^ copper). The ele-
ments with alternative I^atin names are symbolized : Sb for antimony {stibium) ; Cu for
copper {cuprum) ; Au for gold {aurum) : Fe for iron (ferrum) ; Ag for silver {argentum) ;
Pb for lead {plumbum) ; Hg for mercury {hydrargyrum,) ; K for potassium {kalium) ; Na
for sodium {natrium) ; and Sn for tin (stannum).
Naming the compounds. — Each element forms with other elements a group of
COMBINATION BY WEIGHT
117
compounds which are said to contain the respective elements, because the elements
in question can be obtained unchanged from the compounds. Consequently, every
compound has an elementary or ultimate composition. Compounds are symbolized
by joining together the letters corresponding to the different elements in the
compound. Thus, HgO represents mercury oxide, a compound of mercur}^ and
oxygen. When only two elements are united to form a compound, the name of
the second element is modified so that it ends in ide.
The symbol for the element also represents one of its atoms. If more than
one atom is present in a compound, a small figure is appended to the bottom— in
France, generally at the top right-hand — corner of the symbol of the element, to
indicate the number of atoms present. Thus, HgO represents a molecule of water,
i.e. a compound containing two atoms of hydrogen and one of oxygen ; CO repre-
sents a molecule of carbon monoxide — a compound containing one atom of carbon
and one atom of oxygen ; NagCOs represents a molecule of sodium carbonate — a
compound containing two atoms of sodium, one atom of carbon, and three atoms
of oxygen. A letter affixed in front of a group of symbols represents the number of
times that group is contained in the given compound. Thus, crystalUzed sodium
carbonate is symbolized : NagCOs.lOHgO, meaning that this compound contains
one equivalent of NagCOs, and ten equivalents of the group HgO.
J. J. BerzeHus (1814) * represented two atoms of an element in a compound by drawing
a bar through the symbol of the element ; for instance, HO represented HgO ; ¥^0^ ; FegOg ;
OttO represented CugO ; etc. J. J. Berzelius also represented an atom of oxygen united with
an element by placing a dot over the symbol of the element, and an atom of sulphur by a
dash in a simUar position ; thus, Cu represented CuO ; Pb, PbOg ; Ca(5, CaOCOg ; CuS + ofi
represented CuO, SOg + SHgO ; and Fe represented FeSg. This system did not last long
in chemical literature, although the mineralogists used it for a longer time.
Compounds of an element with oxygen are called oxides, and the process of
combination is called oxidation. When an element forms more than one oxide,
a Greek numerical suffix is often prefixed to the word oxide. Thus SOg is sulphur
dioxide ; SO3, sulphur trioxide ; CO, carbon monoxide ; COg, carbon dioxide ;
PbO, lead monoxide ; PbOg, lead dioxide or lead peroxide. The AngUcized Latin
and Greek numerical prefixes are indicated in Table I.
Table I. — Latin and Greek Numerical Prefixes.
Latin.
Greek.
Latin.
Greel£.
1
Uni-
Mono-
17
Septemdeci-
Heptadeca-
2
Bi-
Di-
18
Duodeviginti-
Octodeca-
3
Ter-
Tri-
19
Undeviginti-
Enneadeca-
4
Quadri-
Tetra-
20
Viginti-
Icosi-
5
Quinqui-
Penta-
21
Unvigiuti-
Henicosi-
6
Sexa-
Hexa-
22
Duoviginti-
Docosi-
7
Septa- :
Hepta-
23
Treviginti-
Tricosi-
8
Octo-
Octo-
24
Quattuorviginti-
Tetracosi-
9
Nova-
Ennea-
26
Quinviginti-
Pentacosi-
10
Deca-
Deca-
26
Seviginti-
Hexacosi-
11
Undeci-
Henadeca-
27
Septemviginti-
Heptacosi-
12
Duodeci-
Dodeca-
28
Duodetriginta-
Octocosi-
13
Terdeci-
Trideca-
29
Undetriginta-
Enneacosi-
14
Quattuordeci-
Tetradeca-
30
Triginta-
Triaconta-
16
Quindeci-
Pentadeca-
31
Untriginta-
Henitriconta-
16
Sedeci-
Hexadeca-
32
Duotriginta-
Dotriconta-
Half Whole Equal Many One and a half One third Four thirds
Semi- Omni- Equi- Multa- Sesqui- Tertia- Quadritertia-
Hemi- Hole- Homo- Poly- Hemitri- Trita- Tetratrita-
It is considered bad style to mix Latin and Greek root words and pre6xes. Conse-
quently we usually try to keep Greek with Greek, and Latin with Latin. Thus, we say
Latin
Greek
118 INORGANIC AND THEORETICAL CHEMISTRY
" diatomic," not " biatomic " ; " bimolecular,'* not " dimolecular " ; " bivalent," not
" divalent " ; and " bivariant," not " divariant " ; because " atomic " is derived from the
Greek word, while " molecular," '* variant," and " valent," are derived from Latin words.^B
There are, however, many hybrids universally recognized. E.g. millimetre, centimetre, etc.^H
Monovalent, divalent, etc., are also used in spite of their hybrid character. In the appli-
cation of the Greek numerals in organic chemistry, some hybrids are \ised — -e.g. in the
methane series of hydrocarbons, Greek numerals are generally employed excepting for
C9H20, C19H40, C29H60, • • • and for C11H24, C21H44, . . . where Latin numerals are used.
The series thus runs pentane, C5H12 ; hexane, C6H14 ; heptane, CjHig ; octane, CJin;
nonane, C9H20 ; decane, C10H22 ; undecano, C11H24; etc. For consistency nonane should
be enneadecane, and undecane, hendecane, etc. The custom is so general, and so deeply
rooted in the literature of organic chemistry, that, as F. Beilstein ^ says, the rectification
gegenwdrlig nicht mehr empfehlenswert erscheint. This state of crystallization has not yet
been attained in the naming of inorganic compounds, and the Greek n\imerical prefixes
can be consistently used if thought desirable ; but " sesqui " is generally used whether Greek
or I-.atin aflfixes are employed. However, we cannot always be purists without defying
custom, which, as Horace has said, decides the language we must use.
Sometimes the termination -ic is affixed to the name of the metal for that oxide
which contains the greater proportion of oxygen, and -OUS for the oxide containing
the lesser proportion of oxygen. For instance, SnO is either stannous oxide, or tin
monoxide, and Sn02 is either stannic oxide or tin dioxide ; FeO is ferrous oxide ;
and Fe203 ferric oxide. For historical reasons, the names of some compounds
do not conform to this system because the affix "ic" was assigned to the compound
first discovered, and the compounds subsequently discovered were named accord-
ingly. Consequently, when only one series of compounds is known, the use of
either termination is now avoided — thus, potassium, sodium, and magnesium are
preferred to potassic, sodic, and magnesic respectively. The method of naming
the compounds now under discussion is not always satisfactory when the elements
form more than two compounds. To get over the difficulty, a prefix hypo- (under,
or lesser) is sometimes added to a compound less rich in oxygen than the -OUS com-
pound, and per-, hyper-, or super- (beyond, above) is added to the one with more
oxygen. Thus,
Persulphuric acid
. H2S2O8
Perchloric acid .
. HCIO4
Sulph\iric acid .
. H2SO4
Chloric acid
. HCIO3
Sulphurous acid
. H2SO3
Chlorous acid
. HCIO2
Hyposulphurous acid
. H2S2O4
Hypochlorous acid
. HCIO
The six nitrogen oxides — nitrogen monoxide, dioxide, trioxide, tetroxide, pentoxide,
and hexoxide — would be awkwardly named by this system.
It will be observed that ous from the Latin osus means " richness," so that stannous
means rich in tin, and etymologically stannous oxide means an oxide richer in tin than
stannic oxide, and by implication poorer in oxygen. In actual use, therefore, the etymological
meaning is inverted, and the implied signification has been universally adopted. Etymolo-
gically the term hypo means less rich, so that hypochlorous means less rich in chlorine than
chlorous — in practice the very opposite is the case, for hypochlorous acid has less oxygen
than chlorous acid, and it contains a higher proportion of chlorine. Similar remarks apply
to the prefixes per^ super, and hyper.
Oxides Hke alumina — ^Al203 ; ferric oxide — Fe203, etc., are sometimes called
sesquioxides {sesqui, one-half more). Compounds which have less oxygen than
the normal are sometimes called suboxides [suh, below) instead of hypo-oxides,
e.g. while CuO represents cupric oxide, CU2O represents cuprous oxide, and also
copper suboxide ; similarly, while AgCl represents the normal silver chloride,
Ag2Cl represents silver sw6chloride. Custom has restricted the use of hypo- to the
acids or acidic oxides, and sub- to the basic or indifferent oxides. The oxides can
be roughly divided into two classes. Some oxides, with water, form acids, and
others act as bases. It is not very easy to draw a sharp line of demarcation between
the two. The acidic oxides have a sour taste, and turn a solution of blue litmus
red ; the basic oxides usually turn a solution of red litmus blue, and have a soapy
feel.
The nomenclature of inorganic chemistry is thus based upon the principle that
COMBINATION BY WEIGHT 119
the different compounds of an element with other elements can be named by a simple
change in the beginning or termination of the word — witness ferric and ferrous
oxides ; and also by the addition of a numerical suffix showing the relative number
of atoms of the corresponding element in its compounds. The systematic name of
a compound thus indicates its composition.^ These little artifices, apparently trivial,
are really important advances in the language of chemistry. The method has
some defects, but when the necessity for a modification becomes acute, it will
probably not be difficult to change. Language generally lags in the wake of
progress.
Ebferences.
1 P. Diergart, Journ. prakt. Chem., (2), 61. 497, 1900 ; J. Berendes, Chem}ztg., 28. 103, 663
1899 ; H. Diels, EUmentum, Leipzig, 1899.
2 G. B. Plowright, Pharm. Journ., 20. 289, 726, 1905 ; 22. 583, 1906; A. L. Lavoisier, Mem.
Acad., 492, 1782.
^ A. F. de Fourcroy, Memoires et observations de chimie, Paris, 308, 1784.
*• J. J. Berzelius, Lehrbuch der Chemie, Dresden, 1827 ; Ann. Phil., 3. 51, 363, 1814.
^ F. Beilstein, Handbuch der organischen Chemie, Berlin, 1. 49, 1918.
* W. Whewell, The Philosophy of the Inductive Sciences, London, 1840.
§ 17. The Evolution o! the Chemist's Nomenclature
For a language to be perfect, it is not sufficient that each substance, each idea, each
modification of form, time, place, etc., should be represented by one word, or by one invari-
able symbol, it is necessary in addition, both to aid the memory and to facilitate the opera-
tions of the mind, that analogous words sho\ild designate analogous substances, analogous
ideas, and modifications of ideas. It is thus that the words of our language represent to
us by similar terminations or augments, similar modifications of ideas represented as
when we say : je vois, j'aperQois, je reQois ; nous voyons, nous apercevons, nous recevons.
In like manner do chemists make use of the expressions sulphate, nitrate, chloride, etc. —
A. Laurent (1854).
In the British Association's Report on Chemical Nomenclature,^ it is shown
that the evolution of the chemist's nomenclature is largely conditioned by the history
of chemistry itself. No attempt to name substances systematically appears to
have been made before the time of Geber — about the thirteenth century. The names
in vogue for chemical substances up to the middle of the eighteenth century were
more or less arbitrary, for they were (i) relics of alchemists' terms — for instance,
aquafortis (nitric acid), aqua regia, etc. ; or derived (ii) from the name of their
discoverer — for instance, Cadefs fuming liquid (alkarsine) ; or (iii) from one who had
made a special study of the substance— for example, Glauber's salt (sodium sulphate) ;
or (iv) from the name of the locality where they occurred — for example, Epsom
salts (magnesium sulphate) ; or (v) from some prominent property or quaUty they
possessed— for instance, tartar emetic (potassium antimony tartrate) ; or (vi) the
names were based upon some superficial resemblance, and thus what J. B. A. Dumas
called le langue des cuisinieres — the language of the kitchen — was applied; for
instance, antimonious chloride was called butter of antimony because of its buttery
appearance ; zinc chloride for the same reason was called butter of zinc ; and arsenic
chloride, hatter of arsenic. On account of this superficial resemblance, these sub-
stances for a time were classed along with butter from milk! Similarly, oi/ of
vitriol (sulphuric acid), oil of tartar (deliquesced potassium carbonate), oUve oil,
and the fatty oils generally were classed together ; so also were such unlike sub-
stances as spirit of wine (alcohol), spirit of salt (hydrochloric acid), Libavius' furmng
spirit (stannic chloride), Boijle's fuming spirit (ammonium sulphide), Glaubers
fuming spirit of nitre (nitric acid), and spirits of hartshorn (ammonia) were mcluded
in one class. ' This virtually means that the names of the compounds were the
basis of the classification. The names were arbitrarily assigned, and hence the
classification was almost as arbitrary and confusing as if the compounds had been
120 INORGANIC AND THEORETICAL CHEMISTRY
classified according to the number of the letters in their names. Liquids were
once called mercurys — mercury itself was mercurius communis, alcohol, mercurius
vegetahilis, etc. Salts were distinguished by their taste — satis acida, salis alcalina
— ^and by their volatility — salis alcalina fixa, salis alcalina volatila, etc. There are
here, however, signs of a feeble attempt at a truer classification.
Towards the end of the seventeenth century, chemists began to assign similar
names to salts having the same origin — more particularly in reference to the acidic
component of the salts. Thus, salts derived from sulphuric acid were called vitriols ;
and those from nitric acid were called saltpetres. A century later, P. J. Macquer
and A. Baume, in their Plan d'un cours de chimie experimentale et raisonnee (Paris,
1757), emphasized the need for designating substances similar in composition by
similar names so as to enable chemists to cope with a rapidly growing list of new
compounds. The confused state of chemical nomenclature, even at the beginning
'of the nineteenth century, is shown by an illustration from Joseph Black's Lectures
on the Elements of Chemistry (Edinburgh, 1803), where sometimes a dozen synonyms
for a salt are listed.
About 1770, T. Bergmann advocated a new system of nomenclature which was
described in his Meditationes de systemate fossilium naturali ; the system was based
as far as possible on the terms then in use, and founded on the phlogiston theory.
T. Bergmann also proposed to represent substances of analogous composition by
similar symbols, and so compounded the symbols that each substance had its own
special symbol. For instance, he called potassium sulphate, alkali vegetahile
vitriolatum ; sodium chloride, alkali fossile salitum ; ammonium nitrate, alkali
volatile nitratum ; and similarly for sodium nitrate, sulphate, etc. His
system was excellent for its time, and shortly afterwards (1782), Guy ton de
Morveau 2 gave a consistent nomenclature for the salts which he described as
compounds of acids and bases, and he illustrated the advantages of his system
by applying it to 474 substances — e.g. vitriol de harote (barium sulphate) ; nitre de
mercure (mercury nitrate) ; muriate de cake (calcium chloride) ; fluor de calce (calcium
fluoride) ; etc. In the choice of names for chemical compounds, said G. de Morveau,
the following five principles should be observed :
(1) A phrase like liqueur alkaline aaturie de la matiere colorante de bleu de Prusse is not
a name and it should be replaced by V alkali prussien. Both terms were then in vogue.
(2) The name should correspond as nearly as possible with the object. When a name
is made up of a noun and an adjective, the former should be applied to the least changeable
and more essential constituent. The names of discoverers should be excluded from the
system. (3) If the constitution of a body is not known, a term with no meaning is better
than one which may ultimately prove to be a wrong one. Hence Valkali prussien is
preferable to Valkali phlogistique. (4) New names are best derived from roots of the best
known dead languages — Greek and Latin. (5) Names should be adapted to the peculiari-
ties of the particular language in which they are to be used.
G. de Morveau's system, like Bergmann's, was founded on the phlogiston theory.
These two schemes were probably the first attempts to devise a complete system
of naming inorganic compounds so that each name indicates the qualitative com-
position of the substance for which it stands. These two systems are not very
different, and are not much unlike the one in use to-day.
In 1787, A. L. Lavoisier and G. de Morveau, with the assistance of C. L. Ber-
tholet and A. F. de Fourcroy, presented details of a new Methode de nomenclature
chimique to V Academic des Sciences 3 in Paris. The proposed method was really an
elaboration of T. Bergmann's and G. de Morveau's systems adapted to the duaHstic
hypothesis. Most chemists felt the need for a precise nomenclature independent
of the phlogiston which the French chemists were rapidly driving out of chemical
science. In the proposed system the names assigned to the various compounds
were intended (i) to indicate the compound ; (ii) to define the compound ; (iii) to
recall its constituent parts ; (iv) to classify it according to its composition ; and
(v) to indicate the relative proportions of its constituents.
I
I
COMBINATION BY WEIGHT 121
The French report laid the foundations of the chemical language of to-day— of
course, after making due allowance for the development of the science which has
necessitated many modifications. The terms ic (ique) and ate, ous (eux) and ite,
for respectively distinguishing the higher and lower acidic oxides and their salts,
are employed for the first time. In 1804, T. Thomson * introduced the plan of dis-
tinguishing the different oxides of an element by prefixing the Greek sufl&xes
proto, first ; deuto, second, ... for the first, second, . . . compound of a series —
e.g. CuCl would be the ^/-o^o-chloride of copper ; and CUCI2 the (^to-chloride. In
1808, J. Dalton explained his notation in his New System of Chemical Philosophy
(Manchester, 1808). J. J. Berzelius' modifications ^ followed in 1811 as indicated
above. Berzelius introduced the term ide, or French ure, as a termination for
simple compounds.
Various other systems of nomenclature have been proposed from time to time in which
artificial words replace the arbitrary names applied to well-known substance* — each vowel
or consonant of the artificial word representing either a substance or a number. ^ These
systems have been found to be unworkable. There are also systems based on M. Dewey's
Decimal Classification and Relativ Index (Boston, 1885) ; for example, A. L. Voge, in his
The Inorganic Compounds (Zurich, 1911), arranges 14,000 inorganic compounds on Dewey's
system. He gives
NgO NO N2O3 NO2 N2O4 N2O5
Symbols . . 133211 133311 133411 13361181 13361182 133611
These systems have possible uses in libraries and for card indexes.
The Methode de nomenclature contained as appendices two Memoir es sur de nouveaux
caracteres a employer en chimie devised by J. H. Hassenfratz and P. A. Adet. In these,
54 straight and curved lines representing the combining units, were arranged in various
v/ays to represent possible compounds. The appearance of the combined symbols, in many
cases, recalls some of the modem systems of shorthand writing. The idea of using " short-
hand systems " is revived every now and again, but has never come into general use.
References.
1 Report on Chemical Nomenclature, B. A. Rep., 39, 1884 ; 262, 1885; H. G. Madan, Joum.
Chem. Soc., 23. 22, 1870.
2 G. de Morveau, Journ. Phys., 19. 310, 382, 1782 ; Ann. Chim. Phys., (1), 25. 205, 1798.
3 Methode de nomenclature chimique proposee par MM. de Morveau, Lavoisier, BerthoUet et
de Faurcroy, Paris, 1787 ; London, 1799.
* T. Thomson, A System of Chemistry, Edinburgh, 1804.
5 J. J. Berzelius, Journ. Phys., 72. 266, 1811 ; 83. 253, 1816 ; L. Gmelin, Handbuch der
anorganischen Chemie, Heidelberg, 1. 149, 1870; A. Laurent, Methode de chimie, Paris, 1854;
J. AIr. Newlands, Chein. News, 4. 281, 332, 1861.
CHAPTEE III
HYDROGEN AND THE COMPOSITION OF WATER
§ 1. The History of Pneumatic Chemistry
The history of human knowledge is a history of false inferences and erroneous inter-
pretations of facts. — ^Max Nordau.
The attention of the early workers in chemistry was mainly directed to visible and
tangible liquids and solids, while the gases — spirits, fumes, vapours, and airs, as
they were variously called — which escaped when different substances reacted
together, were usually considered to be unwholesome effluvia, best avoided. Indeed,
about the middle of the eighteenth century J. Black i could say :
In their distillations, chemists have often observed that part of a body has vanished
from their senses, notwithstanding the utmost care to retain it ; and upon further inquiry,
they have always found that subtle part to be air, which, having been imprisoned in the
body under a sohd form, was set free and rendered fluid and elastic by the fire.
In the third century, Clement of Alexandria beheved that the suffocating
properties of some gases were manifestations of a diabolical nature, and J. B. van
Helmont, who was the most advanced student of gases at the beginning of the
seventeenth century, appears to have had a hazy belief that the gases he had
discovered were in some senses living spirits — diabolic or divine. Even as late as
the middle of the seventeenth century, G. Agricola 2 hinted that the gases in mines
were manifestations of malignant imps ; and the idea had not been altogether
exorcised at the beginning of the eighteenth century.
The old chemists used the term spirit or air where we use the term gas generically for
aeriform elastic fluid. Thus, in the first century of our era, Pliny, in his Historia natura
(2. 4), spoke of that spiritus which both the Greeks and the Romans called aero. The
terms sjnritus, flatus, halitus, aura, and emanatio nubila were also applied to aeriform fluids
disengaged by heating other substances, and they are common in the writings of the
alchemists of the Middle Ages. J. B. van Helmont, in speaking of the spiritum sylvestrem
which he had obtained by the combustion of carbon, etc., said, " This spirit, unknown up
to the present, I call by a new name groa," and he says elsewhere ^ that in order to distinguish
the vapour given off by water at ordinary temperatures from the vapour which is derived
from boiling water, " by the Hcence of a paradox, for want of a name, I call the vapour
rising from water at ordinary temperatures, a gas, being not far severed from the chaos of
the auntients (ancients)." Just as the " chaos of the auntients "■ — Hesiod's xaos — was a
confused mixture of elements from which the Creator produced the universe ; so, to van
Helmont, the vapoiu' of water was a confused mass of elements from which all material
substances could be produced. The word chaos was very frequently used by Paracelsus
with a similar meaning. " Chaos," said he, " is an air like the wind. Air is nothing more
than a chaos. What air is, that is chaos. The element air is named chaos." Stephen
Hales (1727 also said that atmospheric air is a veritable Proteus and a chaos. It is an
easy transition from chaos to chas, which has the so\ind of gas. According to M. Speter,
the ch and ao of chaos when converted into Netherland speech become respectively g and a,
so that van Helmont transformed Paracelsus' term to suit the language of his country.
Some derive the word from the geest — spirit, volatile liquid, or refined fluid — of the Dutch ;
or from the gdscht— yeast — of the Germans.*
Near the beginning of the seventeenth century, J. B. van Helmont, in his essay
De flatihus, distinguished gas sylvestre — ^given off by fermenting liquids — from
the inflammable gases which he named gas pingue, gas sicum, or gas fuliginosum.
J. B. van Helmont seems to have adopted the common opinion that gases are
122
HYDROGEN AND THE COMPOSITION OF WATER 123
different combinations of elastic air with various exhalations or impurities, for at
that time chemists regarded the different gases as chaotic mixtures of various
substances with atmospheric air. The term sylvesire was intended to imply that
the artificial gases which he had prepared were untameable and uncondensable.
In a letter to R. Boyle & in 1678, Isaac Newton stated that he considered that the
ferrous gas (hydrogen) which R. Boyle had obtained by the action of acids on iron,
and the cuprous gas (nitrogen oxide), which C. Huygens 6 had obtained by the action
of nitric acid on copper, contained ultimate particles respectively of iron and copper
brought to a state of aerial elasticity ; but the idea of a, ferrous gas from iron, and a
cuprous gas from copper was disproved when H. Cavendish ^ demonstrated the
identity of the gases obtained by the action of acids on iron and on zinc. According
to J. Priestley, " Boyle ^ was the first who discovered that what we call fixed air,
and also inflammable air, are really elastic fluids capable of being exhibited in a state
unmixed with common air." R. Boyle extended his experiments on factitious
(artificial) airs separable from fixed bodies to a variety of substances, and he noticed
the condensability of hydrogen chloride (1676) ; the orange colour of nitrogen
peroxide (1672) ; and the evolution of an air by heating red lead in the focus of
a burning glass (1678). He also obtained an air from oyster shells and red coral
(1661), and noted the inflammability of hydrogen obtained by the action of acids
upon iron (1671). R. Boyle employed the term air generally (1676) in the same
sense that the word gas is used to-day. Tout corps invisible et impalpable, said
R.Descartes (1664), se nomme air. J. Mayow ^ examined the relative elasticities
of the two gases obtained by R. Boyle by the action of nitric and sulphuric acids on
iron, and decided that there exist various elastic fluids other than air. J. Mayow's
conclusion was opposed by the elder Bernoulli,!^ ^Jio claimed that there are no
other elastic fluids besides air ; and, overlooking the constant diminution of volume
which Mayow found to occur when air is breathed or burnt, J. Bernoulli further
claimed that animals are suffocated and flames are extinguished in certain airs
because the airs are charged with miasmata inimical to life and combustion.
It is sometimes said that S. Hales, in his Vegetable Staticks (London, 1727),
confused the different gases which he prepared with atmospheric air. This erroneous
idea has appeared because Hales focussed his attention on the generic physical
properties of gases rather than on their specific chemical characteristics. Thus,
W. V. Harcourt " has pointed out that when Hales states that " the airs generated
by effervescences . . . resemble true permanent air " he really means that they
are true elastic fluids with the same permanence of constitution, and the same
elastic force as common air. Hales heated a number of substances in vessels
arranged so that the gases evolved could be collected over water, and he measured
the proportion of gas furnished by definite weights- of different substances. He also
collected airs furnished by fermentation processes, and airs generated by the action
of acids on metals. S. Hales did not make any special experiments on the chemical
properties of different gases — hydrocarbons, carbon dioxide, nitrogen oxides,
oxygen, nitrogen, hydrogen, cyanogen, and chlorine— which he probably collected,
nor on the aqueous solutions of the more soluble gases^hydrogen chloride, sulphur
dioxide, and ammonia — which he must have prepared. In spite of the experimental
facts which S. Hales thus accumulated, his attention was so preoccupied with their
generic physical properties that he did not observe their specific chemical differences
— oculos habuit et non videbat — and he was thus prevented from making many
capital discoveries.
J. B. van Helmont seems to have believed that while gases could be prepared
artificially in many ways, they could not be caught and held in vessels— r^a^, vasts
incoercible, foras in aerem prorumpit. S. Hales is generally credited with the
invention of the gas-collecting or pneumatic trough. J. B. van Helmont did not
know how to isolate and preserve the gas sylvestre which he discovered near the
beginning of the seventeenth century, and he distinctly stated that the gas cannot be
confined in any vessel, since it overcomes all obstacles and mixes with atmospheric air.
124
INOKGANIC AND THEORETICAL CHEMISTRY
R. Boyle (1661) and J. Mayow (1669) used a glass globe, Fig. 6, Cap. I, inverted in a
basin of water for confining air ; they filled the globe with water and inverted it in the
basin of water so that the gas generated by the action of an acid on some scraps of
iron in the basin displaced the water and collected in the globe. M. d'Element,
in a brochure 12 pubUshed at Paris in 1719, had abready shown that air could Ih
manipulated and measured like other bodies by confining it in vessels over water ;
and in 1621, J. C. Drebbel had noticed the bubbling of gas from a retort heated
with its beak dipping in water. S. Hales devised the apparatus indicated in Fig. 1 ,
for collecting the gases evolved when different substances are heated in a retort —
a glass vessel was used for generating the gases at low temperatures, and a bent
gim barrel for high temperatures. The vessels used for collecting the gas were hung
by strings mouth downwards below the surface of the water. H. Cavendish (1766)
used a similar device. W. Brownrigg 13 used a shelf with two holes larger than
the gas jar and above the level of the liquid in the trough ; the latter were prevented
sinking too deeply by means of wedges. J. Priestley introduced the use of a per-
forated shelf below the level of the liquid in the trough for supporting the vessel to
be filled with gas. Modifications of S. Hales'
and J. Priestley's pneumatic troughs were
employed very effectively in chemical re-
searches on gases by C. W. Scheele (1770) and
A. L. Lavoisier (1772). Joseph Priestley also
substituted mercury for water ; and, by means
of the mercury pneumatic trough, he collected
and isolated gases — ammonia, hydrogen chlo-
ride, sulphur dioxide, silicon fluoride — which
are so soluble in water that their existence had
been overlooked when water was the confining
liquid.
The study of gases began to occupy serious
attention towards the end of the eighteenth
century, so that in 1779, although " only eight
gases were certainly known with respect to
their composition," yet chemists were so proud
of their knowledge that T. Bergmann was
able to write : " During the last ten years
chemistry has not only soared into regions of invisible aerial substances, but it has
dared to explore the nature of these substances, and to search into their constituent
principles." The nineteenth-century chemists devoted a great deal of time and
attention to the imperceptible, intangible gases ignored by the earlier workers.
Indeed, chemistry could never have progressed very far if the gases and vapours
had been ignored. The work of Joseph Priestley, between 1770 and 1780, gave
such a stimulus to the study of gases that G. Cuvier, in his Eloge historique de
Priestley (Paris, 1806), called him un des feres de la chimie moderne.
Fig. 1.— S. Hales' Pneumatic Trough.
Kefebences.
^ J. Black, Experiments upon Magnesia alba, Quicklimey and other Alcaline Substances,
Edinburgh, 1755 ; Alembic Club Reprints, 1, 1893.
2 G. Agricola, De animantibus svbterraneis. Bale, 1657 ; J. B. van Helmont, Opera omnia,
Franckfurti, 1707.
^ J. B. van Helmont, Oriairike, or Physick Refined, London, 1662 ; Orius medicinal, Amsterdam;
1648.
* G. F. Rodwell, Chem. News, 10. 196, 1864 ; M. Speter, Chem. Ztg., 34. 193, 1910 ; E. von Lipp-
mann, ib., 34. 1, 1910 ; 35. 41, 1911 ; Abhandlungen und Vortrdge, Leipzig, 2. 361, 365, 1913.
^ R. Boyle, Works edited by Thomas Birch, London, 1744.
« C. Huygens, Phil. Trans., 10. 443, 1675.
' H. Cavendish, Phil. Trans., 55. 141, 1766.
' R. Boyle, Physico-mechanical experiments to show the spring and effects o/air, London, 1661 ;
HroKOGEN AND THE COMPOSITION OF WATER 125
New experiments touching the rdationbetween flame and air, London, 1671 ; Phil. Trans., iO. 1675 ;
Second continuation of new experiments, physico-mechanical, touching the spring and weight of air,
London, -1676.
^ J. Mayow, Tractatus de parte aerea igneaque spiritus nitri, Oxford, 1669.
^» J.'BeTno\illi,Dissertatiodeeffervescentiaetfermentationen^vahypothesifundata,Base\ 20 1670
" W. V. Harcourt, Phil. Mag. (3), 28. 106, 478, 1846.
^2 M. d'Element, La maniere de rendre Vair visible, Paris, 1719 ; J. C. Drebbel, Een kort tractaei
van de natuere der elementen, Rotterdam, 1621 — German edition, 1624.
13 W. Brownrigg, Phil. Trans., 55. 235, 1765.
§ 2. Hydrogen— Preparation and Properties
It can scarcely be said that pneumatic chemistry was properly begun till Mr.
Cavendish published his valuable paper on carbonic acid and hydrogen gas, in the year
1766.— T. Thomson (1813).
The discovery of hydrogen.— It is inconceivable that the alchemists knew
nothing about this gas, for they were perpetually operating with various metals
in contact with acids. It must therefore have been known for a very long time
that an inflammable air or gas is produced when iron is dissolved in dilute sulphuric
acid. Paracelsus, in the sixteenth century, described the action somewhat
quaintly. He said that when the acid acts on iron " an air arises which bursts
forth like the wind." Near the beginning of the next century, J. B. van Helmont
described this gas as a peculiar variety of air which was combustible and a non-
supporter of combustion, but his ideas were somewhat hazy, for he confused it with
other inflammable gases ; indeed, up to about 1766, writers generally used
inflammable air as a general term to include this gas, as well as the hydrocarbons,
hydrogen sulphide, carbon monoxide, and other combustible gases. Hydrogen
was sometimes specifically distinguished as the inflammable air from the metals.
In 1650, T. Turquet de May erne i reported that the fumes evolved when dilute
oil of vitriol acts on iron are inflammable, and in 1671 Boyle 2 observed that the
flame was extinguished when placed under the receiver of an air-pump, but K.
Boyle's chief concern was to show that the gas, which he called the volatile sulphur
of Mars, was dilatable and compressible, and that it was really an air. Nearly a
century later, J. Priestley also experimented with the gas, and S. Hales ^ found that
iron filings and oil of vitriol gave scarcely any air, but on adding water, there was a
copious evolution of the aeriform fluid. In 1766, H. Cavendish * showed that the
combustible gas produced by the action of dilute sulphuric or hydrochloric acid on
metals like iron, zinc, and tin is a distinct substance with definite properties pecuHar
to itself ; hence, hydrogen was called inflammable air. Cavendish measured the
amount of hydrogen obtained from a given weight of the different metals ; he also
measured the specific gravity of the gas, and found it to be seven times lighter than
atmospheric air ; he also showed that the specific gravity of the gas was the same
whether zinc or iron were used in the preparation. F. de Lassone and C. W.
Scheele discovered almost simultaneously that a solution of zinc in caustic lye
furnishes the same gas. J. Watt (1783),' R. Kirwan (1781), H. Cavendish (1766),
and J. Priestley (1784) identified the gas with the evanescent phlogiston, and
they called it phlogiston, or phlogistic ated air; but neither this name nor
inflammable air persisted very long, for both terms were ousted by the cognomen
hydrogen which A. L. Lavoisier applied to the gas in 1783. In his Considerations
generales sur la dissolution des metaux dans les acides (1784),^ A. L. Lavoisier, follow-
ing a suggestion of P. S. de Laplace, traced the source of the hydrogen which is
evolved when a metal dissolves in a dilute acid, to the decomposition of the water.
He assumed that the oxygen of the water united with the metal to form a calx,
and the hydrogen escaped in the free state. The calx united with the acid to form
water and a salt.
The preparation of hydrogen. — Hydrogen obtained by the action of dilute
126 INORGANIC AND THEORETICAL CHEMISTRY
sulphuric or hydrochloric acid on metallic iron is not very pure, and it possesses
a distinct smell owing to the presence of hydrocarbon gases, etc., formed by the
action of the acid on the carbon compounds associated, as impurities, with com-
mercial iron. The solution remaining after the action of sulphuric acid on the
iron, when put aside in a cool place, soon forms beautiful pale green crystals of
ferrous sulphate. Magnesium and aluminium furnish a fairly pure gas ; with
aluminium the acid should be warmed to start the reaction. In these cases, not
only is hydrogen gas evolved but crystals of magnesium sulphate and of aluminium
sulphate can be obtained from the liquids in which the respective metals have been
dissolved. The action of the acid on tin is rather slow ; granulated zinc is used
for general laboratory work.
Hydrogen gas is made in small quantities in the laboratory by placing granulated
zinc in a bottle fitted with a stopper with two holes — one to take a funnel tube, the
other to take an L-shaped tube for conducting away the gas. Instead, the
granulated zinc may be placed in a two-necked Woulfe's bottle — so named because
these bottles were first described by Peter Woulfe (1784). The one tubulure is fitted
with a one-hole stopper carrying a tube funnel, and the other, with the gas exit tube.
The zinc is covered with water, and sulphuric acid is added a little at a time through
the tube funnel until the gas begins to come off vigorously. For many purposes
there is no need to use the pneumatic trough for collecting hydrogen, since by
bringing the gas- jar mouth downwards over a jet of hydrogen the gas will collect
in the upper part of the jar, and displace the air downwards — hence the term collect-
ing gases by the downward displacement of air — many writers call this collecting
the gas hy wpward displacement. Hydrogen gas so prepared is always tested before
iLse by collecting a test-tube of the gas, and while holding the tube upside down,
applying a lighted taper. If the gas burns quietly at the mouth of the test-tube,
all is well.
Hundreds of different forms of apparatus ^ have been devised for supplying an
intermittent stream of gas by the action of a liquid — e.g. hydrochloric or sulphuric
acid— on a solid — e.g. zinc, ferrous sulphide, or marble. They are all based on the
principle applied by J. W. Dobereiner in his hydrogen lamp. When the gas is no
longer free to escape, the pressure generated by the gas drives the acid away from
the solid ; this stops the further generation of gas. When the pressure is relieved
by allowing the gas to escape, the acid again comes in contact with the solid. In
the better types of apparatus (i) the freshest acid is brought in contact with the
solid ; (ii) the emptying and recharging is simple ; and (iii) a great over-pressure is
avoided.
The properties of hydrogen. — Hydrogen gas is colourless and odourless —
the impure gas may have a smell. The hydrogen gas streaming from the generating
flask can be lighted, and a flame of burning hydrogen is obtained which was formerly
called lumen philosophicum, or the philosopher's flame. To get the flame to burn
steadily it is best to interpose between the exit tube and the jet, a wider tube loosely
packed with granulated calcium chloride to arrest by absorption the water vapoui
carried along with the gas. The hydrogen flame is very hot and melts ordinary
glass ; a jet of hard glass, quartz glass, or platinum can be used. When a lighted
taper is plunged into a jar of hydrogen held mouth downwards, the gas burns with
a scarcely visible blue flame at the mouth of the jar, and the taper is extinguished
showing that the gas is combustible and a non-supporter of combustion. When
J. Black (1766) heard that H. Cavendish had found hydrogen to be much lighter
than air, he thought that possibly a thin bag made from the allantois of a calf, when
filled with hydrogen, would be buoyed up by air. Modifications of Black's idea
are used as illustrative experiments on the lecture table, and not long afterwards
the gas was used for filling balloons. The gas can be poured upwards from one jar
to another, and it can be proved that the gas has actually been transferred from
the one vessel to the other by testing the contents of each jar with a lighted taper
before and after the pouring.
HYDROGEN AND THE COMPOSITION OF WATER
127
The extreme lightness of hydrogen and its combustibility enable many ingenioiis
cperiments to be performed with the gas. For instance, a cardboard box or a light glass
ressel can be coimterpoised bottom upwards, on a balance ; the beam will ascend when
lydrogen is poured upwards into the inverted vessel. Soap-bubbles blown with the gas,
>r collodion balloons filled with the gas, rise to the ceiling very quickly. The gas may be
^syphoned upwards from one vessel to another, or, the gas may be syphoned from, say a
bell-jar and burnt at the long leg of the syphon. An explosive mixture with air is formed
when the hydrogen has nearly all been syphoned away, and the flame at the top of the long
leg of the syphon will then rush back and produce a loud but harmless explosion.
The explosive character of a mixture of hydrogen with oxygen of air can be
illustrated by mixing two volumes of hydrogen gas with either one volume of oxygen
or five volumes of air in a soda-water bottle. A lighted taper applied to the mouth
of the bottle causes the gas to detonate. The combustion of the whole mass is almost
instantaneous. The explosion is so violent that we can understand why N.Lemery,
in his Explication physique et chimique des eclairs et du tonnere (1700), tried to show
that thunder and lightning are produced by the fulminations of hydrogen.'' The
sound obtained when a long glass tube is placed about the flame of burning hydrogen
led to W. Higgins (1777) calling the experiment the chemical harmonicon. The tones
vary with the diameter, thickness, and length of the tube and on the nature of the
jet. The sound appears to be the effect of an extremely rapid series of explosions.
M. Faraday obtained a similar musical flame with inflammable gases and vapours
other than hydrogen. M. Faraday's explanation is that a strong current of air is
established ; this lengthens the flame, and small portions of air are mixed with the
hydrogen in such a manner as to form small quantities of detonating gas, which,
when set on fire, produces slight explosions succeeding each other quickly and
regularly. C. Wheatstone found that while producing sound within a glass tube,
regular intermissions in the intensity of the flame are observed, and these present
a chain-like appearance on a revolving mirror, indicating alternate contractions and
dilations of the flame corresponding with the sonorous vibrations of the column
of air.
Joseph Priestley ^ has told us that in 1776 his friend, J. Warltire, had noticed that
when a flame of hydrogen is allowed to burn in air confined under a bell -jar, the
whole of the receiver appears to be filled with " a fine powdery substance like a
whitish cloud," when the flame was extinguished ; and the air left in the glass was
found to be " perfectly noxious." In the same year P. J. Macquer ^ inquired
whether the flame of hydrogen evolved smoke or soot. He thus described his
experiment :
By placing a saucer of white porcelain in a jet of inflammable gas (hydrogen) burning
tranquilly at an orifice, I foimd that the part of the saucer which the flame licked was
moistened by small drops of liquid as clear as water, and which, in fact, appeared to be
nothing but pure water.
It is probable that J. Warltire's white cloud was not produced by a finely powdered
soHd, but by minute drops of water. In 1779, J. R. Sigaud de la Fond also mentioned
the formation of water during the combustion of inflammable air. P. J. Macquer did
not stop to inquire : Whence came the water ? He has been blamed because he felt
no astonishment at that which is really astonishing, for he merely mentions, with-
out comment, the appearance of the water. P.J. Macquer did not see before him a
great discovery begging for recognition. Hence, asks F. J. Arago (1839), 1° is
genius in the observational sciences to be reduced to the faculty of asking an
appropriate Why 1 The inquiry can be made, (1) What happens to the surrounding
air during the burning of a jet of hydrogen 1 and (2) Is the product of the action
really water ?
J. Warltire's 1776 experiment can be modified by making a jet of dried hydrogen
burn under a bell- jar containing a measured volume of air standing over water.
At first, there is a momentary expansion of the air due to the heating of the confined
air by the flame ; immediately afterwards, the water rises in the jar, and the
hydrogen flame gradually expires. Immediately this occurs the stream of gas is
128 INOKGANIC AND THEORETICAL CHEMISTRY
stopped to prevent it passing into the air in the bell-jar. The gas remaining in the
jar has quite similar properties to the nitrogen gas remaining after mercury is
calcined in air. It is the " perfectly noxious air " alluded to by J. Warltire. In 1777,
C. W. Scheele described an analogous experiment in his Chemische Ahhandlungen von
der Luftundvon dem Feuer (Upsala, 1777), but with other combustible agents. The
experiment shows that when hydrogen bums in air, it unites with the oxygen
and leaves nitrogen behind. If the experiment be carefully made, nearly four-
fifths of the original volume of air remains. The burning hydrogen removes nearly
one-fifth of the original volume of air. Hydrogen does not burn in the residual
nitrogen — although about 7 or 8 per cent, of oxygen is still present. A certain
amount of dew collects on the inner walls of the bell-jar, but that, of course, may
come from the water in the dish below. In fine, the facts give reasons for supposing
that hydrogen, in burning, combines with oxygen to form an oxide of hydrogen in
the same sense that mercury, when calcined in air, combines with oxygen to form
mercuric oxide. It remains to try and isolate a sufficient quantity of the hydrogen
oxide whose existence has just been inferred, but not proved, in order that its
properties may be examined more closely.
The experiment of P. J. Macquer (1778) can be modified so that a jet of dried
hydrogen is burned under a funnel, the stem of which is curved so that it passes into
a two-necked globe ; the other neck of the globe is connected with an aspirator so
that the products of combustion from the hydrogen flame can be aspirated through
the system. The glass bulb is kept cold and a clear colourless liquid collects therein.
This liquid has all the properties of water ; it 'is a clear, colourless, and tasteless
liquid with no smell ; it freezes at 0°, and boils at 100°. The water does not come
from the condensation of the moisture already present in the gas as it rises from
the generating vessel, because the gas is dried by the " scrubbing " it receives as it
passes along the tower of calcium chloride ; this statement can be tested by making
a blank experiment with the un-ignited gas. It is therefore inferred that water
is burnt hydrogen, or the calx of hydrogen ; otherwise expressed, water is hydrogen
oxide formed when hydrogen bums in air. Hydrogen and oxygen are both
gases, and it is therefore more difficult to find the combining ratio Hydrogen :
Oxygen in the formation of hydrogen calx, by direct weighing, than is the case
with the metallic calces. It remains therefore to show how chemists have solved
the problem.
Repeeenoes.
1 T. Turquet de Mayerne, Pharmacopoea, London, L703.
2 R. Boyle, New Experiments touching the relation between flame and air, London, 1671 ;
N. Lemery, Mem. Acad., 101, 1700.
3 S. Hales, Vegetable Staticks, London, 1727.
4 H. Cavendish, Phil. Trans., 56. 141, 1766; 74. 119, 176, 1784; 75. 372, 1785; R. Kirwan,
ib.y 72. 179, 1782; J. Watt, ib., 74. 329, 1784; J. Priestley, Experiments on Air, Birmingham,
6, 1, 1786; F. de Lassone, Mem. Acad., 563, 1776; C. W. Scheele, Chemische Abhandlungen von
der Luft und dem Feuer, Upsala, 1777.
6 A. L. Lavoisier, Mem. Acad., 468, 1784 : (Euvres, Paris, 2. 509, 1862 ; J. B. A. Dumas,
Lecons sur la philosophic chimique, Paris, 158, 1827.
'« C. Cloez, Bull. Soc. Chim., (2), 43. 102, 1885 ; G. Tissandier, ib., (2), 43. 233, 1885 ; V.
Wartha, Ber., 5. 151, 1872; J. Meister, Zeit. anal. Chem., 25. 373, 1886; R. Fresenius, ib., 12.
73, 1873 ; W. Ostwald, ib., 31. 184, 1892 ; R. J. Friswell, Chem. News, 90. 154, 1904 ; 94. 106,
1906; C. Thiele, Chem. Ztg., 25. 468, 1901 ; C. Arnold, ib., 26. 229, 1902; L. L. de Koninck,
ib., 17. 1099, 1893 ; F. M. Perkin, Jovm. Soc. Chem. Ind., 20. 438, 1901 ; H. Hafelin, Pharm.
Ztg., 50. 351, 1905 ; E. Egasse, Dinglers' Journ., 244. 54, 1882 ; H. Arzberger, Pharm. Post,
37. 581, 1904; F. W. Kiister, Journ. prakt. Chem., (2), 48. 595, 1893; J. W. Dobereiner,
Schweiggers' Journ., 38. 326, 1823 ; 39. 159, 1823 ; 63. 468, 1831 ; C. Aschmann, Chem. Ztg., 21,
1049, 1897 ; U. Eebel, ib.,29. 141, 1905; J. D. Edwards, Journ. Ind. Eng. Chem., 11. 961, 1919.
' N. Lemery, Mem. Acad., 101, 1700 ; W. Higgins, Nicholson's Journ., 1. 130, 1777 ;
M. Faraday, Quart. Journ. Science, 5. 274, 1818 ; C. Wheatstone, Phil. Trans., 124. 586, 1834 ;
F. Schaffgotsch, Pogg. Ann., 100. 352, 1857 : 101. 471, 1857 : 102. 627, 1857 ; J. Tyndall, Phil.
Mag., (4), 13. 473, 1857 ; A. Schrotter, Sitzber. Akad. Wien, 24. 18, 1857 ; A. Terquem, Compt.
Rend., 66. 1037, 1868.
HYDROGEN AND THE COMPOSITION OF WATER 129
8 J. Priestley, Experiments and Observations on Different Kinds of Air, London, 3. 367, 1777.
» r. J. Macquer, Dictionnaire de chimie, Paris, 2. 314, 1778; J. R. Sigaud de la Fond,' Kssai
8ur differentes especes d'air, Paris, 1776.
i» F. J. Arago, JiJloge historique de James Watt, Paris, 1834 ; (Euvres, Paris, 1. 454, 1854.
§ 3. Dumas' Experiment on the Composition of Water by Weight
After very careful examination of all the analytical researches made for the determina-
tion of atomic weights, I emphatically declare that the researches of Dumas are the most
important of all, marking as they do the beginning of the analysis of precision, and offering
also the first instance of a true series of determinations, such as is required to furnish the
absolute values of the atomic weights. — -G. D. Hinrichs.
Several determinations of the combining weights of hydrogen and oxygen in
the formation of water have been made. Prior to J. B. A. Dumas' work, there
were the pioneer attempts to find the combining ratio of hydrogen and oxvgen by
M. Monge, A. L. Lavoisier, and M. Meusnier i about 1786. They admitted
measured volumes of hydrogen and oxygen into a globe, exploded the mixture,
and after repeating the process 372 times, weighed the water produced, and
calculated the weights of oxygen and hydrogen employed from the densities of the
gases. It was found that in water the ratio of the weight of hydrogen to that of
oxygen is as 1 : 6*61. In 1791, A. F. de Fourcroy, L. N. Vauquelin, and M. Seguin
repeated M. Monge's work and found the ratio to be 1 : 6'17. In 1803, John Dalton
estimated the ratio of hydrogen to that of oxygen to be 1 : 5' 66, a result further
removed from the truth than the ratios found by the French savants. J. Dalton
corrected his first result in 1808, and gave the ratio 1:7. In 1814, from J. L. Gay
Lussac and A. Humboldt's observation that two volumes of hydrogen and one
volume of oxygen unite to form water, and J. B. Biot and F. J. Arago's observation
of the relative densities of these two gases, W. H. Wollaston calculated the ratio
of the weights of hydrogen and oxygen in water to be 1 : 7 '545. This was followed
by the work of P. L. Dulong and J. J. Berzelius in 1819, and of J. B. A. Dumas
in 1842.
Hydrogen does not combine readily with many of the elements, but it readily
combines with oxygen, chlorine, fluorine, lithium, and a number of others. So
great is the attraction of hydrogen for oxygen that it will very often remove oxygen
from its combinations with the other elements. For instance, on March 6th, 1783,
J. Priestley 2 reported that he had confined lead oxide (minium or red lead) in a
tall cylinder containing inflammable air standing over water ; the red oxide of
lead was heated in the focus of a burning glass. He observed :
The minium became black, and then ran in the form of perfect lead ; at the same time
the air diminished at a great rate, and the water ascended within the cylinder. . . . Seeing
that metal to be actually revived, and that in a considerable quantity, at the same time that
the air was diminished, I could not doubt that the calx was actually imbibing something
from the air ; and from its effects in making the calx into a metal, it could be no other
than that to which chemists had unanimously given the name phlogiston. . . . Con-
sequently, phlogiston is the same thing as inflammable air.
The experiment was varied by confining the gases over mercury in place of water,
and using other calces— e.^r. the oxides of tin, bismuth, mercury, silver, iron, and
copper. He further found that " 1 oz. of lead was revived by 108 oz. measures of
inflammable air, and 1 oz. of tin by 377 oz. measures." The 108 oz. and 377 oz.
measures of inflammable air would weigh nearly 4*4 and 15*4 grains respectively.
Priestley's measurements are good, because these numbers are close to their ideal
values, 4- 6 and 16'3 grains respectively. This remarkable experiment might have
opened J. Priestley's eyes to the insufficiency of the phlogiston hypotheses. A. L.
Lavoisier 3 has pointed out that J. Priestley did not notice that there was a decrease
in the weight of the solid during the reduction, and that water was a product of
the reaction. The true interpretation of the reduction observed by J. Priestley is
due to A. L. Lavoisier.
VOL. I. K
130
INORGANIC AND THEORETICAL CHEMISTRY
In J. Priestley's experiment, the hydrogen is said to be oxidized ; and the
metallic oxide, reduced or deoxidized. The hydrogen is called a reducing agent, that
is, a reducer or deoxidizer ; and the copper oxide an oxidizing agent or
oxidizer J because it oxidizes hydrogen to water. The reaction under consideration
is both an oxidation and a reduction process. All depends upon whether the
hydrogen or the copper be under consideration. In the fifteenth century, Paracelsus
applied the term reduction to the preparation of the metals. During a reduction,
the reducing agent is usually, not always, oxidized ; and during an oxidation, the
oxidizing agent, reduced. If a known amount of copper oxide be reduced by
hydrogen, and the water formed be collected and weighed, the weight of the reduced
copper oxide will show how much oxygen has been used in forming a definite
amount of water. This was done by P. L. Dulong and J. J. Berzelius ^ in 1820, and
by J. B. A. Dumas in his celebrated Recherches sur la composition de Veau in 1843.
J. B. A. Dumas' experiment is not the best of its kind, although it was the best
of its time, and it has long and deservedly held an honoured place in chemical
text-books. The experiment illustrates some important principles, and it is
Purification
Hydrogen. Copper
Protective
Water formed. Tube.
Fig. 2. — Dumas' Experiment (abbreviated).
therefore here described in outline. The first stage of the work involved the
purification of hydrogen.
The hydrogen was prepared by the action of zinc on sulphuric acid. It might be
thought that pure zinc and pure sulphuric acid should be used. Experiment shows,
curiously enough, that under these conditions the action is so very, very slow that some
have jumped to the conchision that " absolutely pure sulphuric acid, even when diluted
with pure water, has no action on perfectly pure zinc." Moreover, it is exceedingly difficult
to prepare pure zinc and pure sulphuric acid. Hence, pure reagents were not used for the
preparation of the hydrogen. Accordingly, the gas may contain nitrogen and oxygen
derived from the air ; sulphur dioxide and hydrogen sulphide derived from the reduction
of the sulphuric acid by the hydrogen ; carbon dioxide ; arsenic hydride (if the acid or the
zinc contained arsenic) ; hydrogen phosphide (if the zinc or the acid contained phosphorus) ;
nitrogen oxides (if the acid contained nitrogen oxides) ; and water vapour. Accordingly,
J. B. A. Dumas (1842) used sulphuric acid, which had been well boiled, to get rid of dis-
solved air, and then passed the hydrogen through a series of U -tubes- — ^Fig. 2- — containing:
(1) pieces of glass moistened with lead nitrate to remove hydrogen sulphide ; (2) solution
of silver sulphate to remove arsenic and phosphorus compounds ; (3) solid potassium
hydroxide to remove sulphur dioxide, carbon dioxide, and nitrogen oxides ; and (4) phos-
phorus pentoxide to remove moisture not absorbed by the solid potassium hydroxide.
J. B. A. Dumas used three potassium hydroxide tubes, and two phosphorus pentoxide
tubes — like (4) — only one of each is in the diagram. The phosphorus pentoxide tubes were
placed in a freezing mixture. The tube marked (5) in the diagram contained phosphorus
pentoxide, and it was assumed that the hydrogen passing through was quite dry — this
tube is accordingly called a temoin tube {temoin, a witness) because it c€ua he employed as
HYDROGEN AND THE COMPOSITION OF WATER 131
evidence that the hydrogen which passed through gave up no moisture to the desiccating
agent.
J. B. A. Dumas passed the purified hydrogen over red-hot copper oxide, and
determined the loss of weight (oxygen) which occurred. He then weighed the
amount of water produced.
The purified hydrogen was passed through a weighed bulb. A, containing copper
oxide, and heated by the spirit lamp underneath. Most of the water condensed in the
bulb B, and the remainder was absorbed in the U-tube G containing solid potassium
hydroxide, and in D and E containing phosphorus pentoxide. The phosphorus pentoxide
tube D was kept cool by a freezing mixture. The three tubes C, 2), E, and the bulb B,
were weighed before and after the experiment. The last U-tube, F^ containing phosphorus
pentoxide was followed by a cylinder, Q, of sulphuric acid through which hydrogen escaped.
The vessels F and Q were not weighed ; they served to protect the other tubes from the
external atmosphere.
The average of nineteen experiments by J. B. A. Dumas (1842) gave :
Copper oxide lost in weight . . . .44*22 grams
Water produced . . . . . . 49*76 „
Hydrogen (by difference) . . 5*54 „
Hence, he inferred that 15*97 parts of oxygen united with two parts of hydrogen
to form water, or 16 parts by weight of oxygen combined with 2 '004 parts by weight
of hydrogen to form water. His nineteen values ranged between 15'892 and
16'024, and his mean value for hydrogen is usually considered to be rather low. or
the mean value for oxygen rather high. A later determination by E. W. Morley gave
16 : 2'016. In approximate work, we may take it that 2 parts by weight of hydrogen
combine with 16 parts by weight of oxygen to form 18 parts of water ; indeed,
J. B. A. Dumas himself expressed his belief that the true value of the ratio Hydrogen :
Oxygen is probably 2 : 16.
It is common to append to the arithmetical mean of a series of observations the
so-called probable error. For example, the mean of Dumas' nineteen determina-
tions of the relative weights of hydrogen and oxygen in water is given as : Oxygen,
15*96 ± 0*007, when hydrogen is 2 ; and 0. L. Erdmann and R. F. Marchand's eight
determinations by a similar method are represented by the average 15*973 ± 0*011.
The probable error in the one place is ±00*07 and in the other ±0*011. This does
not mean that J. B. A. Dumas' results were nearer the true value than 0. L. Erdmann
and R. F. Marchand's. The probable error does noo tell how nearly the average
of a given number of similarly conducted experiments would approach the average
actually found. In J. B. A. Dumas' result, the chances are even that the true average
of the determination by his method lies between (15*96 + 0*007 =) 15*967 and
(15*96 — 0*007 =) 15*953. If an unrecognized constant error affected all the
results, the average actually found would still differ from the true value by this
amount. As a matter of fact, when J. B. A. Dumas had nearly finished his work,
he did find that his numbers were affected by a curious error, previously un-
recognized, so that the concordance of his individual determinations did not ^ prove
that his average was right. This error, if not corrected, makes the result appear
a little low. The reduced copper retains some hydrogen very tenaciously ; 5
similarly, when copper oxide is made, as is usually the case, by calcining the
nitrate to redness in a current of air, it retains an appreciable amount of
nitrogen. As a result, when the oxide is reduced in a current of hydrogen, the
weight of the water formed is less than that which corresponds with the loss of
weight which has occurred during the reduction of the copper oxide, assuming that
water is really formed by the union of hydrogen and oxygen.
The main objections to J. B. A. "^Dumas' work turn on the following facts :
(1) There is a great difficulty in thoroughly removing all the air from a large com-
pUcated apparatus ; (2) The absorption of air by sulphuric acid which is slowly
evolved along with the hydrogen when the acid acts on zinc ; (3) M. Melsens showed
132 INOEGANIC AND THEORETICAL CHEMISTRY
that there is a retention or occlusion of hydrogen by the reduced copper ; (4) T. W.
Richards and E. F. Rogers showed that the copper oxide was probably contaminated
with occluded nitrogen and other gases ; (5) W. Dittmar and J. B. Henderson showed
that there is a slight reduction of sulphuric acid by hydrogen to form gaseous sulphur
dioxide (which is later absorbed by the potash) ; (6) The difficulty in drying the
gas, etc., completely. The last is considered by T. W. Richards (1911) to be
one of the most fertile sources of error in the determination of accurate equivalents.
(7) Before the hydrogen reached the copper oxide, J. B. A. Dumas dried it with
sulphuric acid and phosphorus pentoxide, and used calcium chloride to remove
the aqueous vapour from the excess of hydrogen which left the copper oxide bulbs.
Since phosphorus pentoxide removes more moisture from a gas than calcium
chloride, it is possible that some aqueous vapour escaped. This would tend to
give high results. (8) E. W. Morley has shown that hydrogen from sulphuric acid
and zinc always contains carbon compounds which cannot be removed by absorption ;
(9) J. J. Berzelius emphasized the fact that the displacement of hydrogen by air at
the end of J. B. A. Dumas' experiment, saturated the liquid water with air and made
its weight too large ; and (10) W. Dittmar has stated that J. B. A. Dumas did not
correct his weighings for the buoyancy of air. This would make his weighing of
the water produced appear too low.
In 1892 G. D. Hinrichs^ argued that the combining ratio of oxygen in J. S. Stas'
determinations is a function of the amount of potassium chlorate employed, such
that with 30-35 grms. of chlorate the atomic weight is 16, and with 100 grms. of
chlorate, 15'98. Similar results were obtained with the determinations of J. B. A.
Dumas, J. S. Stas, and J. P. Cooke of the atomic weight of sulphur, chlorine, bromine,
etc. Hence G. D. Hinrichs argues that the atomic weight should be calculated not
from the mean, but from the limiting value corresponding with zero weight of the
substance. P. A. Guye and E. Moles have shown that the relation observed by
G. D. Hinrichs is confined to determinations in which the quantities of substances
employed have been weighed in air and the reduction to vacuum effected by
calculation ; and the results are satisfactorily explained by assuming that the
anomaly is due to the surface condensation of air and moisture, and should there-
fore disappear when the weighings are conducted in vacuo. The average deviation
due to this cause is between 1 in 10,000 and 1 in 20,000. P. A. Guye and E. Moles
found that with silver the error due to surface condensation is 2 X 10^ ^ gram per
gram of metal.
References.
1 M. Monge, Mem. Acad., 78, 1786 ; A. L. Lavoisier and M. Meusnier, ib., 269, 1781 ; A. F. tie
Fourcroy, L. N. Vauquelin, and M. Seguin, Ann. Chim. Phijs., (1), 8. 113, 183, 1791 ; (1), 9. 7,
29. 1791; P. L. Dulong and J. J. Berzelius, ib., (2), 15. 386, 1820; J. B. A. Dumas, ib., (3), 8. 189,
1843; W. H. WoUaston, Phil. Trans., 104. 20, 1814; J. Dalton, A New System of Chemical
Philosophy, London, 1808 ; H. E. Roscoe and A. Harden, A New View of the Origin of Dalton' a
Atomic Theory, London, 1896.
2 J. Priestley, Experiments and Observations on Different Kinds of Air, Birmingham, 1786.
' A. L. Lavoisier, Mem. Acad., 488, 1784.
* P. L. Dulong and J. J. Berzelius, Ann. Chim. Phys., (1), 15. 386, 1820 ; J. B. A. Dumas,
ib., (3), 8. 189, 1843.
5 J. B. A. Dumas, Ann. Chim. Phys., (3), 8. 189, 1843 ; M. Melsens, ib., (3), 8. 205. 1843 ;
W. Dittmar and J. B. Henderson, Proc. Phil. Soc. Glasgow, 22. 1, 1891 ; E. H. Reiser, Per., 20. 2323,
1887; G. S. Johnson, Chem. News, 35. 232, 1879; 59. 272, 1889; W. A. Noyes, Amer. Chem.
Journ., 12. 441, 1890 ; G. Neumann and F. Streintz, Monatsh., 12. 642, 189l'; J. J. Berzelius
Lehrbuch der Chemie, Dresden, 3. 1183, 1848; W. Dittmar, Chem. News, 61. 76, 1890; T. W.
Richards, Journ. Chem. Soc, 99. 1201, 1911.
6 G. D. Hinrichs, Compt. Rend., 115. 1074, 1892; 116. 753, 1893; 118. 528, 1894; P. A.
Guye and E. Moles, Journ. Chim. Phys., 15. 360, 405, 1917.
§ 4. E. W. Morley's Experiment on the Composition of Water by Weight
\n the determination of atomic weights, a small number of values are to be regarded
as fundamental. They are the standards of reference ; and by comparison with them all
HYDROGEN AND THE COMPOSITION OF WATER
133
the other atomic weights are established. The atomic weights of hydrogen and oxygen
are primarj'^ ; that is, one or other of them is the basis of the entire system of
atomic weights.^ — -F. W. Clarke.
It will be observed that P. L. Dulong and J. J. Berzelius, and J. B. A. Dumas
weighed the oxygen and the water, and estimated the hydrogen by difference.
Then followed the work of J. Thomsen, J. P. Cooke and T. W. Richards, and E. H.
Reiser, in which the hydrogen and water were weighed, and the oxygen estimated
by difference. W. A. Noyes and Lord Rayleigh weighed the oxygen and hydrogen,
and estimated the corresponding weight of water. In his memoir On the Density
of Hydrogen and Oxygen, and the Ratio of their Atomic Weights (Washington, 1895),
E. W. Morley first made what J. S. Stas called a synthese cotnplete by weighing all
three quantities— oxygen, hydrogen, and water. He synthesized water by burning
hydrogen in oxygen, and weighed both gases separately and afterwards in combina-
tion. In this way he was able to determine the combining ratio of hydrogen and
oxygen. Since the combining ratio of oxygen with a number of metals has already
been determined, the combining ratios of the same metals with respect to hydrogen
can be computed when once the ratio Hydrogen : Oxygen is accurately known.
Known weights of pure dry hydrogen and pure dry oxygen were stored in two large
glass globes. The vessels containing the hydrogen and oxygen were weighed separately.
The hydrogen was prepared by heating palladium hydride,
and the oxygen by heating potassium chlorate. The
hydrogen was weighed as palladium hydride, and the
oxygen was weighed in a compensated glass globe. The
apparatus for storing and drying the hydrogen and
oxygen is not shown in Fig. 3. The globe containing
oxygen was connected with the apparatus, and the oxygen
passed through a layer of phosphorus pentoxide, and thence
into the glass chamber M, via one of the jets A ; the globe
containing hydrogen was similarly connected with the other
tube containing phosphorus j)entoxide, and the hy-
drogen led into the chamber M rid one of the jets A.
The phosphorus pentoxide was not intended to dry the
entering gases — these had already been dried. The
chamber M was previously evacuated and weighed. One
of the gases, say oxygen, was allowed to enter M, and
electric sparks were passed across the terminals F just
over the jets A. Hydrogen was led into the apparatus
and ignited by the sparks. The rates at which hydrogen
and oxygen entered the chamber were regulated so that
the formation of water was continuous. The water formed
was condensed, and collected in the lower part of the
chamber. To hasten the condensation the apparatus was
placed in a vessel of cold water — dotted in the diagram.
When a sufficient amount of water was formed, the
apparatus was placed in a freezing mixture. The mixture
of unconsumed oxygen and hydrogen remaining in the
tube was pumped away, and analyzed. The weights of
hydrogen and oxygen so obtained wore added to the
weights of unconsumed hydrogen and oxygen remaining
in the globes. The phosphorus pentoxide tubes prevented
the escape of water vapour. The amount of water formed
was determined from the difference in the weights of the
system M before and after experiment. The amounts
of hydrogen and oxygen were determined from the -
weights of the corresponding globes before and after the experiment, ihe ^J^^^^^^\
water formed was determined from the increase in the weight of the above descriDea ve.sei
before and after the combustion.
Pjq, 3. — Morley 's Experiment
—Synthesis of Water.
E. W. Morley, as a mean of eleven experiments, found that :
3-7198 grams
. 29-5335 „
. 33-2630 „
Hydrogen used
Oxygen used
Water formed
Hence, taking oxygen = 16 as the unit for combining weight, it follows that 16
parts by weight of oxygen combine with 2-016 parts by weight of hydrogen
134
INORGANIC AND THEORETICAL CHEMISTRY
to form 18*016 parts o! water — within the Hmits of the small experimental
error, and, adds E. W. Morley : " Until further light is obtained concerning the
sources of error which doubtless afEect all these experiments, this value is the most
probable that can be derived from existing data." It might be added that the
ratio Oxygen : Hydrogen = 16 : a; for twenty -five sets of determinations by other
workers, made since 1821, using different methods, has values of x ranging
between 2-003 and 2*018.
§ 5. The Decomposition o! Water by Metals
If, as I have tried to demonstrate, water is really a compound of hydrogen with oxygen
... in order to obtain hydrogen, it is only necessary to bring water in contact with a
substance for which the oxygen has more affinity than it has for hydrogen, in order to liberate
the hydrogen as a gas. Iron is commonly used for this purpose, and it is necessary to raise
the temperature to a red heat in order to effect the separation. . . . There is une veritable
oxidation dufer par Veau. . . . The oxygen is fixed by unity with the iron, and the hydrogen
is disengaged as an inflammable gas.' — A. L. Lavoisier (1789).
Water remains permanent and stable so long as the balance of the forces between
its constituent elements is maintained, but in the presence of a metal which can
unite with one of these elements, the water may be decomposed. One element — say
hydrogen — is set free, while the other element — oxygen — unites with the agent of
destruction to form a new compound — oxide of the metal. The application of this
principle was suggested to A. L. Lavoisier by the illustrious P. S. de Laplace ; and
as a result, the first conscious analysis of water was made by A. L. Lavoisier, assisted
by M. Meusnier, about 1784. This particular process has the disadvantage of
isolating only one of the two elements of water. In their Memoire oil Von prouve
par la decomposition de Veau, que ce jiuide n'est point une substance simphy et qu'il
y a plusieurs moyens d'obtenir en grand Vair inflammable qui y entre comme princife
constituant, A. L. Lavoisier and M. Meusnier (1771) ^ passed steam over hot iron,
and found that the metaUic iron was converted into a *' black oxide precisely
similar to that produced by the combustion of iron in oxygen gas " ; otherwise
expressed, the iron is oxidized by the water, and the water is reduced by the iron,
forming " a peculiar inflammable gas," which Lavoisier named hydrogen, because
" no other term seemed more appropriate." The word signifies the generative
principle of water, from the Greek vSwp, water, and ycwaw, I generate or produce.
The German word for hydrogen is Wasserstqff — the stufi from which water is made.
The following is a modernized form of M. Meusnier and A. L. Lavoisier's experiment :■ —
Fill an iron, porcelain, or hard glass tube- — 60 cm. long and 1'5 cm. diameter- — with bright
iron turnings or bright iron nails. In Fig. 4 a hard glass tube is used. This is drawn out
at one end as shown in the
diagram. This end is fitted
with a delivery tube dipping in
a gas trough. A roll of pre-
viously ignited asbestos paper,
6 cm. long, is inserted in the
opposite end. This end is closed
with a red rubber stopper and
the exit tube of the flask so
arranged that it passes a short
distance into the core of the
asbestos paper. The asbestos
roll, later on, prevents the
liquid water from coming into
contact with the hot glass and
breaking the tube. Water is
boiled in the flask, and the
When all the air has been driven
Fig. 4.
-Decomposition of Steam by Hot Iron — A.
Lavoisier and M. Meusnier's Experiment.
steam passing through the iron turnings is decomposed.
out of the apparatus, hydrogen may be collected in the gas jar
If zinc be used in place of iron, the temperature need not be much higher
HYDROGEN AND THE COMPOSITION OF WATER 135
than the boiling point of water, since zinc reduces steam and forms zinc oxide at
a comparatively low temperature. H. V. Regnault 2 found the zinc oxide is crystal-
line if the reaction occurs at red heat. If a strip of magnesium ribbon be placed
in a bulb of a hard glass tube and heated, in a current of steam, at a red heat, the
metal appears to burst into flame, forming magnesium oxide. The resulting
hydrogen can be ignited if the jet of steam be not too vigorous. According to
A. Ditte (1871), magnesium decomposes water slowly at 70° ; and according to
H. Fleck and H. Bassett (1895), magnesium amalgam decomposes cold water. In
A. W. Knapp's experiment (1912) powdered magnesium is added to ten times its
weight of water with a little palladious chloride in solution ; metallic palladium is
formed and this metal acts catalytically or electrolytically on the water. The
decomposition is then so vigorous that the water appears to boil, and the escaping
hydrogen ignites spontaneously. Metallic calcium decomposes cold water and
gives off hydrogen, but the action slows down very soon, probably because the
calcium hydroxide is not all dissolved by the water, and in consequence a protective
crust of this substance forms over the surface of the metal. The calcium can be
advantageously warmed with water in a flask which is connected directly with a
delivery tube leading to the gas trough. If the water is not free from carbonates,
a crust of calcium carbonate also forms over the surface of the metal. Calcium
hydroxide is formed as well as hydrogen. The reaction with strontium is rather
more vigorous than with calcium ; and with barium more energetic than with
strontium. The metal sodium decomposes cold water, giving off hydrogen, and
forming sodium hydroxide. So much heat is generated during the reaction that
the metal melts, showing that its temperature has risen over 95°. The experiment
is liable to unpleasant explosions when the sodium is confined so as to enable the
resulting hydrogen to be collected. The cause of the explosion has not been
definitely established ; it has been attributed to the formation of a dioxide or a
hydride. 3 It is more likely to be due to the formation of a film or bubble of water
superheated above its boiling point. Potassium alone reacts so violently with
water that the temperature rises high enough to set fire to the hydrogen. The
hydrogen burns with a violet-tinged flame, owing to the presence of the vapour
of potassium ; the hydrogen produced by the action of sodium on water burns
with a yellow flame, owing to the contamination of the hydrogen with the vapour
of sodium. According to J. J. Berzelius, a solution of metallic sodium in mercury —
sodium amalgam — decomposes water much less turbulently than sodium alone ;
the result is similar when a small piece of potassium amalgam — 3 or 4 mm.
diameter — is placed on water. J. J. Berzelius says the gas obtained by the alkali
amalgam is odourless, but if an acid or ammonium chloride is also present, the
product smells like the gas derived from the dissolution of zinc in acids.
This set of experiments gives a series of metals which appear to react with
water with increasing violence ; the metals — iron, zinc, magnesium, calcium, sodium,
potassium — seem to have an increasing avidity or affinity for oxygen so that they
are able to tear the whole of the oxygen from the water, fix the oxygen, and thus
liberate half or all the hydrogen as a gas. Under suitable conditions, by treatment
with fluorine, chlorine, or bromine, the hydrogen is fixed and the oxygen liberated
as a gas. Still further, by passing an electric current through water, both components
are liberated in the gaseous state.
References.
1 M. Meusnier and A. L. Lavoisier, Mem. Acad., 269, 1784 ; A. L. Lavoisier, ib., 468, 1784.
2 F. G. Benedict, Chemical Lecture Experiments, New York, 1901 ; M. Rosenfeld, Ber ,\b.
161, 1882 ; 26. 59, 1893 ; Journ. praU. Chem., 12). 48. 599, 1893 ; G. T. Uoody, P roc. C hem.
Soc., 7. 20, 1891; H. V. Regnault, Ann. Chim. Phys., (3), 43. 477, 1855; A. W. Hofmann,
Introduction to Modern Chemistry, London, 1865; Ber., 15. 2663, 1882; J. B. Mevick, Amer.
Chem., 7. 276, 1877 ; A. Senier, Chem. News, 91. 87, 1905 ; A. W. Knapp, ih., Ij^. 253, 1912 ;
A. Ditte, Compt. Rend., 73. 108, 1871 ; H. Fleck and H. Bassett, Journ. Amer. Chem. 6oc., 17-
789, 1895 ; J. J. Berzelius, Lehrbuch der Chemie, Dresden, 1. 769, 1825.
3 R. Bottger, Journ. prakt. Chem., (1), 85. 397, 1862 ; M. Rosenfeld, ib., (2), 48. 699, 189.^.
136
INORGANIC AND THEORETICAL CHEMISTRY
§ 6. The Decomposition of Water by Electricity
Electricity is a key which will open a way into the innermost parts of nature.-
RiTTER (1798).
W.
In 1758, G. B, Beccaria ^ exposed water to powerful electric sparks, and although
he must have decomposed this substance, he does not seem to have been aware ot
it ; in 1789, the Dutch chemists P. van Troostwijk and J. R. Deiman noticed that
when an electric charge from a powerful electric
machine was passed through water, bubbles of gas
were obtained. They showed that the gases
were not due to the expulsion of air dissolved
by the water, since the same result was obtained
by using distilled water, and water freed from
dissolved air by a prolonged boiling. Hence,
it can be inferred that water is decomposed
into its constituent gases by the electric dis-
charge. On May 2nd, 1800, W. Nicholson and
A. Carlisle ^ happened to put a drop of water
in contact with two wires from a voltaic battery,
and noticed the formation of small bubbles of
gas about the tips of the wires provided the wires
were not in contact. They then immersed the
two wires in a glass of water, and found that
Fig. 5.-J. W. Ritter's Apparatus ^^f,^^ were formed about both wires; the gas
(1800)for the Electrolysis ot Water collected about one wire was hydrogen, and
— Gases separated. about the other wire, oxygen. Hence, hydrogen
and oxygen are produced during the electrolysis
of water. The gases were mixed and exploded. The result was water. This is
very interesting — chemical combinations can produce an electric current ; here an
electric current is used to produce chemical decomposition. H. Davy (1807) also
showed that the hydrogen and oxygen liberated during the decomposition of water
are in the proportions in which they combine to form water.
The experiment of W. Nicholson and A. Carlisle appears to have excited a great
deal of attention at the time, and many substances were treated
in a similar manner. This culminated in the brilliant discovery of
the alkali metals by H. Davy in 1807. J. W. Ritter's form of
apparatus, shown in Fig. 5, is the prototype of the many ingenious
forms of apparatus which have been devised for illustrating
W. Nicholson and A. Carlisle's experiment. In place of J. W.
Ritter's electrodes a and h, Fig. 5, adapted for the discharge from
an electric machine, plates of gold or platinum, in communication
with an accumulator or galvanic battery, are used. During the
passing of the electric current, bubbles of gas accumulate on the
metal plates and then rise into the test-tubes. More gas is given
off at one plate than the other. In fact, the volume of the oxygen
obtained approximates very closely to half the volume of the
hydrogen. The gas in each tube can be identified by means of a
lighted taper or otherwise. In the one tube, the taper burns with
the " blinding brilliance " characteristic of oxygen ; and the gas in
the other tube burns with the blue flame characteristic of hydrogen.
The water to be decomposed or electrolyzed is usually acidified
with a few drops of, say, hydrochloric or sulphuric acid. Some of
the water disappears during the electrolysis, but no change can be
detected in the amount of acid mixed with the water. Hence it is
inferred that the water, not the acid, has been decomposed. The experiment
succeeds equally well if a solution of sodium or potassium hydroxide be used with
Fio. 6.— Elec-
trolysis of
Water — Gases
mixed.
HYDROGEN AND THE COMPOSITION OF WATER 137
nickel or iron electrodes. Here again the water, not the alkali, is decomposed.
The acid or alkali is used because water alone does not conduct an electric current
very well. In fact, pure water is said to be a non-conductor of electricity.
Dilute solutions of acids or alkalies are good conductors. If iron electrodes are
used in the acidulated liquid much of the oxygen formed during the decomposition
of the water is used in oxidizing the metal.
A mixture of one volume of oxygen and two volumes of hydrogen, called electro-
lytic gas or detonating gas — A. Volta (1776) called it aura tonante — is often
wanted in gas analysis, etc. This is easily provided by placing both electrodes
under one receiver. The apparatus illustrated in Fig. 6 is often used for this
work — it explains itself. The outer jacket keeps the electrolyte cool. Many
forms of apparatus have been devised for the electrolytic preparation of small
quantities of hydrogen and also of the mixed electrolytic gas. 3
Are hydrogen and oxygen the sole products of the electrolysis of water ?
—Electrolytic oxygen often contains a little ozone and the electrolyte some hydrogen
peroxide ; both th^se compounds are formed by the electrolysis of acidulated
water, but not if a solution of barium hydroxide be electrolysed. Besides oxygen
and hydrogen, the early chemists noticed that an acid and an alkali are respectively
formed about the positive and negative poles during the electrolysis of water.
W. Cruickshank (1800) supposed the acid to be nitrous acid, and the alkah ammonia ;
J. B. Desormes (1801) considered that hydrochloric acid and ammonia were the
products ; while M. Brugnatelli (1802) explained the phenomenon by asserting
that it is the nature of electricity to produce these substances, and he called the
acid product electric acid. In 1807, Humphry Davy sought the origin of the acid
and the alkali, and published an account of his experiments in a most important
memoir entitled. On Some Chemical Agencies of Electricity . T. Thomson has styled
this investigation " the finest and completest specimen of inductive reasoning
which has appeared in the age in which Dav}^ lived."
While accepting H. Cavendish's demonstration that water is a compound of
oxygen and hydrogen, H. Davy considered the possibiHty that some product might
result from the unexpected decomposition of oxygen and hydrogen, and he then
divested the common experiment of every imaginable source of fallacy. It seemed
to H. Davy that the acid and alkali are most likely produced : (1) from the water ;
or (2) by the decomposition of the glass ; or (3) by the electrolysis of sodiimi
chloride derived from the hands touching the instruments ; or (4) from substances
derived from the ambient air which are decomposed by contact with the electrical
apparatus. Instead of conducting the electrolysis in glass vessels, Davy tried
vessels of gold, and by taking precautions to eliminate disturbances produced by
the contact of the vessels with the hands, and by the presence of impurities in the
water, H. Davy found that while an acid still continued to be formed, no alkali
appeared, and he showed that the alkali is derived from the solution of the glass
vessels during the electrolysis. H. Davy next conducted the electrolysis in an
atmosphere of hydrogen, and he then found that neither an acid nor an alkali was
developed, and hence he inferred that the acid which appears in the electrolyte is
derived from the nitrogen in the atmosphere. Consequently, when precautions
are taken to prevent the introduction of impurities from external sources,
no acid or alkali is produced during the electrolysis of water. It has since
been found that the volume of oxygen obtained during the electrolysis of a solution
of lithium, potassium, sodium, barium, or calcium hydroxide is sensibly less than
half that of the corresponding hydrogen. The hydrogen obtained is rather more
than double the volume of the oxygen when an electric current of low density is
used for the electrolysis.
It must be emphasized that the decomposition of water by the electric current
is not the same in kind as that produced by the disruptive discharge of an electric
machine in P. van Troostwijk and J. R. Deiman's experiment. In the latter case,
oxygen and hydrogen are evolved at both the poles dipping in the liquid, while in
138 INORGANIC AND THEORETICAL CHEMISTRY
the former case, oxygen is evolved at the one pole, and hydrogen at the other. The
electrolysis of water by the disruptive discharge is largely masked by the thermal
decomposition of the water (A. L. Lavoisier and M. Meusnier's experiment). J. W.
Ritter decomposed water the same year as W. Nicholson and A. Carlisle, but
apparently in ignorance of their work. J. W. Ritter modified
the experiment. He half-filled the two legs of a V-tube with
concentrated sulphuric acid, aa, Fig. 7, so as not to wet the sides
of the tube with acid ; he then carefully poured distilled water,
WW, into each leg of the tube so as not to disturb the acid, when
he found the water in the upper part of the legs of the tube did
not affect litmus paper. When gold wires, gg, connected with a
battery were dipped in the water, hydrogen collected at one pole,
oxygen at the other. J. W. Ritter said that the water in one leg
of the tube is not in communication with the water in the other
Fig. 7.— J. W. Hit- leg. He therefore inferred that water is an elementary body, and
ter's Experiment, that about the one pole water unites with negative electricity to
form oxygen, and with positive electricity to form hydrogen,
about the other pole. This conclusion conflicts with the evidence obtained when
water is decomposed by agents other than electricity, and it was explained by
Faraday's experiment on electrolysis.
The formula for water used to be written HO when the atomic weight of
hydrogen was taken unity, and oxygen 8. This agrees quite well with the deter-
minations of E. W. Morley and of J. B. A. Dumas. But we naturally ask for an
explanation of the result of the electrolysis of water. Does an atom of hydrogen
occupy twice the volume of an atom of oxygen ?
References.
^ G. B. Beccaria, Lcttere delVElettricismo, Bologna, 1758 ; G. Pearson, Phil. Trans., 87. 142,
1797 ; P. van Troostwijk and J. R. Deiman, Observations sur la physique, 35. 369, 1789.
2 W. Nicholson and A. Carlisle, Nicholson's Journ., 4. 179, 1800 ; H. Davy, Phil. Trans.,
97. 1, 1807 ; J. W. Ritter, Voigt's Mag., 2. 356, 1800 ; Gilbert's Ann., 9. 284, 1801 ; 10. 282, 1802.
' R. Bunsen, Gasometrische Methoden, Braunschweig, 72, 1857 ; 76, 80, 1877 ; A. W. Hofmann,
Ber., 2. 244, 1869 ; J. N. von Fuchs, Schweigger's Journ., 15, 494, 1815 ; J. W. Dobereiner,
Gilbert's Ann., 68. 55, 1821 ; A. Ehrenberg, Zeit. anal. Chem., 26. 226, 1887 ; E. W. Magruder,
Amer. Chem. Journ., 19. 810, 1897 ; J. L. Beeson, Journ. Amer. Chem. Soc, 26. 324, 1904 ; S. S.
Mereshkowsky, Centrb. Bakter., 11. ii. 786, 1904 ; M. Vezes and J. Labatut, Zeit. anorg. Chem.,
32. 464, 1902 ; J. J. Berzelius, Lehrbuch der Chemie, Dresden, 1. 185, 1825; M. BrugnateJli, Ann.
chimica, 18. 136, 1800; J. B. Desormes, Ann. Chim. Phys., (1), 37. 284, 1801; H. Davy, Phil.
Trans., 97. I, 1807; A. Volta, Letiera sulVaria inflammabile, M.ila,n, 1777; Strasbourg, 1778;
F. Richarz and C. Lonnes, Zeit. phys. Chem., 20. 145, 1896; Lord Rayleigh, Journ. Chem. Soc,
71. 181, 1897.
§ 7. Cavendish's Experiments on the Synthesis of Water by Volume
It is curious to note the changing fortunes of water in the history of chemistry. First
the matrix of the whole universe ; then only one of the four elements, though the chief of
the quaternion ; and at last discovered to be itself nothing but a liquid product of combustion,
one oxide among many, the mere ash, rust, or calx of so much burnt hydrogen.' — S. Brown
(1851).
From the earliest dawn of scientific speculation, water has been regarded by-
natural philosophers as one of the four primal elements, and they were quite right,
so far as their knowledge went, because they did not know how to decompose it
into simpler substances. The dogma had been reiterated so frequently that, down
to the days of the first French revolution, no one appears to have entertained any
doubts of the simple elementary nature of this liquid ; Basil Valentine called it
'* the mother of the metals." At the beginning of the seventeenth century the
sagacious J. B. van Helmont ^ planted a sprig of willow in a vessel suspended in
HYDROGEN AND THE COMPOSITION OF WATER 139
air, and fed it on nothing but water ; he found the plant to grow apace — new
branches, leaves, and roots sprouted forth. He said :
I placed 200 livres of dried soil in an earthenware pot, and planted therein a sprig of
willow weighing 5 livres. At the end of five years, the willow had increased to nearly
69 livres, 3 onces. The vase had never been watered with anything but rain water or dis-
tilled water.
Hence it was inferred that the constituents of plants — wood, foliage, acids, salts,
and earths — are embodied within elemental water in some mysterious inscrutable
way. The experiment seemed to him a crucial one, but J. Woodward's researches 2
showed the conclusion was fallacious because the parts played by the substances
dissolved in the water, and by the atmospheric air surrounding the plant, were not
recognized. The composite nature of water was not suspected until over a century
after J. B. van Helmont's time. Even the shrewd Robert Boyle in his Sce/ptical
Chymist (Oxford, 1661) lauded, beyond his predecessors, the importance of water :
It seems evident that water may be transmitted into all the other elements . . . not
only plants, but animals and minerals may be produced out of water.
Near the end of the seventeenth century, Isaac Newton 3 noticed that while
the refractive indices of various non-combustible substances increase proportionally
with their densities, the increase with the refractive indices of combustibles —
camphor, turpentine, oils — is greater than corresponds with their densities. " Water
has a refractive index in a middle degree between those two sorts of substances,
which consist as well of sulphureous, fat, and inflammable parts, as of earthy,
lean and alcalizate ones." After the compound nature of water had been discovered,
commentators read into Newton's statement a prediction that water would be found
to contain an inflammable substance as one of its constituents, although it may be
questioned if Newton intended to make any such assertion.
In 1782, J. Priestley * thought that he had proved that water is converted into " air
of the same purity as the atmosphere " by heating it in porous earthenware vessels so long
as there is free access of air to the outside of the retort, but he found the following year,
in agreement with a hint he had received from Josiah Wedgwood, that the supposed con-
version was a mal-observation, because the air was transmitted from the outside to the
interior through the pores of the retort.
In 1776, p. J. Macquer ^ noticed the formation of a liquid resembling water when
hydrogen burns in air, and the flame is allowed to impinge on a cold slab of porcelain.
Soon after the discovery of oxygen, J. Priestley (1775) noticed that when hydrogen
is mixed with certain proportions of oxygen, a violent detonation occurs when
ignited by a flame. In the spring of 1781, J. Priestley ^ made what he called " a
random experiment to entertain a few philosophical friends," in which a mixture
of inflammable air with dephlogisticated air or oxygen was exploded in a closed
vessel by means of an electric spark, as had been effected by A. Volta in 1777. The
sides of the glass vessel were found to be bedewed with moisture after the explosion.
Neither P. J. Macquer nor J. Priestley appears to have paid any particular attention,
at the time, to the phenomena ; they both seem to have thought that the dew
" was nothing else than the mechanical deposit of the moisture dispersed in common
air." According to J. Priestley, John Warltire repeated this experiment with a
copper vessel, and obtained a slight loss of weight which he thought might be due
to the escape of ponderable matter in the form of heat, through the pores of the vessel.
In the light of subsequent events, il est clair, said A. L. Lavoisier (1781), q^^e M.
Priestley a forme de Veau sans s'en douter. Meanwhile, H. C. Cavendish ^ looked
upon the deposition of the dew as a fact " well worth examining more closely " ;
H. Cavendish also wished to find what became of " the air lost " during the com-
bustion of hydrogen in common air. He tried (i) if the air had been changed into
carbon dioxide ; (ii) if it had been changed into nitric acid ; and (iii) if it had been
changed into sulphuric acid. He negatived these hypotheses one by one. In the
summer of 1781, H. Cavendish followed up the subject by exploding mixtures of
140 INORGANIC AND THEORETICAL CHEMISTRY
dephlogisticated air with inflammable air in closed vessels. A certain amount of
the gaseous mixture lost its elastic form, and produced a certain amount of liquid
water. In the fourth experiment on exploding gases — dated July 5th, 1781, in his
laboratory notebook — H. Cavendish demonstrated the relations between the
volumes of inflammable air and common air consumed in the formation of water,
for he showed that by exploding a mixture of 7,344 volumes of inflammable air
(from zinc and an acid) with 17,361 volumes of common air, there was a contraction
of 10,630 volumes, and a gas -^-^^d of the specific gravity of common air remained.
Before the end of the month, H. Cavendish had proved the liquid product to be
pure water, for his notebook says :
The liquid was not at all acid, nor gave the least red colour to paper tinged with red
flowers, it yielded no pungent fumes on evaporation, and yielded scarce any sediment
on evaporation to dryness.
H. Cavendish stated his conclusion from these experiments when they were
described in his paper Experiments on Air (London, 1784) :
When inflammable air (hydrogen) and common air are exploded in proper proportion,
almost all the inflammable air, and near one-fifth of the common air, lose their elasticity,
and are condensed into dew. And by this experiment it appears that this dew is plain
water, and consequently that almost all the inflammable air and about one-fifth of the
common air, are turned into pure water.
Cavendish repeated the experiment with a mixture of inflammable air (hydrogen)
with nearly twice its volume of pure dephlogisticated air (oxygen), and found that
almost the whole of the mixture in the globe formed pure water ; a quantity of
water was collected by repeatedly introducing more gas into the globe and exploding
the mixture. The vessel and its contents underwent no change in weight or parted
with anything ponderable during the explosion, while a certain volume of gas was
replaced by a certain weight of water. Hence, as A. L. Lavoisier ^ has expressed
it : Veau n'est point une substance simple, et qu'elle est composee poids pour poids d'air
inflammable et d'air vital — otherwise expressed, water consists, weight for weight,
of the hydrogen and oxygen gases lost in its production. The results of H. Caven-
dish's experiments, 1781-2, were communicated to J. Priestley not later than March,
1783, and also to A. L. Lavoisier in June, 1783, and published in 1784 ; the delay
in publication was occasioned by the need for investigating the puzzling appearance
of nitric acid along with water when oxygen was substituted for atmospheric air.
There appears to have been some extensive alterations in Lavoisier's paper before
it was pubUshed, but there is no means of determining precisely the extent of the
additions. J. Watt wrote a letter to J. Priestley, April 26th, 1783, containing an
outline of a theory of the composition of water, and on June 19th, 1783, Joseph
Priestley ^ read a paper before the Royal Society in which he stated in reference
to the bedewed glass in his experiment :
I carefully weighed a piece of filter paper, and then, having wiped with it all the inside
of the glass, weighed it again, and always found, as nearly as I could judge, the weight of
the decomposed air in the moisture acquired by the paper,
H. Cavendish's public statement that he had previously communicated to
J. Priestley every experiment which was needed to determine the composition of
water was publicly acknowledged by J. Priestley. These facts have never been
impugned, and they are supported by the entry in the Minute Book of the Royal
Societ}^ 10 which was confirmed at the meeting on the 26th June, 1783. J. Priestley
said :
These arguments received no small confirmation from an experiment of Mr. Cavendish,
tending to prove that the reconversion of air into water, in which pure dephlogisticated air
and inflammable air were decomposed by an electric explosion, and yielded a deposit of
water equal in weight to the decomposed air.
The work of H. Cavendish was soon confirmed by M. Monge,ii in a memoir
Sur le resultat de Vinflammation du gas inflammable et de lair dephlogistique dans des
HYDROGEN AND THE COMPOSITION OF WATER 141
vaisseaux clos. M. Monge exploded measured volumes of hydrogen and oxygen
in an exhausted glass globe, and by admitting fresh quantities of gas for explosion,
he collected a relatively large amount of water. He calculated the weight of the
original gases from their known densities, and weighed the liquid product. The
results showed :
L'air inflammable .
L'air dephlogistique
Total weights of components
Total weight of product .
Deficit .
Onces. Gros. Grains.
6 10-03
3 0 58-53
3 6 68-56
3 5 101
1 67-55
Owing to the use of moist gas, M. Monge over-estimated the weight of the
hydrogen which he had employed, and consequently there was a small deficiency
between the observed weight of water and the estimated weights of gas required
for the synthesis. The water produced was very slightly acid, and he assumed
that the acidity is due to " the small quantity of vitriolic (sulphuric) acid which
inflammable air carries when prepared by the dissolution of iron " in that acid ; H.
Cavendish had already proved that nitric acid is a by-product of the reaction
under certain conditions. M. Monge concluded (i) that the volume of hydrogen
required for the formation of water is about twice as great as that of the oxygen ;
and (ii) that " when inflammable air and dephlogisticated air, both pure, are
exploded, there is no other product but pure water, heat, and light." Experiments
similar in principle, but with highly purified materials, were made by A. Scott
and E. W. Morley ^^ over a century later.
J. Watt's claims to the first trae conception of water. — In 1783, James Watt,i3
of engineering fame, expressed the opinion that " according to J. Priestley's experi-
ments, dephlogisticated air unites completely with about twice its bulk of inflam-
mable air . . . and therefore water is composed of dephlogisticated air and phlo-
giston." It is possible, though doubtful, that J. Watt had in mind inflammable
air or hydrogen when he used the term phlogiston, and by dephlogisticated air,
what is now called oxygen. In his Thoughts on the constituent parts of water and
of dephlogisticated air communicated to the Royal Society, November, 1783,
J. Watt said that he was convinced by the arguments of R. Kirwan and J. Priestley
that inflammable air is either wholly pure phlogiston or at least that it contains no
admixture of any other matter ; but he added that in his opinion inflammable air
contains a small quantity of water and much elementary heat. He regarded heat
as a material substance, and invested it with the capacity of combining with other
substances like other material elements, and of becoming the basis of those sensible
qualities by which bodies are permanently distinguished from each other. According
to his theory, dephlogisticated air is composed of water deprived of its phlogiston
and united to elementary heat. He believed that dephlogisticated air and phlogiston
can unite in certain degrees to form, not water, but fixed air, while under other
circumstances they can unite to form neither water nor fixed air, but rather
phlogisticated air. In spite of this, it has been claimed that J. Watt was the first
to form the conception that water is a compound of dephlogisticated and inflam-
mable air, and that H. Cavendish made the proposition good by unassailable ex-
periments. Naturally, J. Watt's claims have been disputed, and the so-called waier
controversy has been waged upon the rival claims of J. Watt, A. L. Lavoisier, and
H. Cavendish.
The controversy is exceedingly involved. The three rival claimants almost
simultaneously arrived at analogous conclusions by different paths. H.
Cavendish was at work on the products of the combustion of hydrogen ;
J. Watt was speculating on the latent heat of steam ; and A. L. Lavoisier
was studying the production of acids by the oxidization of inflammable substances.
All these paths ultimately converged into the one line of inquiry which culminated
142 INORGANIC AND THEORETICAL CHEMISTRY
in the discovery that water is a compound of hydrogen and oxygen. The issue is
confused by the fact that, while the date of publication or receipt by a scientific
society is usually taken to be decisive in questions of priority, this is not always
satisfactory. In the present case, the observed results of the one were communicated
to others before they were published, and alterations were made in some of the
original papers, after they had been read and before they were published. Lord
Jeffrey has shown that the case cannot be decided by those narrow and jealous
canons of evidence derived from the rigid maxims of law or the precedents in cases
of patent :
Courts of law must proceed on inflexible rules, and can make no distinction of persons ;
and are forced therefore peremptorily to reject all evidence proceeding from the parties
concerned, or from those having any interest in the issue ; though it is certain by so doing
they must occasionally decide against the truth, and against the conviction of all unpro-
fessional observers. The question m a court of law, in short, is never really what the truth
of a case is, according to the actual and conscientioiis belief of the judges or jury, after
considering every atom of producible evidence that is in existence, but merely what is the
import of the evidence that is legally admissible. ... In all questions before the public
no evidence is inadmissible.
J. Priestley tried to repeat the experiment on the formation of water which he
said had been described to him by H. Cavendish. In order to ensure the absence of
moisture, Priestley prepared his dephlogisticated air from nitre ; and his inflammable
air, by heating what he called " perfectly made charcoal " in a retort. The gas from
the charcoal would obviously be obtained by the diffusion of air and furnace gases
through the walls of the retort and the reaction between these gases and the charcoal.
J. Priestley failed because, through an extraordinary blunder, the wrong inflammable
air was used. According to W. V. Harcourt (1846), " neither the phlogiston nor
the inflammable air of Priestley and Watt were convertible terms for hydrogen,
their notion of the change of air into water, and of water into air, had no reference
to hydrogen, but first to nitrogen, and afterwards to a mixture of gases, the chief
of which was carbon monoxide. J. Priestley's paper was communicated to the Royal
Society on April 19th, 1783 ; H. Cavendish's communication to J. Priestley must
therefore have been anterior to the speculation which J. Watt addressed to J.
Priestley on the 26th of the same month, as well as to Lavoisier's experiments the
following June." J. Priestley thus comes as an intermediate link, for through him
an account of the experiments and conclusions of H. Cavendish were transferred to
Watt.
Lavoisier's claims to the discovery of the composition of water. — For a time,
some claimed A. L. Lavoisier to have discovered the composition of water inde-
pendently of H. Cavendish. According to M. Berthelot, the laboratory journal shows
that as early as March, 1774, the attention of A. L. Lavoisier was directed to the
product of the combustion of hydrogen since he believed that every inflammation
ought to be attended by an increase in weight, and in 1777, he burnt hydrogen in
air, and, in 1781, oxygen in hydrogen ; but A. L. Lavoisier's mind was preoccupied
with the conviction that oxidation means acidification, and the production of water,
which must have occurred, seems to have passed unheeded. Lavoisier said after-
wards that he did not then know about Macquer's experiment. A. L. Lavoisier
was always on the alert as to the nature of the products of the combustion of
hydrogen, and in 1783 he was in such a position that the slightest hint would enable
him to comprehend its true nature. This hint was furnished by the rumours of
H. Cavendish's experiment which spread through the scientific world in the spring
of 1783. C. Blagden communicated the result of H. Cavendish's experiment to
A. L. Lavoisier, at Paris, on June 24th, 1783. A. L. Lavoisier confirmed the fact
with a single hasty experiment made a few days after C. Blagden's communication,
and described before the French Academy — partly in November and partly in
December, 1783. Although the account of H. Cavendish's Experiments on Air was
not read before the Royal Society until January 15th, 1784, Lavoisier i* said that on
HYDROGEN AND THE COMPOSITION OF WATER
143
June 24th, 1783, " Mr. Blagden has informed us that Mr. Cavendish had burnt
inflammable air in closed vessels, and that he had obtained a very sensible quantity
of water." Consequently, H. Cavendish was undoubtedly first in the field, and
he furnished his rivals with the grounds of their conclusions — J. Watt through
J. Priestley, A. L. Lavoisier through C. Blagden.
H. Cavendish was certainly not clear about the character of the reaction involved
in his synthesis because his mind was unconsciously mystified by the phlogiston
hypothesis. He seems to have rather incUned to the opinion that the indifference
of hydrogen to oxygen at ordinary temperatures impUed the presence of some
substance in the former which lessened the intensity of its affinity for oxygen, and
he conceived that this substance could be water alone, because water is the sole
residue of the combustion of hydrogen and oxygen. Thus, H. Cavendish remarked :
From what has been said, there seems the utmost reason to think that dephlogisticated
air is only water deprived of its phlogiston, and that inflammable air, as was before said,
is either phlogisticated water, or else pure phlogiston, but in all probability the former.
The indifference of free hydrogen and oxygen to one another at ordinary tempera-
tures was a source of perplexity to others besides H. Cavendish. Thus, J. Watt is
said :
Priestley accounts for the facts by supposing that the two kinds of air, when formed
at the same time and in the same vessel, can unite in their nascent state ; but that, when
fully formed they are incapable of acting upon one another, unless they are first set in
motion by external heat.
It certainly required Lavoisier's system to give a significance to Cavendish's capital
discovery, by showing that water is a definite oxide or calx of hydrogen formed
whenever hydrogen is burnt in air or oxygen.
The synthesis of liquid water. — The following is a modernized form of
H. Cavendish's elegant experiment, although it is not any more demonstrative. In
Cavendish's original experiment, the explosion vessel was weighed before and after
the gases were exploded :
A stout glass vessel, A, Fig. 8, is fitted with a stopcock, C, at one end, and with a piece
of strong pressure tubing, D, con
nected with a reservoir, at the other
end. A pair of platinum wires, T,
are sealed into the stout glass measur-
ing vessel just below the stopcock.
These wires are put in commimica-
tion with an induction coil, which
in turn is connected with an accumu-
lator. The tube A is called the
eudiometer^ or the explosion tube.
This is filled with mercury by ad-
justing the levelling tube B and the
stopcock C. A mixture containing one
volume of oxygen and two volumes
of hydrogen is introduced into the
explosion tube vid the stopcock G
and by depressing the levelling tube.
When the explosion tube is about half
or three-foiu*ths filled, read the volume
of its contents by bringing the mercury
to the same level in both levelling tube
and explosion tube. Then depress the
levelling tube so that the mercury falls
nearly to the bottom of the explosion
tube. Pass a spark from the induction
coil through the wire terminals of the
explosion tube. The gases explode,
and the level of the mercury is again • u . x
adjusted after the apparatus has stood for a few minutes m order to regain the temperature
The mercurv rises nearly to the level of the stopcock. The mixed gas probably
Explosion
Tube "
Fig,
8. —Modern Form of Synthesis of Liquid Water
by Volume — Cavendish's Experiment.
of the room.
144
INORGANIC AND THEORETICAL CHEMISTRY
contained a trace of air, and probably also a slight excess of either oxygen or hydrogen. The
advantage of this form of explosion vessel lies in the fact that the explosion takes place under
diminished pressure, and is not so liable to fracture the apparatus because it is less
violent.
The result shows that two volumes of hydrogen unite with one volume of oxygen
to form water. Suppose the experiment he repeated a number of times with, say,
one volume of oxygen and three volumes of hydrogen — one volume
of hydrogen remains after the explosion ; again try the experiment with
two volumes of oxygen and two volumes of hydrogen — one volume
will remain uncombined after the explosion. It is inferred from this
experiment, that two volumes of hydrogen and one volume of oxygen
combine to form water, and if an excess of either oxygen or
hydrogen be present, the excess will remain uncombined after the
reaction.
Gas analysis. — If a known volume of gas containing hydrogen be
mixed with an excess of air or oxygen ; or if a known volume of a
gas containing oxygen be mixed with an excess of hydrogen and
exploded in a eudiometer, the contraction represents the volume of
water formed, and the corresponding amount of the gas under
investigation can be computed. A. Volta i^ utilized these .facts in
devising a process to estimate the two gases. A metal cap, B, was
fitted to the upper part of a graduated tube, A, Fig. 9, which con-
stituted J. Priestley's eudiometer. The metal cap carries an insulated
wire, C, which enabled a spark to be passed in the interior of the tube.
A rubber ring, Z), was used in reading the level of the liquid in the
tube. The funnel, F, connected with the stopcock. E, was used in
fiUing the eudiometer with gas in the pneumatic trough. The Hmits of
explosibility of mixtures of hydrogen and oxygen are approximately Hydrogen :
Oxygen=5'4 : 94'6 ; and 94*7 : 5'3. No explosion occurs if the proportions of
these two gases are outside these limits.
Fig. 9—
Vol t a's
Eudiometer,
Example. — 20 c.c. of air were mixed with 20 c.c. of hydrogen and exploded. The
mixed gases, after the explosion, occupied 28 c.c. Hence, the contraction shows that 12 c.c.
of the mixture combined to form water. Of this two -thirds must have been hydrogen, and
one-third oxygen. Hence, the original 20 c.c. of air contained 4 c.c. {i.e. one-third of 12 c.c.
of oxygen). This illustrates an important principle used in gas analysis.
J. Priestley was led astray by the presence of nitric acid in the water formed
by the union of hydrogen with oxygen. According to H. Cavendish's notebooks,!^
he found in September, 1781, that the liquid formed by exploding oxygen
with twice its volume of hydrogen contained some nitric acid. H. Cavendish also
found that this acid was obtained whether the oxygen was prepared from mercury
nitrate, from mercuric oxide, or from plants under the action of solar light ; and
consequently he inferred that the nitric acid was not present as an impurity in the
oxygen. In January, 1783, he showed that, if hydrogen is burnt in the presence
of an excess of oxygen slightly contaminated with nitrogen, the excess of oxygen
unites with the nitrogen forming nitric acid ; but if the hydrogen is burnt with
oxygen mixed with a large proportion of nitrogen, " the heat of the explosion is
80 much diminished that though the affinities of hydrogen and oxygen are sufficient
to determine at that temperature the formation of water, the affinities of nitrogen
and oxygen are not sufficient to determine the production of nitric acid." H.
Cavendish thus demonstrated that the only product of the explosion of hydrogen
and oxygen is water.
E.. Bunsen (1857) ^^ noticed that when electrolytic gas is exploded with air, some
nitric oxide is formed, and if an excess of oxygen be present, some nitrogen peroxide
is also formed. According to K. Finckh (1905), the amount of nitric oxide so
HYDROGEN AND THE COMPOSITION OF WATER
145
formed depends upon the temperature and pressure of the admixed gases. For
instance,
Initial pressure of mercury
Electrolytic gas per 100 vols, air
Nitric oxide formed
450
460
455
88
124
220
0-22
1-02
2-45
Le^effina
Tube.
750 mm.
210 vols.
3*01 per cent.
To reduce the proportion of nitrogen oxides formed during the explosion of hydrogen
(or hydrocarbons) with air, R. Bunsen found it best to keep the amount of pure
hydrogen between 3'81 and 1-55 per cent., for
the resulting error is then negligibly small, If
wider eudiometer tubes than those employed
by R. Bunsen are used, these limits must
be raised. According to A. SchuUer (1882),
when hydrogen is exploded with an excess
of oxygen, some hydrogen peroxide is formed
at the same time.
The volumetric synthesis of steam.—
When hydrogen unites with oxygen to form
water, is the product equal to the joint
volume of the constituents when measured
in the same state of aggregation, without
allowing the gaseous water to condense to
the liquid state ? Water is a gas — often
called steam — when its temperature is a
little above 100° at ordinary atmospheric
pressures. In 1865 A. W. Hofmann modified
an old experiment of J. L. Gay Lussac (1808)
by placing a hot vapour jacket about the
explosion tube so that the water remains Fig. 10. — Synthesis of Steam by Volume,
in the gaseous condition and does not con-
dense to a liquid after the explosion. A. W. Hofmann's experiment was described
in his Introduction to Modern Chemistry (London, 1865), and a modification is
illustrated in Fig. 10.
The upper end of the glass jacket surrounding the explosion tube. Fig. 10, is connected
with a flask, M, containing toluene, boiling at about 110°, or amyl alcohol, boiling at about
130°. The lower end of the jacket is connected with a flask and condenser, N, so that the
amyl alcohol can be recovered. When the amyl alcohol is steadily boiling, and the ex-
plosion tube has been filled as described in the preceding experiment, the gases are sparked.
In a few minutes, when the temperature has had time to adjust itself, bring the levelling
tube in position for a reading.
The result of this experiment is to demonstrate that two volumes of hydrogen
unite with one volume of oxygen to form two volumes of steam, for the steam
occupies just two-thirds the original volume of the mixed gases. Hence, A. W.
Hofmann's form of J. L. Gay Lussac's experiment demonstrates that when water
is synthesized at a temperature above its point of condensation — 100° — two
volumes of hydrogen react with one volume of oxygen to form two volumes
of steam. Several types of chemical problems are based on this fact. It is necessary
to correlate the different results obtained when water is synthesized by volume
and by weight.
References.
1 J. B. van Helmont, Orfus Medicince, Amsterdam, 1648; Lugduni Batavorum, 68, 1656.
2 J. Woodward, Phil. Trans., 21. 193, 1699 ; H. Braconnot, Ann. Chim. Phys., (1), 61. 187,
1807.
* Isaac Newton, Opticas, London, 75. 1704,
4 J. Priestley, Phil. Trans., 23. 426, 1783.
VOL. I. L
146 INORGANIC AND THEORETICAL CHEMISTRY
* A. L. Lavoisier, CEuvres, Paris, 2. 335, 1862 ; P. J . Macquer, Dictionnaire de chimie, Paris,
2. 314, 1778.
* J. Priestley, Experiments and Observations on Different Kinds of Airy London, 2. 30, 1775 :
3. 382, 1777 ; 5. 395, 1781.
' H. C. Cavendish, Phil. Trans., 74. 119, 176, 1784 ; 75. 372, 1785 ; Alembic Club Reprints, 3 ;
1893 ; R. Kirwan, Phil. Trans., 74. 154, 1784.
« M. Lavoisier, Mim. Acad., 473, 1781 (printed 1784).
» J. Priestley, Phil. Trans., 73. 414, 1783; Experiments on Air, Birmingham, 6. 29, 1780.
i» W. V. Harcourt, B. A. Rep., 22, 1839.
" M. Monge, Mem. Acad., 78, 1786.
12 A.. Scott, Phil. Trans., 184. 643, 1893 ; E. W. Morley, Zeit. phys. Chem., 20. 68, 242, 417,
1895.
" G. Wilson, The Life of the Honorable Henry Cavendish, London, 265-446, 1851 ; H. Kopp,
Beiirage zur Geschichte der Chemie, Braunschweig, 1876 ; E. Grimaux, Lavoisier, 1743-1794,
Paris, 1888 ; T. E. Thorpe, B. A. Rep., 761, 1890 ; M. Berthelot, La revolution chimique, Paris,
1890 ; Notice historique sur Lavoisier, Paris, 1889 ; J. P. Muirhead, Correspondence of the late
James Watt on his discovery of the theory of the composition of water, liondon, 1846 ; Lord Brougham,
Lives of the Philosophers of the time of George III., London, 1855 ; W. V. Harcourt, B. A. Rep.,
22, 1839; Phil. Mag., (3), 28. 106, 478, 1846; J, W&tt, Phil. Trans., 74. 329, 1784; Anon.,
Quart. Rev., 77. 105, 1846; Lord Jeffrey, Edin. Rev., 57, 1848.
1* A. L, Lavoisier, Mem. Acad., 468, 1784.
15 .J. Watt, Phil. Trans., 74. 334, 1784.
16 A. Volta, Ann. Chimica, 1. 171, 1790 ; 2. 26, 1791 ; 3. 36, 1791.
17 W. V. Harcourt, B. A. Rep., 1, 1839 ; H. Cavendish, Phil. Trans., 74. 130, 1784.
1^ R. Bunsen, Gasometrische Methoden, Braunschweig, 72, 1857 ; K. Finckh, Zeit. anorg. Chem.,
45. 116, 1905 ; A. Schuller, Wied. Ann., 15. 290, 1882.
CHAPTER IV
TEE PHYSICAL PROPERTIES OF OASES
§ 1. The Atmosphere
The atmosphere in which we live and breathe is really a part of the globe on which we
stand. We are not surrounded by mere empty space. On the contrary, we live and move
at the bottom of a vast ocean of air, which is just as material as the water which surrounds
the flat-fish living at the bottom of the sea {1914)."
Air was once considered to be a thin, pellucid, evanescent, inscrutable, and im-
ponderable spirit — the spirit of life. Even to-day, air is still used as a symbol
for what is spiritual and divine ; but to early man the analogy between the im-
palpable breath of the physical heavens and the inscrutable spirit of God, was
very real. It was quite a long time before air was recognized to be a gravic
material essentially ponderable like earth and sea.^
The physical properties of air were studied long before its chemical properties
were investigated. Anaxagoras, who lived about the sixth century B.C., cited two
experiments to show that air is material : (i) A blown bladder resists compression,
and (ii) the inside of an inverted drinking glass when plunged beneath the surface
of water remained dry showing that the presence of air prevented the ingress of
the water. These are among the earliest experiments on record. Aristotle (b.c.
384), in spite of some confused ideas on the nature of gases, considered air to be a
material substance which possessed weight, because he found that a blown bladder
weighed less when empty than when inflated with air. Simphcius, a writer of the sixth
century, commenting on Aristotle, said that Ptolemy showed that air has no weight
when weighed in air, and that Aristotle's conclusion was vitiated by the condensa-
tion of moisture in the bladder derived from the air blown from the lungs during
the inflation of the bladder. About a century before Christ, Hero of Alexandria,
in an important work on Pneumatics, described some experiments to prove that
air is a material substance. For instance, he said :
Let a vessel which seems to be empty be inverted, and, being carefully kept upright,
pressed down into the water ; the water will not enter it even though it be entirely immersed ;
so that it is manifest that the air, being matter, and having itself filled all the space in the
vessel, does not allow the water to enter. Now if we bore the bottom of the vessel, the water
will enter through the mouth, but the air will escape through the hole. Again, if before
perforating the bottom, we raise the vessel vertically, and turn it up, we shall find the inner
surface of the vessel entirely free from moistiu-e, exactly as it was before immersion. Hence,
it must be assumed that the air is matter.
A similar experiment was mentioned by Empedocles 2 (c%Vca 430 B.C.), and a
correct explanation given. The same experiment was also described in the essay,
De ingeniis spiritualibus, by Philo of Byzantium, about 300 B.C.
The weight of air.— In his Book of the balance of wisdom, written in the fifteenth
century, the Arabian Al-Khazoni recognized clearly that air has weight. He said :
When a heavy body of whatever substance is transferred from a rarer to a denser air,
it becomes lighter in weight ; and when transferred from a denser to a rarer air, it becomes
heavier. . . . Although the weight of a substance in air does not appear to vary, there is
an actual variation, owing to a difference of atmospheres at different times.
However, GaHleo dei Galilei, in 1632, is usuaUy credited with having first demon-
strated satisfactorily that air possesses weight ; and he made a rough determmation
147
148 INORGANIC AND THEORETICAL CHEMISTRY
of the specific gravity of air by comparing the relative weights of equal volumes of
air and water. G. Galilei found water to be 400 times heavier than air ; and twenty
or thirty years later, R. Boyle (1661) found water to be 938 times heavier than air.
Both measurements were very crude, and are quite unreUable; G. Galilei's result
is too low, R, Boyle's too high. Refined experiments show that 1000 c.c. of dry
air weigh 1*293 grms. under standard conditions — 760 mm. pressure, 0°, and at
sea level in latitude 45°. Hence the specific gravity of air is 0001293 if water be
unity. This means that a normal litre of dry air freed from carbon dioxide and at
0° and 760 mm. weighs 1-2930 grams at sea-level and a latitude of 45°. The
actual numbers are : 1 29276 (H. V. Regnault, 1847) ; 1-293085 (P. von JoUey,
1879) ; 1-29284 (Lord Rayleigh, 1888-93) ; 1*29273 (A. Leduc, 1898) ; 1*2930
(P. A. Guye, J. Kovacs, and E. Wourtsel, 1912) — vide Cap. on atmospheric air.
The accidental or experimental errors affecting the number 1-2930 amount to
less than one in ten thousand. The variations which have been observed show
that the density of air is not constant but variable both with respect to place and
time. This conclusion is in harmony with the variations which have been observed
on the relative proportions of oxygen and nitrogen in air. Thus, P. A. Guye, J.
Kovacs, and E. Wourtsel found the weight of a normal litre of air, collected during
a rising barometric pressure, to be 1-2927 grm., and 1*2932 grm. when collected
during a falUng barometric pressure. The former number is taken to mean that
the air has a shght deficit in the proportion of oxygen, and the latter, a sHght deficit
in the proportion of nitrogen — when the normal Utre is taken as 1-2930 grms. The
specific gravity of air, referred to the standard hydrogen 2, is taken to be 28*75 ;
or it oxygen 32 be the standard, 28*95.
The terms atmosphere and air are sometimes taken to be synonymous and interchange-
able, but the word air is often used when reference is made to a limited portion of the
atmosphere. The word air was formerly used in the same general sense that the word gas
is to-day. Later, the meaning of the word air was narrowed to connote the atmosphere.
The word atmosphere is derived from the Greek aTfx6s, vapour ; acpaipa, the sphere.
The term atmosphere is also applied to the gaseous envelope or medium surroiinding any
body, whatever be the nature of the gas- — air, oxygen, carbon dioxide, etc. Hence the term
atmospheric air is often used to emphasize the fact that air is the enveloping medium.
Both Anaxagoras and Aristotle believed that there is no vacuum and this belief
crystallized into the phrase : Nature abhors a vacuum. For instance, when a
glass cyHnder, closed at one end, is filled with water ; then closed at the open end
with the hand ; turned upside down ; and the hand removed while the open end of
the cyHnder is under water, the water remains in the cyHnder. The rise of water
in pump barrels was explained by the same hypothesis. When it was found that
water could not be pumped higher than about 34 ft., it followed that the hypothesis
required modification, for nature's horror of a vacuum obviously could extend
only to the equivalent of 34 ft. of water.
The pressure of the air.— In 1644, E. TorriceUi,^ a pupil of G. Galilei, pubHshed
an account of an experiment which puzzled the philosophers of the time because
they were obsessed by the hypothesis that nature abhorred a vacuum.
In E. Torricelli's experiment, a glass tube — about four feet long, and closed at one end
■ — was filled with mercury, the open end was closed with the thumb, and the tube inverted
so. that, when the thumb was removed, the open end was immersed in mercury . No air was
allowed to enter the tube during the operation. Instead of the mercury remaining suspended
in the tube, the column of mercury fell to such an extent that its height above the surface
of the mercury in the dish was nearly 30 inches, or 760 mm. The vacuous space in the tube
above the mercury is called Torricelli's vacuum.
Nature's horror of the vacuum at the top of the tube extended only to the equivalent
of 30 inches of mercury. It did not appear probable that nature should have a
particular whim of this character, and E. Torricelli suggested the alternative hypo-
thesis that the column of mercury was maintained by the air pressing on the surface
of the 7nercury in the outer vessel. B. Pascal, in his New experiments concerning the
vacuum (1647), argued that since mercury is nearly 13 J times as heavy as water.
THE PHYSICAL PROPERTIES OF GASES 149
30 inches of mercury will be equivalent to 34 ft. of water, and he accordingly-
repeated E. Torricelli's experiment with a tube 46 ft. long, using water instead of
mercury. He obtained a column of water 34 ft. long. When the experiment was
repeated with other liquids, he found, in every case, that the height of the column
was inversely as the density of the liquid. Hence, it was inferred that the height
of the column of mercury is a measure o{ the pressure of the atmosphere, and that
fluctuations in the pressure of the air are accompanied by a corresponding rise or
fall in the column of mercury. R. Boyle (1665) apphed the term barometer to
Torricelli's instrument — from the Greek ^dpo^, weight ; and fxirpov, a measure.
In 1647, B. Pascal persuaded M. Perier to repeat Torricelli's experiment at the
bottom and at the summit of the mountain Puy-de-D6me. On September 23rd,
1648, M. Perier wrote that the result nous ravit tous d'admiration et d' etonnement,
for the mercury sank lower in the tube the higher up the mountain the vessel was
carried. This confirmation of what was anticipated by Torricelli's hypothesis was
taken to prove that the pressure of the air per sq. cm. is greater at the bottom
than on the top of the mountain, and not as Aristotle and his followers would teach
that Nature has a greater horror of a vacuum at sea-level than at higher altitudes.
In a posthumous work, On the weight of the mass of air, published in 1663, B. Pascal
summarized arguments which proved conclusively that all those effects, previously
attributed to Nature's horror of a vacuum, are really produced by the pressure, that is,
by the weight of the air.
After the discovery of Torricelli's vacuum a group of philosophers — Thomas Hobbes,
Franciscus Linus,* etc. — refused to abandon a favourite hypothesis they had formed that
the world is everywhere full and a vacuum is impossible. They were called plenists in
contradistinction to the vacuuists — O. von Guericke, B. Pascal, Robert Boyle, ^ etc. — who
believed that a vacuum was possible, and capable of being obtained by certain physical
processes. A controversy followed, not always in the choicest of language ; thus, Thomas
Hobbes, addressing Drs. Ward and Wallis, said :
But I here dismiss you both together. So go your ways, you uncivil Ecclesiastics,
inhumane Divines, Dedoctors of morality, unasinous Collegues, egregious pair of
Issachars, most wretched Vindices and Indices Academiarum, and remember Vespasian's
law {maledici senatoribus non opportere ; remaledicere fas et civile esse) that it is uncivil
to give ill language first, but civil and lawful to return it.
The facts finally conquered an erroneous hypothesis.
Units o! pressure. — The pressure of the air in any given locality varies within
comparatively narrow limits. The normal or standard pressure of the atmosphere
is equal to the weight of a column of mercury of unit area, and 760 mm. high. This
pressure is sometimes called " one atmosphere." It is merely necessary to know
the height of the barometric column to know the weight or pressure of the air per
unit sectional area. The standard corresponds with a weight of (76 X 13*596 =)
1033-3 grms. per sq. cm., or 14' 7 lbs. per sq. ip. The word pressure is generally used
in preference to weight, because air, like all other fluids, not only presses down-
wards, but also equally in all other directions.
The selection of the atmosphere as the unit of pressure is quite arbitrary, and
other units are used — e.g. the kilogram per sq. cm., and the pound per sq. in. The
pressure of a dyne per sq. cm. was recommended by the International Physics
Congress at Paris in 1900, because it is consistent with the C.G.S. system of units.
This unit was called a barie ; a similar unit, the barad, was proposed by a com-
mittee of the British Association in 1888, and there has been some controversy as
to whether the unit had better be referred to a dyne per sq. cm. or to a pressure a
million times greater. The density of mercury is 13*596, and in latitude 45° the force
of gravity is equivalent to 980*6 dynes. Hence, a barometer column 76 cm. high
will be maintained by a pressure equivalent to 76xl3-596=1033-3 grms., or 1033*3
X980*6=l,013,300 dynes per sq. cm., or in round numbers, 10^ dynes per sq. cm.
This number — called a megabar— may be inconveniently large, and a ten-thousandth
part of 10^ is called a bar, hence, a bar is equivalent to 100 dynes per sq. cm. ; a
150 INORGANIC AND THEORETICAL CHEMISTRY
centibar to one dyne per sq. cm.; and a millibar to 01 dyne per sq. cm. This
unit, the millibar, has been recommended for recording barometer readings. One
megabar is equivalent to 750 mm. of mercury, under standard conditions. The
approximation is correct to one part in 5000. Since 13"596x980"6=13332 ;
and one ten- thousandth of this is 1*3332, it follows that to convert centimetres
into bars, multiply by 133'38 ; and, to convert bars into centimetres, multiply by
0-0075. Since there are nearly 2-54 cms. in an inch, 133'33 X 2-54 =338-63, therefore
to convert inches into bars, multiply hy 338'63. A pressure of one megabar is almost
2 per cent, greater than a kilogram per sq. cm. ; and 1*3 per cent, less than the
atmosphere unit.
The extent of the atmosphere. — The air gets less and less dense at higher and
higher altitudes, and I. Newton (1704) estimated air to be four times rarer at an
elevation of about 7 J miles than at sea level ; 1,000,000 times rarer at a height of
76 miles ; and 1,000,000,000,000,000,000 times rarer at an altitude of 228 miles ;
and so on. If ^q be the pressure and Dq the density of air at sea level, E. Halley's
formula 6 becomes
_^
Pressure of air at an altitude h = p^e, Po
Variations in the value of the gravitation constant, g, and the rotation of the
planet are neglected. Under actual conditions, the earth's atmosphere is in-
cessantly agitated by convection currents — winds and storms — so that there is a
continual transfer of air from one part to another. The limiting height at which
the atmosphere is in convective equilibrium is about 29 kilometres, and the tem-
perature falls roughly about 10° per kilometre as we ascend. Above this region,
the temperature of the air is constant. It is indeed impossible to place a limit to the
height the atmosphere extends. G. J. Stoney showed that, because the molecules
of some gases attain certain high velocities, these gases are able to escape from the
atmosphere of the earth and the other planets. At a height of 100 to 125 miles, there
is sufficient air to ofEer enough resistance to the passage of meteorites to raise their
temperature to incandescence. Whatever be the height, the weight of the normal
barometric column (per square centimetre of mercury) measures the normal weight
of a column of air of the same sectional area and extending from sea level upwards.
B. Pascal (1663) appears to have been the first to calculate the total weight of all
air about the globe. His estimate is 8,283889,440000,000000 hvres— where a livre
is equivalent to 1 lb. 1 oz. 10^ dr. avoirdupois.
References.
1 S. Brown, Essays, Edinburgh, 1858; G. F. Rodwell, Chem. News, 9. 14, 26, 50, 242, 1864;
10. 74, 1865 ; 11. 74, 1865.
2 T. Gomperz, Griechische Denker, Leipzig, 1. 191, ]896 ; London, 1. 238, 1901.
' E. Tomcelli, Opera geometrica, Firenze, 1644.
* F. Linus, De corporum inseparabilitate, London, 1661 ; Thomas Hobbes, Collected Works,
London, 1845.
^ R. Boyle, An Examen of Mr. T. Hobbes, his Dialogues physicus de natura aeris, London,
1662 ; Animadversions upon Mr. T. Hobbes' Probletnata de vacuo, London, 1674 ; A defense of
the doctrine touching the spring of air proposed by Mr. Buyle in his new physico -mechanical experi-
ments against the objection of Franciscus Linus, London, 1662 ; T. Hobbes, Lessons of the Principles
of Geometry, Appendix to Elementorum philosophiae, London. 1655.
« G. H. Bryan, Phil. Trans., 196. A, 12, 1901 ; G. J. Stoney, Of Atmospheres on Planets and
Satellites, Dublin, 1897; E. Halley, Phil. Trans., 31. 116, 1723; B. Pascal, Traitez de Vequilibre
des liqueurs et de la pesanteur de la masse de Vair, Paris, 1663.
§ 2. The Influence of Pressure on the Volume of Gases — Boyle's Law
At the bottom of all cosmic order lies the order of mathematics, the law that twice
two is always four. — P. Carus.
The quantity of matter in a gas is most frequently determined by the measure-
ment of its volume. The volume of a gas is very sensitive to changes of pressure.
THE PHYSICAL PROPERTIES OF GASES 151
and it is therefore very pertinent to inquire : What is the effect of variations of
pressure on the volume of a gas ? About the time Pascal and Torricelli demonstrated
the weight and pressure of the atmosphere, 0. von Guericke (1650) invented the air
pump. The new instrument attracted much attention, and the effect of the
*' vacuum " (reduced pressure) was tried on all kinds of animate and inanimate
objects. In his memoir, Nova experirmnta physico-mechanica de vi aeris elasticce
(London, 1660), Robert Boyle says that he placed a partially inflated lamb's bladder
in the vacuum produced by the air pump, and noticed that the bladder became
fully distended to its former size. Boyle thus established the important fact that
the less the pressure exerted upon a given mass of air, the greater its volume.
In 1661, Boyle continued his work on the elasticity or spring of air, as he called it,
and stated that R. Townley,^ after reading about Boyle's experiments on the
determination of the density of air from the height of a colimin of mercury which
it supports, propounded the view that " the pressures and expansions are in
reciprocal proportions." On August 2nd, 1661, R. Hooke made some experiments
which confirmed Townley's hypothesis, and W. Croone and R. Boyle, at a meeting of
the Royal Society on September 11th, 1661, gave an account of some experiments
on the same subject. In his Defense against lAnus (London, 1662), Robert Boyle
pubhshed an account of the experiments which clearly estabhshed R. Townley's
hypothesis. Accurate experiments of this nature, said Boyle, " have not been
previously made (that I know) by any man." Boyle's result can be expressed in
words : the volume of a gas kept at one uniform temperature varies inversely
as the pressure. This is Boyle's Law. Some years afterwards, E. Mariotte, in
his Discours de la nature de Vair (Paris, 1679), reported analogous results which he
and M. Hubin obtained in 1676 by means of an apparatus similar to that employed
by Robert Boyle, which led him to take it pour une regie certaine ou hi de la
nature, que Vair se condense a proportion des poids dont il est charge, and thus to
confirm R. Boyle's deduction made fourteen years earlier. On the Continent, ignoring
a priority of at least fourteen years, the law is sometimes improperly designated
la hi de Mariotte, or Mariottesches Gesetz. At the time of the discovery of the law,
air was the only gaseous body known, and therefore the accuracy of the law was
established by Boyle and Mariotte for one body only. The law of Boyle may
therefore be expressed : The product of the pressure and the volume of a gas kept
at one uniform temperature is always the same. Or, for a given mass of air,
pv = constant. The numerical value of the constant, of course, depends upon what
units are selected for representing the pressures and volumes. Pressures may be
expressed in atmospheres, miUimetres of mercury, pounds per square inch, etc. ; and
the volume in litres, cubic centimetres, cubic feet, etc. Boyle's law assumes yet
another guise. If pi be the pressure of a gas occupying a volume Vi ; and p, the
pressure when the volume is v, then, since the products pv and piVi are equal to
the same constant, they are equal to one another. Consequently pv=piVi. If any
three of these magnitudes be known, the fourth can be calculated directly. A large
number of measurements are summarized in these formulae, any one of which,
indeed, contains the essence of all Boyle's observations condensed into a simple
equation.
Example.— A eudiometer holds 4-5 litres of gas when the barometer reads 755 mm.
What will be the volume of the same body of gas when the barometer stands at 760 mm. ?
Here, pi = 755, Vi=4:-5, p = 760, hence, ^=4-47 litres. The most common problem is to
calculate— reduce— the volume of a gas at any observed pressure, to the correspondmg
volume at normal pressure, 760 mm. Given 4-5 litres of gas at 755 mm. pressure, there is
no need for any formula to calculate the corresponding volume at 760 mm. The Pressure
760 mm. is greater than 755 mm., hence the volume will be less, hence miUtipJy 4-5 by the
fraction i^^ and the result is 4*47 litres.
When the volume of gas, collected over mercury, is to be measured when the
pressure of the atmosphere is 760 mm., and the difference in the levels of the
mercury in the gas jar and in the pneumatic trough is 56 cm., it foUows that the
152 INORGANIC AND THEORETICAL CHEMISTRY
pressure of the gas in the narrow tube is 760 mm. less 560 mm. =200 mm. When-
ever practicable, of course, the mercury inside and outside is brought to the same
level before the gas is measured.
Suppose that the confining liquid is water, not mercury. Water is frequently
used when the gases are not appreciably soluble in that liquid. Suppose that the
external pressure is 760 mm. (barometer), and there is a difference of 10 cm. between
the level of the water exposed to the gas, and the level of the water exposed to the
air. The weight of 10 cm. of water is not the same as the weight of 10 cm. of mercury.
Mercury is 13'596 times as heavy as water, hence, a 10 cm. column of water is equi-
valent to the weight of a column of mercury 10-f-13*596 or 0-74 cm. or 7'4 mm.
high. The pressure of the gas is therefore 760 — 7*4 = 752-6 mm. But water vapour
exerts a definite pressure at any given temperature, and a still further reduction
must be made if we want the pressure actually due to the gas and not to the mixture
of vapour and gas. This will be investigated later.
Test for the equilibrium of gases. — If the gas be confined under such conditions
that the product pv at any fixed temperature is not con-
stant, the system will not be in a state of equilibrium.
If a gas were confined in a cyhnder with a sliding piston
moving without friction and if the constant in Boyle's
equation be p (in atm.) v (in litres) =12, then, if the piston
supports a weight of 6 atms., the gas will expand or con-
tract until the product pv satisfies the test. Consequently,
Boyle's law describes the necessary condition for the
volume and pressure of a gas to be in a state of equi-
o Pressures Ubrium whcu thc temperature is invariable. In practice
Fig. 1. — Duhem's Ex- there is no such thing as a frictionless piston, and if
periment. Boyle's law were to be tested in a real cylinder an allowance
would have to be made for the friction of the piston by putting
an extra weight a on the descending piston and a less weight ft on the ascending
piston ; Boyle's law would then be {p4-a.)v or (p-\-ft)v is equal to a constant.
P. Duhem (1902) ^ has used an interesting illustration. The dotted curve, Fig. 1, repre-
sents the relation between pressure and volume as defined by Boyle's law. If the volume ,
corresponding with any given pressure be observed when the rising piston has come to rest,
the observed volume will appear to be less than that corresponding with the pressure as
defined by Boyle's law, because friction will prevent the piston rising to the point corre-
sponding with the equilibrium position on the dotted curve. Similarly, on a descending
piston, friction prevents the volume attaining that indicated on the equihbrium curve.
The friction thus corresponds to what J. W. Gibbs (1876) called the passive re-
sistance of a system to assume a state of equilibrium. The nature of the passive
resistance can here be recognized, but in some cases we feel sure that something
analogous retards the movement of a system to the condition called stable equi-
librium, although we know nothing of the character of the passive resistance or
hysteresis— from uo-rcpew, I lag behind — which opposes the change.
References.
1 G. F. Rodwell, Che7n. News, 9. 14, 26, 50, 242, 1864 ; 10. 74, 1865 ; 11. 74, 1865.
2 P. Duhem, Traite elementaire de mecanique chimique fondee sur la thermodynamique, Paris,
1897 ; Theorie thermodynamique de la viscosite, dufrottement, et desfaux equilihres chimiques, Pari.<?,
1896; Thermodynamique et chimie, Paris, 1902; J. W. Gibbs, Trans. Connecticut Acad., 3. 108,
343, 1876-8.
§ 3. Deviations from Boyle's Law
Experimentally we do not know of any gas behaving in strict conformity to the law of
Boyle ; but in the case of many gases, and of nearly all gases at very high temperatures,
the deviation from uniformity is very slight.- — ^J. B. Stallo.
The pressures used by Boyle extended over a range varying from 3 cm. to 300
cm. of mercury. It is hazardous to infer that because the product pv is constant
THE PHYSICAL PROPERTIES OF GASES
153
over a limited range of pressures, it will remain constant for pressures widely different
from those actually measured. The method of measurement used by R. Boyle,
though excellent for its time, is now considered somewhat crude. In the middle of
the eighteenth century, P. van Musschenbroeck (1729), J. H. Sulzer, and J. Robinson
tried to find if Boyle's result could be extended to all pressures, but with no very
definite results. In 1799, M. van Marum i called attention to the deviation of
ammonia from Boyle's law at high pressures. H. C. Oersted and C. Suensson (1826),
and C. Despretz (1827) extended the observations to other gases, and it was found
that the easily condensable gases like ammonia, hydrogen sulphide, and cyanogen
began to deviate appreciably from Boyle's law at pressures exceeding two atmo-
spheres, and with air, the constancy of the product began to fail at pressures exceed-
ing 20 atm., for it diminished with increasing pressures. Similar conclusions
were estabHshed for other gases by F. J. D. Arago and P. L. Dulong (1831) and by
C. S. M. PouiUet (1844).
Later on, many careful investigations were made by H. V. Regnault (1847),
J. 0. Natterer (1850-4), L. Cailletet (1870-9), E. H. Amagat (1869-93), and others,
to find if the simple law of R. Boyle correctly describes the l3ehaviour of gases at
pressures far removed from the normal pressure of the atmosphere — 76 cm. of
mercury. The general results show that no two gases behave precisely in the same
way. The deviations for many gases are significant. By differentiating the re-
lation pv = constant, Jc, or rather v = k/p, dv/dp = — k/p'^, and if k be taken unity,
and j9 = 2, 3, 4, . . . be substituted.
dp p^'
dp
4' 9' 16'
meaning that the greater the pressure to which a gas is subjected the less the
corresponding decrease in volume, — dv, for any subsequent increase of pressure.
With most gases, the concentration increases more, that is, the volume increases
less than Boyle's law describes ; and at high
pressures, the concentration increases less, that is, p
the volume is greater than Boyle's law indicates.
This is illustrated by plotting Boyle's law. 50
Boyle's law, when graphed, furnishes the con-
tinuous curve shown in Fig. 2. This curve is 40
a rectangular hyperbola. The deviations with
nitrogen from this ideal condition are indicated 30
by the dotted line in the same Fig. 2. If it
were not for this phenomenon, the density of 20
the gas would increase so that while oxygen at
one atm. pressure weighs about 0*0014 grm. per 'o
C.C., at a pressure of 3000 atm. the gas would be
four times as heavy as water, and at 10,000 atm. °^
pressure over 13 times as heavy as water.
According to Boyle's law, the volume of a gas
should diminish indefinitely as the pressure is • j c -i. 1
increased, and in time the volume would approach zero, or become indehnitely
small. This is absurd. Pressure can diminish only the space between the mole-
cules and not the actual substance of the molecules. Hence, if h denotes the
volume occupied by the molecules the changes in the volume of the gas with
variations of pressure will be represented by p{v-h) ^constant, not by ^v=constant.
It does not follow that h represents the actual volume of the space occupied by the
matter in the molecules. The effect of the volume of the molecule on the compressi-
bility of a gas was dimly recognized by D. Bernoulli, 1738 ; and by M U . Lomanos-
soff, 1750 ; it was studied by A. Dupre, in 1865 ; and by J. D van der \\ aals in 1872.
In his important Memoires sur relasticite et la dilatabilUe des flukes jusqu aux
1
1
1
•
-Idea
/Gas
(Graoh
nf
Boyle's Law)\
.__
- Nitrogen
\
\
^
— -
10
20 30 4.0
YiQ,. 2. — Volume : Pressure Curves.
154
INORGANIC AND THEORETICAL CHEMISTRY
trh hautes pressions, embodying the results of work extending from 1878 to 1893,
E. H. Amagat showed that while the product pv remains fairly constant at low
pressures for many gases, the numerical value of pv changes in a remarkable manner
as the pressures increase in magnitude. E. H. Amagat's measurements for carbon
dioxide show that the product pv is not constant, for when
p '
. 1
50
100
125
150
200
500
1000 atms.
pv .
. 1
0-92
0-49
0-31
0-31
0-50
1-02
1-81
Notice how the product pv at first diminishes in magnitude and then steadily
increases. This is brought out very clearly on plotting the numbers. If the
products pv were constant for all values oip,we should get the straight line, dotted
and marked ideal gas line in Fig. 3 ; with carbon dioxide, however, the curve descends
below the line for an ideal gas, and then steadily rises, passing above the ideal gas
line when the pressure is nearly 500 atmospheres.
The curves for hydrogen, helium, argon, and neon, at ordinary temperatures,
do not descend below the ideal gas line, but take a path resembhng the hydrogen
line shown in Fig. 3. However, even
'•®o 1 I i 1 \ ^ i I ^^ these gases exhibit the same peculiar
behaviour at lower temperatures.
Thus, according to H. K. Onnes and
C. Braak (1907), with hydrogen at
—140°, the product pv reaches a mini-
mum when the pressure is about 25 at-
mospheres ; at —195°, 45 atmospheres ;
and at —213°, 51 atmospheres. In
1886, C. Bohr reported that oxygen
behaved in a peculiarly abnormal
manner at a pressure of about 0*7
mm. of mercury. The pressure-
volume curve gaVe an abrupt change
of direction which was ascribed to
the transformation of oxygen into
another variety ; but some careful
measurements by Lord Rayleigh
(1907) and M. Thiesen (1901) indicate
that the statement is probably
summarize these results at a constant
l«40
I'OO
0*60
0*20
y
y.
•V.
^
/^
V
1
Ar^
^
/
^co^
t
'"id
1
A
w
y
Line
\
J
Y
)
\/
i*^
V
Press
tre p
200
400
600
BOO
Fig. 3. — pv-Pressure Curves (Amagat).
based upon a mal-observation. To
temperature :
(1) With small pressures, the product pv decreases with increasing pressure
showing that the volume of gas, at relatively small pressures, is less than
is described by Boyle's law. At very low pressures, the gas will follow
Boyle's law pv = piVi. Lord Rayleigh (1901-2) found no appreciable
variation with oxygen, hydrogen, and nitrogen between O'Ol and 1*5 mm.,
showing that between these pressures the deviations from Boyle's law are
too small to be detected.
(2) With large pressures the product pv increases with increasing pressure,
showing that the volume of the gas, at relatively great pressures, is greater
than is described by Boyle's law.
(3) All gases, in consequence, show a minimum value for the product pv. At
0°, for example, the minimum value of pv for air and nitrogen occurs at
100 atm. pressure ; for oxygen at about 200 atm. ; for carbon dioxide,
at about 35 atm. ; and for ethylene at about 42 atm. The pressure
corresponding with the minimum depends on the nature of the gas and
on the temperature. The minimum is less prominent with the more
permanent gases than with the more condensable gases.
THE PHYSICAL PROPERTIES OF GASES 155
Gases which obey Boyle's and Charles' laws under ordinary atmospheric con-
ditions usually remain gaseous at comparatively low temperatures and are accordingly
called permanent gases.
Befebences.
1 M. van Marum, Gilbert's Ann., 1. 145, 1799 ; P. van Musschenbroeck, Elements phyaicce,
Lugduni Batavorum, 1734 ; H. C. Oersted and C. Suensson, Edin. Journ. Science, 4. 224, 1826 ;
- Omluftens sammentrykkelighed, Forosg over den Mariotteske Lov, Kaobenhavn, Oversigt, 13, 1825;
C. S. M. Pouillet, rMments de physique, Paris, 1. 327, 1844 ; Compt. Rend., 24. 915, 1847 •
J. Robinson, System of Mechanical Philosophy, Edinburgh, 3. 637, 1822 ; C. Despretz, Ann
Chim. Phys., (2), 34. 335, 443, 1827 ; Cmipt. Rend., 14, 239, 1842 ; 21. 216, 1845 ; J. 0. Natterer,
Sitzber. Akad. Wien., 5. 351, 1850 ; 6. 557, 1850 ; 7. 557, 1851 ; 12. 199, 1854 ; Liehig'a
Ann., 54. 254, 1845; Pogg. Ann., 62, 139, 1844; 94. 436, 1855; L. Cailletet, Compt.
Rend., 70. 1131, 1870; 74. 1282, 1872; 75. 77, 1271, 1872; 90. 210, 1880; 88. 61, 1879;
Ann. Chim. Phys., (5), 19. 386, 1879: J. H. Sulzer, Mem. Acad. Berlin, \\&, 1753; F. J.D.Arago
and P. L. Dulong, Mem. Acad., 10. 193, 1831 ; Ann. Chim. Phys., (2), 43. 74, 1830 ; Bull. Soc.
Encour., 29. 295, 1836 ; H. V. Regnault, Mem. Acad., 21. 1, 329, 1847 ; 26. 229, 1862 ; Compt.
Rend., 13. 1077, 1841 ; Ann. Chim. Phys., (3), 4. 5, 1842 ; D. Bernoulli, Journ. Phys., (3), 8, 521,
1899 ; Compt. Rend., 128. 1229, 1899 ; J. D. van der Waals, Die Continuitdt des gasformigen und
fliissigen Zustaiides, Leiden, 1873 ; A. Dupre, Theorie mecanique de la chaleur, Paris, 1869 ; Compt.
Rend., 56. 960, 1863 ; 57. 774, 1863 ; E. H. Amagat, Compt. Rend., 68. 1170, 1869 ; 71. 67, 1870 ;
73. 183, 1871 ; 74. 1299, 1872 ; 75. 479, 1872 ; 77. 1325, 1873 ; 82. 914, 1876 ; 85. 27, 139, 1877 ;
87. 342, 1878 ; 88. 336, 1879 ; 89. 437, 1879 ; 90. 995, 1880 ; 91. 428, 1880 ; 93. 306, 1881 ; 94.
847, 1882 ; 95. 281, 638, 1882 ; Ann. Chim. Phys., (4), 28. 274, 1873 ; (4), 29. 296, 1873 ; (5), &
270, 1876 ; (6), 11. 520, 1887 ; (5), 19. 345, 1880 : (5), 22. 353, 1881 ; (5), 28. 456, 464, 480, 500,
1883 ; Archiv, Sciences Geneve, (4), 35. 169, 1869 ; (4), 40. 320, 1871 ; (4), 49. 246, 1873 ; (5), 8.
270, 1876 ; Lord Rayleigh, Phil. Trans., 196. 205, 1901 ; 198. 417, 1902 ; Zeit. phys. Chem., 37.
713, 1901; 41. 71, 1902; 52. 705, 1905; Proc. Roy. Soc, 73. 153, 1904; M. Thiesen, Ann.
Physik, (4), 6. 280, 1901 ; C. Bohr, Wied. Ann., 27. 459, 1886 ; H. K. Onnes and C. Braak,
Comm. Lab. Phys. Leiden 97, 99, 100, 1907.
§ 4. Dalton's Law of Partial Pressures
Accurate and systematic investigation has brought to light the infinite complexity of
nature ; the fineness of the dovetailing of every event into many others ; the never-ending
response of all things to changes in the conditions that encompass them ; the imiversal
orderliness of natural occurrences ; and the absolute interdependence of cause and effect.
— M. M. P. MuiR (1894).
When two gases, which do not act chemically on one another under the con-
ditions of the experiment, are brought together, the gases mix intimately, by diffusion,
so as to form a homogeneous mixture. Furthermore,^ John Dalton (1802) found
that each gas seemed to exert the same pressure as if it occupied the space alone, and
the total pressure of the mixture of gases was the sum of the several pressures due
to each gaseous component of the mixture. If P be employed to denote the total
pressure of a mixture of gases, and^i the partial pressure exerted by one of the gases,
P2 the partial pressure exerted by another gas, pg the partial pressure of a third gas,
Dalton's discovery means that T P =i?i +i?2 + Pa + • • • ^^ words, in a mixture
of gases which exert no physical or chemical action on one another, each gas exerts
the same pressure as if it alone occupied the entire vessel, and the total pressure
is the sum of the partial pressures due to each of the gases. This is Dalton's law
of partial pressures. If the four volumes of nitrogen and one volume of oxj^gen m
the atmosphere be under normal pressure, the nitrogen gas will sustain a pressure
approximately 608 mm. and the oxygen gas 152 mm. of mercury. J. Dalton added :
Since two gaseous fluids which exert neither attraction nor repulsion on one another,
distribute themselves so that their imited pressure is equal to the pressure of the atmo-
sphere, all the components of the atmosphere are arranged together at a given pressure and
temperature, and by a paradoxical though true disposition, each of them occupies aU the
space destined for the aggregate.
156 INORGANIC AND THEORETICAL CHEMISTRY
It might be added that Dalton's partial pressure law is quite independent of Boyle's
law, and can be extended to mixtures of any number of gases.
Examples.— (1) Moist hydrogen gas is confined over water under a pressure of 747-2 mm.
of mercury at 15 "3°, the partial pressure of water vapour at that temperature is J 2*9 mm. of
mercury. Then, from Dalton's law of partial pressures, it follows that the hydrogen gas
itself is under a partial pressure equivalent to 747*2 less 12*9, or 734-3 mm. of mercury.
(2) If atmospheric air contains a mixture of four volumes of nitrogen and one volume
of oxygen, show that if the manometer records a pressure p^ the partial pressure of the
oxygen gas will be ^jo, and of the nitrogen gas ^p.
(3) If a moist gas of volume Vi be confined in a vessel at a pressure p^, show that the
volume V of dry gas at normal pressure, 760 mm., and the volume v^ of the water vapour
at normal pressure, are respectively v=v^{p■^^—f)|lQ(i, and V2=ViflJQ0, where / denotes
the vapour pressure of the water at the temperature of observation.
There are many reasons for supposing that the molecules of a substance exert
some kind of attraction on one another. This intermolecular attraction gives rise
to phenomena of cohesion, viscosity, capillarity, surface tension, etc. The inter-
molecular attraction is probably very powerful in solids, weaker in liquids, and very
small with gases ; but it is highly probable that the molecules of nearly all gases
do exert some attractive influence on one another, and the gases, in consequence
of this physical action, " deviate " from Dalton's law to an extent dependent upon
the magnitude of the intermolecular attraction. Many mixtures of gases show
slight, but not marked deviations from the law, e,g. carbon dioxide and sulphur
dioxide ; hydrogen with air, and with nitrogen ; etc.
P. Fuchs 2 has investigated the change in volume which occurs on mixing chemi-
cally indifferent gases — ^nitrogen with nitrous oxide, carbon dioxide, or oxygen ;
nitrous oxide with carbon dioxide or oxygen ; and oxygen with carbon dioxide.
In every case there is an expansion which is greater the more the two components
differ in physical properties. The change in volume does not correspond with the
ratio of the two gases, but reaches a maximum which is beyond the 1 : 1 ratio,
so that the maximum change occurs with mixtures containing more than 50 per
cent, of the gas with the lower critical temperature ; and the maximum lies nearer
to the 1 : 1 ratio, the more the components resemble one another. The change in
volume 8v is qualitatively but not strictly quantitatively represented by J. D. van
der Waals' equation
8v
x(\-x) {^^^^^^ - (h + h- h,))
where x denotes the number of gram-molecules of the one gas, and (1 ~x), of the
other ; ai represents the attraction constants of the molecules of the one com-
ponent ; a2, of the other component ; and ai2, of the molecules of the different
components for one another ; bi, 62? ^^^ ^12 represent the corresponding volume
constants. Accordingly, the theoretical results agree more closely with the
observed results when an allowance is made for the effect of the attraction of the
molecules for one another.
J. Dalton's law is thus a limiting law for ideal gases. A. Leduc ^ prefers to state
the law for actual gases in the form : The volume occupied by a mixture of gases is
equal to the sum of the volumes which the component gases would separately occupy
at the same temperature and pressure as the mixture. If two gases, originally at the
same pressure, are mixed so that the temperature and total pressure remain unaltered,
the pressure of the mixture can be calculated if the coefficients of deviation from
Boyle's law. Ay be known between the common pressures p and pi for the mixture
and for each of the two gases, where
V 739^9 A V9. — V-i '
^2'^2 J^V^-V\
There is usually found to be a slight increase of pressure on admixture which is
scarcely measurable with the less condensable gases. The value of A at 16° between
THE PHYSICAL PROPERTIES OF GASES 157
1 and 2 atm. is 0*000143 for a mixture of nearly equal volumes of carbon and sulphur
dioxides ; 0*000005 for air ; —0*000002 for equal volumes of hydrogen and oxygen.
The law had been applied to test if chemical action occurs on mixing certain
gases, e.g. to find if any sign of chemical action occurs when nitric oxide (NO) is
mixed with nitrogen peroxide (NOg) resulting in the formation of nitrogen trioxide
(N2O3). It is assumed that if no chemical combination takes place, the mixture
will obey Dalton's law, and conversely.^ The conclusion can be valid when it has
been shown that the molecules of the two gases exert neither attraction nor
repulsion upon one another. If they did, the test might lead to wrong conclusions with
respect to chemical action. A slight contraction, for instance, might be evidence
of molecular attraction, not of chemical combination.
References.
1 J. Dalton, Mem. Manchester Lit. Phil. 80c., 5. 635, 1802 ; Ann. Chim. Phys., (1), 44 40
1802. ■ '
2 P. Fuchs, Zeit. phys. Chem., 92. 641, 1918 ; J. D. van der Waals, Binare Gemische, Leipzig,
1900.
3 A. Leduc, Compt. Rend., 123. 805, 1896 ; 126. 218, 1859, 1898 ; A. Leduc and P. Sacerdote,
ib., 126. 218, 1853, 1898 ; A. Leduc, Recherches sur les gaz, Paris, 105. 820, 1898 ; D. Berthelot
and P. Sacerdote, Compt. Rend., 128. 820, 1899 ; D. Berthelot, ib., 126. 954, 1030, 1415, 1703,
1877, 1898; 128. 1159, 1899.
* H. B. Dixon and J. D. Peterkin, J own. Chem. Soc., 75. 613, 1899.
§ 5. The Laws of Nature
We must confess that physical laws have greatly fallen off in dignity. No long time ago
they were commonly described as the Fixed Laws of Nature, and were supposed sufficient
in themselves to govern the universe. Now we can only assign to them the humble rank
of mere descriptions, often erroneous, of similarities which we believe we have discovered.
—J. H. POYNTING (1899).
Nature, always working by law, is always consistent, always inexorable ; her laws are
invariable.- — ^A. Simmons.
This is a convenient place to further emphasize the meaning of the term " law "
in chemistry. The laws of a country may be the enactments of a ruling power,
the ukases of a czar, or the regulations of the police superposed upon a people
compelling them to act in particular ways, but it is of course absurd to say that
Dalton's law and Boyle's law must be obeyed, implying that these laws are com-
mands imposed upon gases which they are compelled to obey. The laws of nature
describe, they do not compel. A substance does not act in a particular way because
there is a law, but the law originated when it was found that substances acted
in that particular way. Consequently, law is a useful term which the careless
sometimes personify ; it is a figure of speech, and is employed by scientific men
purely in a metaphorical sense. The term has led to some confusion, for it has
led to the belief that the uniformity described by the law has been imposed on
nature by the will of a rational being — God himself. As previously indicated, a
law in science is a kind of summary of the present state of our knowledge of the
phenomena described by the law, and it is always subject to revision with the
growth of knowledge. Laws do not necessarily establish facts. Consequently,
the term would be replaced by another word, if we could think of a better.
Rule would perhaps lead to less misunderstanding. The German equivalent--
Gesetz, statute — is perhaps worse. A law of nature can have authority only in so
far as it is based on facts. As indicated previously, the term " law of nature " is
applied to a comprehensive generalization which " methodically and systematically
describes certain natural phenomena." The laws of chemical combination describe
what the elements do under definite conditions ; and generally, the laws of
chemical and physical phenomena are collocations of those circumstances
158 INORGANIC AND THEORETICAL CHEMISTRY
which have been found by experiment and observation to accompany all
chemical and physical changes included in the statement of the law. The test
of the " law " is that the statement holds good without exception. A broken
law, said J. H. Poynting, is a false description.
It is sometimes said that a law of nature has never been disproved ; this can
only mean that if a law of nature is disproved, it ceases to be a law. The common
meaning attached to the saying, " The exception proves the rule," is wrong, and
it is an instance of confusion arising from the double meaning of words. In the old
Latin form, Exceptio prohat regulam, the word prohat means tests, just as to-day
proving wines means testing them. The proverb therefore meant that the apparent
exception furnishes a means of trying, testing, or proving the rule, and if the ex-
ception cannot be explained, then the rule breaks down, for the exception disproves
the rule. The exception annihilates the rule, for, said J. W. Ritter in 1798, a law
must be abandoned immediately a real exception is discovered — it is no longer a law.
When the exact conditions are set up, the law describes the phenomenon without
variableness or shadow of turning. The law is then regarded as an objective power.
This power is called a force, and further, the force is said to be the cause of the
phenomenon. Thus gravitation is regarded as an attractive force causing one
particle to attract every other particle in the universe ; chemical affinity is regarded,
in this sense, as a selective chemical change. If therefore we find a gas deviating
from Boyle's law, or a mixture of gases " disobeying " Dalton's law, the alleged
laws may be false, incomplete, or imperfect descriptions, or some perturbing influence
is at work which masks the simple phenomena described by these laws.
§ 6. The Influence of Temperature on the Volume of Gases — Charles'
Law
According to the schools of philosophy, it has been proved that the effect of cold is to
make bodies contract while heat makes them expand.- — G. Galilei (1615).
The expansion of air by heat has long attracted the attention of chemists. Hero
of Alexandria (c. 117 B.C.), G. B. Porta (1616), C. Drebbel (1608), and G. Galilei
(1615) experimented on the subject. H. Boerhaave considered the effect of tempe-
rature on the volume of gases, and, in his Elementa chemice (Lugduni Batavorum,
1732), he stated that when air is heated, it becomes so rare that -neither the measure
nor the limit of its dilation has been yet discovered ; and added :
Air of unequal masses but of the same density, is always expanded in the same measure
by the same degree of fire ; so that these expansions in the same density of air, by a constant
law of nature, are always proportional to the augmentations of heat.
Influence of temperature on the volume of gases — pressure constant. — In 1790,
Joseph Priestley concluded " from a very coarse experiment " that " fixed and
common air expanded alike with the same degree of heat ; " J. Dalton, in 1801,
inferred from his experiments : " Upon the whole, I see no sufficient reason why
we may not conclude that all elastic fluids, under the same pressure, expand
equally by heat ; " and J. L. Gay Lussac,i in 1802, quoted some experiments in
support of the generalization : The same rise of temperature produces in all gases
the same increase in volume, provided the pressure and mass be kept constant.
This law is generally designated Charles' law, in honour of J. A. C. Charles, who,
according to Gay Lussac, made some crude experiments on the subject fifteen
years before Gay Lussac's publication. Some call this relation Ga?j Lussac' s law.
It might, perhaps, with more propriety be called Volta's law, because A. Volta,2
described it in his Memoria sulla uniforme dilatazione deiVaria, in 1793.
G. Amontons had an inkling of this law in 1702.
The increase in volume which occurs when one litre of nitrogen at 0° is heated
in a suitable vessel is shown in the following table (R. Chappius, 1888) :
THE PHYSICAL PROPERTIES OF GASES
Table I. — Thermal Expansion of Nitrogen .
159
Temperature 0«*.
Volume V litres.
Expansion per litre per
degree.
0
10
20
30
40
10000000
10367781
10735396
1-1102875
1-1470244
0-0036778
00036770
0-0036763
00036761
The numbers in the last column — called the coefficients of thermal expansion —
mean that the volume ^; of a litre of nitrogen, when heated through 6° can be repre-
sented very closely by the expression : v = {l +0*003676^) htres. In other words,
nitrogen increases 0*003676, or very nearly 273rd part of its volume at 0° for every
degree rise of temperature. More generally, if Vq be used to denote the volume of gas
at 0°, we have, instead of the preceding expression, v = '^0(1 + 273^)' otv = VQ{l-\-aB).
This is very nearly true for most of the common gases, and it therefore represents
a condition which must be satisfied by the temperature and volume of a gas,
under constant pressure, in order that the system may be in stable equihbrium.
While solids and liquids have their own characteristic coefficient of expansion,
gases have nearly the same coefficient of thermal expansion. This is the meaning
of Charles' law. The coefficients of thermal expansion (pressure constant) for
the gases run something like this for one atmosphere pressure and variations of
temperature between 0° and 100° :
Air .
. 0-003671
Hydrogen .
. 0-003661
Carbon dioxide
. 0-003728
Carbon monoxide .
. 0-003669
Sulphur dioxide .
. 0-003903
Nitrous oxide
. 0003719
These numbers are close enough to -^^ for most practical purposes. In general,
the more easily a gas is liquefied and the greater its molecular weight, the greater
the deviation from the constant 0*003665 found for air— witness carbon dioxide,
0-003728; hydrogen bromide, 0-00386 ; etc.
For every degree centigrade the temperature falls, the volume of the gas
decreases by 273rd. If -^^^id part of a gas be taken away 273 times, no more gas
remains. This is illustrated by plotting the above equation.
If the temperature be less than —273°, the gas would have a negative volume, that is,
a volume less than nothing ! If the temperature be —273°, the gas would occupy no voluine !
rt is impossible to imagine a substance occupying no space, but such is a logical conclusion
from Charles' law. Where is the fallacy ? Whenever a natural process is represented by
mathematical symbols, it is well to remember that the artificial statement often expresses
more than actually obtains in nature, because, in the physical world, only changes of a certain
kind occur. We must therefore limit the generality of the mathematical expression.
Charles' law includes a simplifying assumption. The apparent volume of a gas may be
resolved into at least two parts ; (1) the volume occupied by the molecules of the gas ; and (2)
the space betvjcen the molecules. If b denotes the space occupied by the molecioles, and v
the observed volume of the gas, the space between the molecules will be represented by
V —b. Although for the sake of simplicity, we assiune v to represent the total volume occupied
by the gas, Charles' law refers to v—b, that is, to the space between the molecules, and
in that case, the conclusion that v=0 when the temperature is -273° involves no absurdity.
Moreover, the gas would liquefy before the temperature -273° was attained, and the simple
gas law of Charles would not then be applicable.
It has been urged that J. L. Gay Lussac's statement of Charles' law means that
the increase in the volume of a gas at any temperature, for a rise of 1°, is a constant
fraction of its initial volume at 0°— in symbols, v=ro{l-\-ad) ; while J. Dalton's
statement of the law means that the increase in the volume of a gas at any tempera-
160 INOKGANIC AND THEORETICAL CHEMISTRY
ture, for a rise of 1°, is a constant fraction of its volume at that temperature— in
symbols dvjdd^av ; hence by integration, v=VQe^^ . If the latter expression be
expanded, v=Vo(l+ct0+Ja2^2_|_ ^ ^ j^ and if the second and higher powers
be outside the range of measurement, the two statements amount to the same
thing. R. Mewes and L. Neumann 3 proposed to replace v=VQ{\-\-a6) by
v=x-\-{vQ—x)(\-\-a^), or approximately v=VQ(\-\-a)^. The results at ordinary
temperatures are good, but they become less accurate with decreasing tempera-
tures. The discrepancies are in fact attributed to errors in the measurements at
low temperatures which are introduced by surface condensations, etc.
Influence of temperature on the pressure of a gas — volume constant.— About
1682, R. Boyle made some experiments on the influence of " cold and heat " on
the pressure, or the spring of air, as he called it, and found that the effect of the
greatest degree of cold he could produce did not " weaken the spring by anything
near so considerable as one would expect." The subject did not attract much
attention until G. Amontons (1702-3) ^ published two memoirs in which he
demonstrated that equal masses of air, measured at the same initial pressure,
acquire equal increments of pressure when heated to the boihng-point of water
provided the volumes are maintained at their initial value ; and if the pressure of
the air before heating be doubled or tripled, the additional pressure produced when
the air is heated to the boiling-point of water is likewise doubled or tripled. Other-
wise expressed, the ratio of the total pressures {jp and f{) of air at two definite
temperatures {T and Tj), and kept at a constant volume, has always the same
value R and is independent of the initial pressure. In symbols, fjT =pilTi ; which
can be written p = RT, where R is the constant of proportion. In words, the same
rise of temperature produces the same increase of pressure provided the volume
and mass of the gas be maintained constant. This relation might be called
Amontons' law. It can be very simply deduced from Charles' and Boyle's laws,
expressed in an analogous form, p=Po{l -\-27 3^)> or f = Po{l -{- Pd) , where
/3 denotes the coeflB.cient of increase of pressure (volume constant). J. L.
Gay Lussac thought that all gases had the same values of a and j3 ; and it was
thought that a = j3. More exact measurements have shown that neither statement
is true. The coefficient j3 for the above-named gases between 0° and 100° are :
Air
. 0-003665
Hydrogen ....
. 0-003663
Carbon dioxide
. 0-003688
Carbon monoxide
. 0-003845
Nitrous oxide
. 0-003676
Absolute zero. — J. Amontons (1703) ^ argued that air would exert no pressure
at all if it were cooled below freezing-point of water to about 2 J times the range of
temperature between the freezing- and boihng-points of water. In 1779, J. H. Lam-
bert ^ repeated Amontons' experiment and estimated that air would occupy no volume
at all, if cooled to — 270° ; more accurate measurements make this temperature — 273°.
This temperature, —273°, is supposed to be a non ultra plus, or limiting temperature
— the nadir or lowest possible temperature — a kind of primum frigidum. Hence,
—273° is sometimes called the absolute zero ; and temperatures reckoned from this
zero are called absolute temperatures. U Association Internationale de Froid "^
recommended that the letter K — from Lord Kelvin — be employed to denote absolute
temperatures so that 0° C. =32° F. =273° K. On the absolute scale of temperatures,
0° C. will be 273° K. If J be employed to denote the temperature on the absolute
scale, and 6 the temperature on the centigrade scale, we have T=273+^. Hence,
if V be the volume of a gas when the absolute temperature is T, and Vi the volume
when the temperature is Jj, from the preceding equation (3) v : Vi=T : Tj, which
is but another way of stating Charles' law. The volume of a gas varies directly
as the temperature, so that v=RTy where R is the constant of proportion. The
THE PHYSICAL PROPERTIES OF GASES 161
arbitrary nature of the absolute zero deduced from the coefficient of thermal ex-
pansion of air, will appear when it is remembered that a similar train of reasoning
would furnish —5000° as the absolute zero, if the coefficient of expansion of mercury
were made the standard. It must be remembered, however, that the coefficient
of thermal expansion of all gases, unlike liquids and solids, has nearly the same
value ; and further, the gaseous state probably represents the simplest form in
which matter can exist. There are, however, other reasons for selecting —273° as
the absolute zero which are discussed in works on thermodynamics.
The combined influence of temperature and pressure on the volume of a gas. —
According to Boyle's law, the volume of a gas varies inversely as the pressure, so
that if a pressure pi and volume t'l change to a volume a; at a pressure p2, then,
from the relation ;Pi^i=j32^ (Boyle's law). Again, according to Charles' law, the
volume of a gas varies directly as the absolute temperature, so that if a gas whose
volume is a? at a temperature Ti changes to a volume V2 when the temperature rises
to T2, we have from the above relation, xT2=V2^i- On substituting the value of
X from the preceding relation
If 2>2j ^2j ^2 represent the volume of the gas under standard conditions of tem-
perature and pressure, f^^'^jT^ will have a constant numerical value, say R ; and it
follows at once that when both temperatures and
pressure vary, the effect on the volume will be
given by the equation pv=RT, where R is the
constant of proportion — generally called the gas
constant. An equation which attempts to express
the relation between the pressure, temperature,
and volume of a gas is sometimes called the equa-
tion of state — Zustandsgleichung, or equation
caracteristique — or the characteristic equation or
the gas equation. The equation of state is applic-
able to ideal gases. If an arbitrary value be
assigned to the constant R, and corresponding
values of p and v be plotted for a series of values
of T, say T=l,2,3, . . ., a series of curves. Fig. vo/umes
4, are obtained. These curves may be supposed ^^^^ 4._Surface showing the Rela-
to have been drawn on a surface abed. While a tion between the three variables :
plane suffices for showing the relation between two Temperature, Pressure, and
variables, a surface in three dimensions is needed Volume of Gases.
for three variables. These formulae are used a
great deal in calculations involving the variations in the volumes of gases
owing to variations in temperature and pressure. For mstance, m reducmg tne
volume of a gas at any observed temperature and pressure to the con-espondmg
volume at the standard or normal pressure and temperature-O C and 7bu mm.
pressure — often represented by n.p.t., or N.P.T., or S.T.P., or S..L., b.r.
ExAMPLE.^(l) If a gas measures 170 c.c. at a pressure of 735 mm. "mercury, and a
temperature of 15°, what is the volume of the gas at normal temperature ajd Pye^"je .
Here it is required to find v in the preceding formula where p = 7bO ; l -£ia, ^i ^ »
Vi = 170; and ^1 = 735; hence, iy = ||i X^^^f X 170 = 155-8 c.c. ^^lo-dor at
(2) Show that 13-8 c.c. of a gas at 747^6 mm. pressure at 19° reduce to 12 4 c.c. at
760 mm. and 0°.
Approximations can be used for general calculations,8 and books on gas analysis
have tables for converting unit volume at 6° and pressure p X '/''^'''^^/r^^
standard conditions. It will be observed that the fraction T/i o lor tf ^^^^o^^
(273 + d)l{213 + <9o), and if ^o be 0°, the fraction reduces to 1 f oyyr^, or 1 +^Y:f^^J^^'
The numerical value of the gas constant R.-The numerical value of i2 depends
M
VOL. I.
162 INORGANIC AND THEORETICAL CHEMISTRY
upon the units of pressure and volume ; if unit mass of gas be taken, the value of
R will depend upon the molecular weight of the gas. If one gram-molecule be taken,
j)V=RT, and if n gram-molecules be taken, j)v=nRT. If the litre he taken as the
unit of volume and the atmosphere as unit of pressure, and since a gram-molecule of
gas occupies nearly 22*4 litres at 0° and under one atmosphere pressure, 1 X 22*4
=1 X 22x273 ; or i2=0082 litre-atmosphere. If the gram and cubic centimetre be
taken as unit, it follows that if % represents the volume of a gram-molecule of any
gas at n.p.t., i;i=22,400 c.c. ; pi is 1033-3 grms. per sq. cm. ; and Ti=273. Hence,
22=pv/T = (1033-3 x22,400)/273 = 84,760 gram-centimetres of energy. From
measurements of the mechanical equivalent of thermal energy, it is known that
one gram-centimetre of mechanical energy is equivalent to 42,650 calories. Hence,
R=:pvlT = l'd cals., or 2 cals. nearly.
References.
^ J. Dalton, Mem. Manchester Lit. Phil. Soc., 5. ii, 695, 1802 ; J. L. Gay Lusaac, Ann. Chim.
Phys., (1), 43. 137, 1802 ; J. Priestley, Experiments and Observations on Different Kinds of Air,
Birmingham, 1777-
* A. Volta, Annali di Chimica, 4. 227, 1793 ; reprinted in J. Guareschi's Legge della dilatazione
dei gaz di Alessandro Volta, Turin, 1914 ; G. Amontons, Mem. Acad., 50, 1703.
» R.Mewes and L. Neumann, Zeit. Sauerstoff Stickstoff Ind., 11. 13, 1919; T. Box, Practical
Treatise on Heat, London, 1876.
* E. Mach, Die Principien der Wdrmelehre, Leipzig, 1900 ; G. Amontons, Mtm. Acad., 50,
1703.
* J. Amontons, Mem. Acad., 50, 1703.
•^ J. H. Lambert, Pyrametrie, Berlin, 29, 40, 74, 1799.
' Chem. Ztg., 35. 3, 1911.
8 G. J. Stoney, Proc. Roy. Soc. Dublin, (2), 6. 387, 1890; Wa. Ostwald, Zeit. angew. Chem.,
32. 359, 1919.
§ 7. Deviations from Charles' Law
Nature abhors the straight line.- — R. Ross (1914).
We have already seen that the coefficients of thermal expansion of all gases
are only approximately the same. The coefficients for the individual gases differ
a little among themselves as indicated above. The variation in the coefficient of
thermal expansion at temperatures and pressures not far removed from normal
atmospheric temperatures and pressures, is not very marked, and for regular gas
calculations can be ignored. It remains to indicate the variation, if any, in the
coefficient of thermal expansion with large variations of temperature and pressure.
H. Flaugergues (1825) showed that the coefficient of expansion of moist air is rather
larger than that of dry air. Charles' law was also tested by P. L. Dulong and A. T.
Petit (1815), F. Rudberg (1837), H. V. Regnault (1841), G. Magnus (1842), E. H.
Amagat (1873), P. Jolly (1874), P. Chappius (1888), H. K. Onnes and M. Boudin
(1900), etc.i The more important results are as follows :
(1) The exact coeflacient depends on the nature of the gas.— H. V. Regnault,
about 1850, proved that different gases have not the same coefficients of thermal
expansion, as Charles' law assumes, but that each gas has its own specific constant.
For ordinary calculations, particularly with gases which cannot be liquefied in the
neighbourhood of atmospheric temperatures, the coefficient is taken to be
a=^ = ^.
(2) The influence of pressure. — The coefficient of expansion of most gases is
increased by augmenting the pressure of a gas until a maximum value is attained,
after that, the coefficient diminishes with increased pressure. For instance, E. H.
Amagat (1893) found that the coefficients of expansion of carbon dioxide at tem-
peratures between 50° and 60° assumed the following values when the pressure
changed from 30 to 1000 metres of mercury :
Pressures
30
60
126
200
500
1000 metres
CoeflQcients
. 00069
0-0085
0-0410
0-0086
00033
0-0018
THE PHYSICAL PROPERTIES OF GASES
163
r-oo
Carbon dioxide thus shows a marked variation in the coefficient of thermal expansion
at high pressure. In agreement with these facts, the coefficient also diminishes as
the pressure is reduced, even as low as 0"077 mm. of mercury. The variation is
not so marked with gases like nitrogen, oxygen, and hydrogen which are not easily
condensed to the liquid condition. The general result of H. V. Regnault's and
Amagat's work is to show that if a gas is more compressible than is represented by
Boyle's law, the coefficient of thermal expansion is increased by pressure ; and
conversely for gas less compressible than is indicated by Boyle's law, the coefficient
of thermal expansion decreases with an increase of pressure. The value p which
furnishes the greatest coefficient of thermal expansion is that same value of j)
which gives the minimum product pv. At ordinary temperatures, therefore, hydro-
gen and helium do not exhibit this variation in the value of their coefficients of
expansion. With these gases, the coefficient of expansion steadily diminishes with
increasing pressure ; although even these resemble other gases if the temperature
be low enough. Consequently, at high enough pressures^ when the minimum pv is
reached, the coefficient of thermal expansion of all gases decreases with an increase of
pressure.
(3) The influence of temperature. — The general effect of raising the temperature
is to lower the coefficient of expansion. For instance, Hirn (1862) found that for
water vapour from 0° to
118-5° 162° 200° 246-5°
Coefficient of expansion . 0-004187 0-004071 0-003938 0-003799
Similarly, L. Troost and P. Hautefeuille (1876) found the coefficient for silicon
tetrachloride fell from 0-00449 between 100° and 125° to 0-00399 between 125° and
180° ; while between the same temperatures
the coefficient for carbon tetrachloride fell
from 000470 to 0-00414 ; and for phosphorus
trichloride, from 0-00489 to 0-00417.
The changes in the coefficient of expansion
with increasing pressure become less and less
as the temperature is raised, and finally dis-
appear. So does the minimum value of the
product pv become less and less marked as the
temperature is raised. The gradual flattening
of the carbon dioxide curves as the temperature
rises from 40° to 100° is brought out very clearly
in Fig. 5. All gases exhibit a minimum value fiq. 5.— Amagat's pv-T — Curves for
for pv. The pressure required for a mini- Carbon Dioxide,
mum depends on the temperature as well
as on the nature of the gas. The minimum is most marked when the gas is near
its temperature of liquefaction. If the temperature is much above this critical
point, the minimum is very small— with hydrogen the minimum is inappreciable
at 0°— Fig. 3. All other gases show a minimum at ordinary temperatures. Hence,
H. V. Regnault, who discovered this peculiarity of hydrogen, was led to say ironically
that hydrogen is a gas plus que parfait—a. gas more than perfect ; but hydrogen
also shows a minimum at reduced temperatures. Similar remarks apply to helium
and neon.
References.
1 H. Flaugergues, Gehler's Physik. Worterbuch, 1. 637, 1825 ; P. L. Dulong and A T Petit.
Ann. Chim. Phys., 7. 117, 1815; F. Rudberg, Fogg- ^^^^ 41- 271, 1837; ^. 119, 1W8; U.
Magnus, ib., 55. 1, 1842; H. V. Regnault, Mem. Acad., 21. 25, 1841 ; A7in ^?*'"-./^y^-» j^), 5.
52, 1842 ; E. H. Amagat, ib., (4), 28. 274, 1873 ; L. Troost and P. Hautefeuille. tb (6). 7. 464.
1876 ; P. Jolly, Pogg. Ann. Jubelbd., 82, 1874 ; P. Chappius, Arch Sciences Phys. Oefu, (3). ^J.
5. 153, 248, 1888; H. K. Onnes and M. Boudin, Versl. Akad. ^'"^'^'I^'J' 224^^/ f- t*'
Amagat, Ann. Chim. Phys., (6), 29. 68, 1893; G. A. Hirn, Cosmos, 22. 283, 413, 734, 1803;
Theorie micanique de la chaleur^ Paris, 607, 1862.
0*20
300
164 INORGANIC AND THEORETICAL CHEMISTRY
§ 8. The Critical State of Gases
The ordinary gaseous and liquid states are only widely separated forms of the same con-
dition of matter, and may be made to pass into one another by a series of gradations so
gentle, that the passage shall nowhere present any interruption or break of continuity.
Gas and lic(uid are only distinct stages of a long series of continuous physical changes.- —
T. ANDREW.S (1869).
The fact that some elements occur as gases, others as liquids, and yet others as
solids is a mere accident of temperature or pressure. Similar remarks apply to
chemical compounds which do not decompose when the temperature is augmented.
If the prevailing atmospheric temperature were 100° higher than it is, water would
be a gas ; and if 100° lower, water would be a solid. Similarly, if the atmospheric
pressure were ten times as great as it is, chemistry books would describe sulphur
dioxide and many other so-called gases either as liquids or solids ; while if the
pressure were much less than it is, many so-called liquids would be styled gases.
Every substance is potentially solid, liquid, and gas. The solid, liquid, and gaseous
states of matter are merely phases assumed by virtually all kinds of matter as the
temperature rises from absolute zero upwards. The three forms which the elements
and their compounds can assume are called the three states o£ aggregation. The
three states of water are : Gas above 100° ; liquid between 100° and 0° ; and solid
below 0° under ordinary atmospheric pressure. These facts are symbolized :
0° 100°
Waterice — Waterjiquid ^Watergteam (760 mm.)
Investigators who have special facilities for working at high temperatures, report
that gold has a melting point, 1062°, and a boiling point, 2530°, or :
1062° 2530°
Goldsoiid ^ Goldiiquid ^ Goldgas (760 mm.)
Similarly, those working in laboratories specially equipped for measurements at
low temperatures, report that oxygen has a melting point, —227°, and a boiling point,
-182-5°, or
— 227° - 182-5°
Oxygensoiid — Oxygenuquid ^ OxygeUgas (760 mm.)
M. Faraday (1819) ^ emphasized the fact that when any form of matter passes
from the solid to the liquid state, or from the liquid to the gaseous state, its physical
properties diminish in number and variety. Thus, solids in becoming liquids lose
their hardness, crystalline form, etc. ; and in passing to the gaseous state, the
phenomena are still more marked, thus, all gases have nearly the same coefficient
of thermal expansion. " The varieties of density, hardness, opacity, colour, elas-
ticity, and form which render the number of solids and fluids almost infinite, are
in gases supplied by a few slight variations in weight, and some unimportant shades
of colour."
The critical state. — T. Andrews demonstrated in his paper On the continuity of
the gaseous and liquid states of 7natter,^ in 1869, that if gaseous carbon dioxide be
gradually compressed in a vessel suitable for the observation, the volume diminishes
more rapidly than would occur if Boyle's law correctly described the behaviour of
the gas ; and when the pressure attains a certain value, the gas begins to liquefy.
A further decrease in the volume does not change the pressure, but only increases
the quantity of gas liquefied. At length, when all the gas has liquefied, a large
increase of pressure only causes a minute decrease in the volume of the liquid, since
liquids in general undergo but a small change of volume on compression.
If the experiment be made with carbon dioxide at 0°, the gas commences to
liquefy when the pressure has attained 35'4 atmospheres ; if at 13'1°, liquefaction
commences at 49"8 atmospheres pressure ; if at 30°, 70 atmospheres pressure ;
while if the temperature exceeds 31°, or, more accurately, 31' 35°, no pressure,
however great, will liquefy the gas. Other gases exhibit analogous phenomena. This
THE PHYSICAL PROPERTIES OF GASES
165
is in agreement with M. Berthelot's conclusion ^ in 1850 that pressure will not
liquefy gases under all conditions of temperature. For each gas there is a particular
temperature above which Uquefaction is impossible however great be the applied
pressure. Andrews called this the critical temperature of the gas ; the corresponding
pressure, or the critical pressure, is the least pressure which will liquefy the gas at
the critical temperatures ; the volume of unit mass of the substance at the critical
temperature and pressure is the critical volume ; and the reciprocal of the volume
is the critical density. Consequently, a substance at the critical temperature and
pressure is at its critical density, and is said to be in its critical state. The critical
constants of a few substances are indicated in Table II., where the atmosphere is
the unit of pressure ; and the volume refers to one gram of the gaseous
substance in litres at 0°, and 760 mm. is the unit of volume ; and the density is
referred to water at 4°.
Table II. — Critical Constants of some Gases.
Substance.
Critical
Critical
Critical
Critical
pressure.
temperature.
density.
volume.
Hydrogen . . . . .
140
-240-8°
0043
000264
Nitrogen .
350
-1460°
0-400
0-00517
Oxygen
50-8
-118-8^
0-650
0-00426
Carbon dioxide .
770
31-35°
0-450
0-00660
Nitrous oxide
74-5
38-8°
0-454
000436
Sulphur dioxide .
78-9
155-4°
0-520
000249
Water
194-6
364-3°
0-208
000386
Carbon disulphide
75-0
273-0°
0-377
0-00900
Air . . .
35-9
-140-7°
0-344
0-00468
Helium
2-3
-267-8°
0-066
000299
Argon
52-9
-117-4°
0-509
000404
Ammonia .
110-3
131-0°
0-239
000481
Methane
45-6
-82-85°
01623
0-00488
Acetylene .
64-5
35-25°
0-2346
000690
Ethylene .
54 0
11-0°
0-210
0-00752
Tin tetrachloride
32-9
318-7°
0-7414
000060
Hydrogen chloride
83-0
52-3°
0-462
0-00520
Nitric oxide
71-2
-93-5°
0-524
000347
Chlorine
93-5
146-0°
0-547
0-00616
In 1883, J. Dewar * showed that the ratio of the critical temperature to the
critical pressure of many gases is nearly proportional to the molecular volume,
and that the quotient TclVe for the common gases generally lies between 3 J and 5.
Other dependent relations have been indicated by E. Aries and W. K. Fielding.
T. Andrews' critical temperature was forshadowed by D. I. Mendeleeff in 1861
in a paper on the expansion of liquids above their boiling points,^ when he said :
The absolute boiling point of a liquid is the temperature at which the cohesion and
heat of vaporization become zero. At this temperature, the liquid changes to vapour
regardless of pressure and volume.
D. I. Mendeleeff's absolute boiling point thus corresponds with T. Andrews' critical
temperature. D. I. Mendeleeff estimated the absolute boiling point of water to be
580°, and of ethyl alcohol, 250°. It is interesting to notice the influence of tem-
perature on carbon dioxide, partly liquid, partly gaseous. Observe the upper surface
of the ^as confined in a glass tube containing partly liquefied carbon dioxide
over mercury at 18°. The surface of the liquefied gas has a sharply defined meniscus.
On raising the temperature, the meniscus of the liquid becomes flatter and flatter
until, at 31-35°, the surface of the Uquid seems to disappear. The sharp line of
demarcation between the liquid and gas vanishes at the critical temperature In
the words of T. Andrews, as the temperature of the liquefied gas approaches 31 :
The surface of demarcation between the liquid and the gas became fainter, lost its
166 INORGANIC AND THEORETICAL CHEMISTRY
curvature, and at last disappeared. The space was then occupied by a homogeneous fluid,
which exhibited when the pressure was suddenly diminished, or the temperature slightly
lowered, a peculiar appearance of moving or flickering striae throughout the entire mass.
At 40°, the tube contains a homogeneous gas. Liquid carbon dioxide cannot exist
at this temperature, however great the pressure. Small tubes of liquid carbon
dioxide for illustrating the phenomena by lantern can be obtained. Thin sections
of quartz found in many granites contain cavities with liquid carbon dioxide which
can be seen to pass through the critical point when the sections are warmed on the
stage of a microscope.
A blue opalescent mist appears in the tube before the meniscus of the liquid
can be detected when the temperature of the gas is gradually lowered. The converse
series of changes occur on heating. According to D. KonowalofE (1902),® the
critical opalescence is due to the scattering of light by fine particles of liquid
spontaneously formed about dust particles ; or, according to M. von Smoluchowsky
(1908), to accidental aggregations of molecules produced by molecular collisions.
The appearance of the blue mist is connected with slight disturbances which have
been observed in the equation of state when applied to observations in the neigh-
bourhood of the critical temperatures. P. de Heen (1888) argues that there are two
kinds of molecules — molecules liquidogeniques, and molecules gasogeniques — and that
the former can persist in the vapour phase. I. Traube (1892) and P. Villard favoured
this view. If P. de Heen means that the pressure of a saturated vapour of a pure
substance, like that of a mixture, depends upon the relative masses of liquid and
vapour phases, the hypothesis is contrary to all experience. This was emphasized
by G. G. Stokes and M. Prud'homme. According to M. von Smoluchowsky, the
ceaseless to-and-fro agitation of the molecular particles of a gas will produce,
spontaneously and continuously, minute inequalities in the density of different
parts. A given cube of dimensions /x, for example, will contain sometimes
more and sometimes less molecules. Usually these differences are inaccessible to
measurement. The case is different when the fluid is not rarefied such as occurs when
it is near the critical state. There is then a permanent condition of fine-grained
heterogeneity where contiguous regions of notably different density are almost in
equilibrium. Owing to molecular agitation, the denser swarms of molecules break
up slowly, and, at the same time, others are forming elsewhere. The opalescence is
produced by the molecular swarms causing a lateral diffraction of the light. The
fluctuations of density increase as the compression increases and are very much more
accentuated with a compressed gas than with a gas of normal density. The smaller
the aggregates, the shorter the wave length of the light undergoing diffraction.
Hence the opalescence may appear blue. M. von Smoluchowsky's theory has been
extended by A. Einstein and confirmed by the work of H. K. Onnes and W. H.
Keesom. It has also been applied to explain the opalescence of liquid mixtures in
the neighbourhood of the point of critical miscibility, and the blueness of the sky.
The relation between the pressure and the volume of, say, carbon dioxide, at
different temperatures — T, Tq, Tj, T2 — is represented diagrammatically in Fig. 6.
The portion of the curve K^T^, or K^Ti, represents the behaviour of the gas when
liquid is present ; the portion K2M2, or KiMi, the behaviour of the gas in the
presence of its own liquid and M2?^2» ^^ -^i?1j ^^^ behaviour of the liquid when no
gas is present. It will be observed that K2^2 o^ ^1^1 i^ ^^^ ^^ of constant
vapour pressure which is horizontal with the v-axis. It illustrates in a graphic
manner the well-known law : At any fixed temperature, the pressure of a gas in
the presence of its own liquid is always the same. The curve TqKqPq represents the
relation between pressure and volume at the critical temperature ; and the curve
T, the relation between p and v at a temperature when the gas does not liquefy.
The line K^KiK^B represents the condition under which the gas, compressed at the
stated temperatures T^, Tj, and T2, begins to liquefy, and hence it is the curve for
saturated vapour, and also the curve for the liquid at its vaporization temperature ;
it is not quite accurately called the dew curve, or ligne de rosee, because a gas
b
THE PHYSICAL PROPERTIES OF GASES
167
MofTIt
under a gradually increasing pressure, jfirst shows signs of liquefaction under con-
ditions represented by a point on this line ; similarly, the line K^M^M^A is called
the boiling curve, or Ugne d' ebullition, because a liquid under a gradually diminishing
pressure first shows signs of vaporization under conditions represented by a point
on this line. Note also that the lines K^A, KqB, and KqPq divide the plane of the
paperinto three regions. Everypointto the right of BKqPq represents a homogeneous
gas ; every point in the region AKqB represents a heterogeneous mixture of
gas and liquid ; and every point to the left of AKqPq, a homogeneous liquid. The
gas in the region KqBVTq is below its critical temperature, and, in consequence, is
sometimes called a vapour as distinct from a gas. The diagram, Fig. 6, thus repre-
sents the conditions of equilibrium of a liquid or gas under different conditions of
temperature, pressure, and volume.
The continuity of the liquid and gaseous states.— It is interesting to
note historically that C. Caignard de la Tour (1822),7 long before Andrews' experi-
ment, noticed that when a liquid is heated
in a sealed tube there is a definite tem-
perature at which the surface of separa-
tion between the gas and liquid disappeared
and the whole contents of the tube become
homogeneous. C. Caignard de la Tour's
experiments thus demonstrate that the
critical temperature is the upper limit of
the liquid state ; and Andrews' experiments
prove that the critical temperature is the
lower limit to the homogeneous gaseous
state. The passage from the one state to
the other proceeds in a continuous manner.
The liguid and gaseous states are con-
tinuous, not abrupt. The properties —
density, surface tension, viscosity, refractive
power, heat of vaporization, compressibility,
etc. — of a liquid gradually lose their distinctive character as the temperature is
raised, until, at the critical temperature, the properties of liquid and gas are the same.
There is no evidence of a change in molecular structure when, say, carbon dioxide
passes from one state of aggregation to another ; nor is there any evidence of a poly-
merization of the molecules when the common gases condense to Hquids. Nitrogen
peroxide, water, and some other substances, however, do appear to polymerize and
form compound molecules on passing from the gaseous to the liquid state of aggression.
The properties of the condensing gases do not then exhibit that continuity
shown by carbon dioxide and other gases which do not polymerize or dissociate.
The difference between liquids and gases below the critical temperature seems
to be a question of molecular attraction. If the molecules of a substance in the
liquid state have essentially the same motions as in the gaseous state, the specific
heat of a vapour should be nearly the same as that of the corresponding liquid.
This is by no means the case. For example, the specific heat of liquid mercury is
twice as large as that of the vapour; and the specific heat of liquid water is three times
that of steam. There is, however, usually less difference between the specific heats,
densities, and coefficients of thermal expansion of solids and the corresponding hquids.
The condensation of binary mixtures of gases. — In a posthumous memoir
presented to the Royal Society in 1886, T. Andrews » showed that some extra-
ordinary phenomena occur when certain binary mixtures of gases are subjected to
a gradually increasing pressure. A mixture of 6 parts of carbon dioxide and one
of nitrogen commences to liquefy at 3-5° under a pressure of 48*3 atm. Here
nitrogen condenses to a liquid at a temperature nearly 150° higher than its critical
temperature, and at 102 atm. pressure, the whole of the nitrogen hquefies along with
the carbon dioxide. The individual properties of the gases are thus profoundly
Volume
Fig. 6. — Pressure - Volume Curves
Carbon Dioxide.
for
168
INORGANIC AND THEORETICAL CHEMISTRY
modified in the presence of other gases which are supposed to be chemically
indifferent. The conception which has crystallized from Dalton's law of partial
pressures, namely, that the two components of a mixture of gases are perfectly
independent of one another, each preserving its own individuality, and each
behaving as if it were an isolated individual, is quite erroneous. The explanation
turns on the existence of a definite relation between the composition of a condensed
liquid and of the vapour during, say, the distillation of a binary mixture of two
volatile liquids which exert no chemical action on one another. L. P. Cailletet dis-
covered this remarkable phenomenon during his Experiences sur la compression
des melanges gazeux in 1880. If a mixture of one volume of air and nine volumes
of carbon dioxide be subjected to a gradually increasing pressure at about 2°, the
gas begins to liquefy at a pressure of about 72 atm.; and on increasing the pressure,
still keeping the temperature constant, the liquid again passes into the gaseous
state when the pressure reaches 149 atm. ; and the liquid does not reappear again
however great the pressure. If the pressure at which the liquid appears and
disappears be plotted with the corresponding temperature, we get the dew curve
BKC, Fig. 7. For the same abscissa Ti, there are two ordi-
nates, pi and p2, between which the mixture is in a hetero-
geneous condition. At temperatures above Tq, no condensa-
tion will occur at all ; below Ti, only normal condensation
takes place ; at temperatures between Ti and Tq, both
normal and retrograde condensation as P. Kuenen (1893)
named the phenomenon, will occur. The dotted line AC repre-
sents the boiling curve ; above AC, the system will be in the
liquid state. K corresponds with the critical temperature of
Fig. 7.— Diagrammatic, the mixture ; C is called a plait-polnt. For mixtures of two
gases, therefore, (1) there is a critical zone of temperature above
which complete or partial liquefaction is impossible. (2) Within the temperature of
the critical zone itself a part of the mixture can be brought by pressure to the liquid
state, and in the region of retrograde condensation, condensation is produced by
diminution of pressure, and evaporation by an increase of pressure. The phenomena
with mixtures thus appears quite different from what obtains with single gases. (3)
Below the temperature of the critical zone, the whole of the mixture can be
liquefied by pressure.
The phenomenon occurs only with mixtures of a certain composition ; above
and below these limits, the dew curves are quite normal. The curves can be taken
in three dimensions with the three variables — pressure, volume, and temperature.
The two dew points of a given mixture approach one another as the temperature
rises. Thus, F. Caubet (1901) found with a mixture of 74' 58 per cent, of carbon
dioxide and 254:2 per cent, of sulphur dioxide gases :
Table III. — Betrogbade Condensation of Mixtures of Carbon Dioxide and
Sulphur Dioxide.
p
c
..^^-'
/
To
.><
70"
72°
74-2»
Pressure.
Volume of liquid.
Pressure.
Volume of liquid.
Pressure.
Volume of liquid.
610
0
600
0
75-0
0
66-2
*
70-0
*
78-2
*
69-4
0-066
74-5
0-066
80-6
0-066
77-0
0-164
84-8
0-164
83-4
0-099
83-4
0-250
87-8
0-184
86-2
0-066
87-8
0-428
89-6
0-099
88-0
♦*
89-6
0-263
89-8
Xc*
93-0
0
90-0
*♦
95-6
0
105-0
0
91-0
0
105-0
0
105-0
0
One asterisk * represents the first dew point — ler point de roaie ; and two asterisks **
the Becond dew point — 2e jioinl de rosee.
THE PHYSICAL PROPERTIES OF GASES 169
L. P. Cailletet and E. Mathias (1866) » found empirically that the mean values
of the densities of a liquid, D^, and of its saturated vapour, Z)^, at a constant
pressure, vary with the temperature in a very simple manner. If the densities be
plotted with the temperature, a closed curve AKB, Fig. 8, is obtained. The
mean values of the densities of the co-existing liquid and vapour, plotted against the
temperatures, fall on a straight line, KC, Fig. 8. The density of the liquid decreases
while that of the vapour increases as the temperature rises, until, at the critical
point, K, the two densities are equal to the critical density. Hence, the rule was
called the hi du diametre rectiligne, or Cailletet and Mathias' law of rectilinear
diameters. According to this empirical rule, the mean
values of the densities of a liquid and of its saturated
vapour is a linear function of the temperature ; so that
if D represents the mean value of the two densities,
D=a-{-bd, where a and h are constants, and 6 denotes the
temperature on the centigrade scale. For argon, the equa-
tion of the mean density curve is Z)=0'209 56 —0-00262 35^ ; ~'^°°
and in cases where the curve has a slight curvature, the - '^o'
equation D=a-\-hd-}-cd^ usually represents the observed ^
results. The law has been tested with carbon dioxide, o o 4. q oa
sulphur dioxide, nitrous oxide, hydrocarbons, alcohols, car- Yiq. 8. Variations in the
bon tetrachloride, tin tetrachloride, oxygen, helium, xenon, Densities of Co-existing
etc. In all cases the empirical law was found to be re- Liquidand Gaseous Oxygen,
markably exact, except in the case of those substances —
e.g. the alcohols, fatty acids, etc. — which are known to exhibit molecular aggrega-
tion. The curvature of the line is taken to indicate molecular association, although
the absence of curvature does not necessarily mean that molecular association is
absent. E. Mathias and H. K. Onnes' results for oxygen are indicated in Fig. 8 ;
the curve is plotted from the following observations :
Temperature . . -2104° -1820° -154-5° -1402° -1299° -123-3° -120-4**
Density of liquid, jD< . 1-2746 1-1415 0-9758 0-8742 0-7781 0-6779 0-6032
Density of vapour, A' . 0-0001 0-0051 0-0385 0-0805 0-1320 0'2022 0-2701
Mean density, D . . 06373 05733 05072 0-4773 0-4550 0-4400 0*4366
The mean densities calculated from the linear expression, Z)=0- 1608— 0*002265^,
do not deviate from the observed values more than ± 0*003. The law does not hold
good for substances whose molecules in the liquid and gaseous states have a different
complexity.io
References.
1 B. Jones, The Life and Letters of Faraday, London, 1. 308, 1870.
2 T. Andrews, Phil. Trans., 159. 575, 1869 ; B. A. Eep., 76, 1861 ; W. A. Miller, Chemical
Physics, London, 1863.
3 M. Berthelot, Compt. Bend., 30. 166, 1850.
* J. Dewar, Phil. Mag., (5), 18. 210, 1884 ; Nature, 28. 561, 1883 ; E. Arifes, Compt, Rend.,
166. 193, 1918 ; W. R. Fielding, Chem. News, 117. 379, 1918.
5 D. I. Mendeleeff, Liehig's Ann., 119. 1, 1861.
« D. Konowaloff, Ann. Physik, (4), 10. 360, 1903; (4), 12. 1160, 1903; L Traube, tb., (4),
8. 289, 1902; M. von Smoluchowsky, ib., (4), 25. 205, 1908; Bull. Acad.. Cracow, 1057, 1907;
P. de Heen, Recherches touchant la physique comparee et la theorie des liquides, Paris, 1888 ; Bull,
Acad. Belgique, (3), 25. 695, 1893 ; P. Villard, Ann. Chim. Phys., (7), 10. 429, 1897 ; A. Einstein,
Ann. Physik, (4), 33. 1275, 1910; (4), 36. 1572, 1910; W. H. Keesora, ib., (4), 35. 591, 1911 ;
H. K. Onnes and W. H. Keesom, Comm. Lab. Phys. Leiden, 104, 1908 ; Lord Raylcigh, Phil.
Mag., (4), 41. 107, 1871 - (5), 47. 375, 1899 ; M. Prud'homme, Journ. Chim. Phys., 14. 445,
1917.
' C. Caignard de la Tour, Ann. Chim. Phys., (2), 21. 127, 178, 1822 ; (2). 22. 140,
1823
«'t. Andrews, Phil. Trans., 178. 45, 1887 ; L. P. Cailletet, Compt. Rend., 90. 210, 1880 ; Journ.
Phys., (1), 9. 192, 1880 ; (2), 2. 389, 1883 ; J. P. Kuenen, Phil. Mag., (5), 40. 173, 189o ; Zext.
phys. Chem,, 11. 38, 1893 ; 24. 667, 1897 ; F. Caubet, ib., 40. 257, 1902 ; Liquefaction des melanges
170 INORGANIC AND THEORETICAL CHEMISTRY
gazeux, Paris, 1901 ; J. Dewar, Proc. Boy. 8oc., 30. 538, 1888 ; P. Duhem, Journ. Phys. Chem., 1.
273 1897.
» E. Mathias and L. Cailletet, Journ. Phys., (2), 5. 679, 1886 ; Compt Rend., 102. 1202, 1886 ;
104. 1563, 1887 ; E. Mathias, ib., 112. 85, 1891 ; Ann. Fac. Sciences Toulouse, (1), 6. 1, 1892 ;
S. Young, Phil. Mag., (5), 33. 263, 1892 ; (5), 50. 291, 1900 ; Journ. Chem. Soc., 59. 37, 126,
929, 1891 ; K. Tsuruta, Phys. Rev., (1), 10. 116, 1900; E. Mathias and H. K. Onnes, Comm.
Lab. Phys. Leiden, 117, 131, 1911.
1" P. E. Guye, Arch. Sciences Oenhve (3), 31. 176, 1894 ; E. Mathias, Le point critique, des
corps purs, Paris, 1904.
CHAPTER V
COMBINATION BY VOLUME
§ 1. Gay Lussac's Law of Combining Volumes
Omnia mensura et numero et pondere disponsuisti — Thou hast ordered all things in
measure, and number, and weight. — Liber Sapientiae.
Not very long after John Dalton had directed the attention of chemists to the
relations subsisting between the weights of bodies which combine in different
proportions, J. L. Gay Lussac i established a similar correspondence between
volumes of combining gases. A. von Humboldt, the naturalist and eirplorer,
collected samples of air from different parts of the world, and with the aid of
J. L. Gay Lussac, analysed the different samples with the idea of finding if the
composition of air was variable or constant. J. L. Gay Lussac used Cavendish's
process — explosion of a mixture of air and hydrogen gas. As a preliminary, A. von
Humboldt and J. L. Gay Lussac investigated the proportion by volume in which
hydrogen and oxygen combine, and found the ratio of hydrogen to oxygen, by
volume, to be nearly as 2 : 1. If either hydrogen or oxygen was in excess of these
proportions, the excess remained after the explosion, as a residual gas. A. von
Humboldt and J. L. Gay Lussac (1805) found :
Vols, of oxygen.
Vols, of hydrogen.
Vols, of residue.
100
300
101*3 hydrogen
200
200
101-7 oxygen
After making corrections for impurities, etc., in the gases, J. L. Gay Lussac and
A. von Humboldt stated that " 100 volumes of oxygen required for complete satura-
tion 199-89 volumes of hydrogen, for which 200 may be put without error."
A. Scott (1893) found, as the result of twelve experiments on the volumetric com-
position of water, that oxygen and hydrogen combine very nearly in the ratio
1 : 2-00245 by volume.
Struck by the simplicity of the relation thus found, J. L. Gay Lussac (1808)
followed up the subject by numerous experiments with different gases. As a
result, in his Memoire sur la comhinaison des substances gazeuses, les unes avec les
autres (1809), he concluded that " gases always combine in the simplest proportions
by volume." For instance, one volume of hydrogen combines with one volume
of chlorine forming two volumes of hydrogen chloride ; two volumes of hydrogen
combine with one volume of oxygen forming two volumes of water vapour (which
condenses to liquid water if the temperature be below 100°).
There are slight deviations with the gases which show deviations from the
laws of Boyle and Charles, but the experimental results are such as to leave no
doubt that J. L. Gay Lussac's generalization is valid, and accordingly, we define
Gay Lussac's law : when gases react together, they do so in volumes which bear
a simple ratio to one another, and to the volume of the gaseous product of the
action. It is assumed, of course, that the initial and final products of the reaction
are under the same conditions of temperature and pressure.
The remarkable way in which elements unite by weight was traced to a
peculiarity in the constitution of matter ; so here, we are tempted to make a
similar quest. It follows at once (1) if elements in a gaseous state unite m simple
171
172 INORGANIC AND THEORETICAL CHEMISTRY
proportions by volume, and (2) if the elements also unite in simple proportions
by atoms, then the number of atoms in equal volumes of the reacting gases must
be simply related. With John Dalton, in his A New System of Chemical Philosophtj
(Manchester, 1808), let us make a guess. Assume that equal volumes of the
different gases under the same physical conditions contain an equal number
— say n — of atoms. Then, when two volumes of hydrogen react with one volume
of oxygen to form two volumes of steam, we have 2n atoms of hydrogen reacting
with r? atoms of oxygen to form 2n " compound atoms " of steam. Hence, two
atoms of hydrogen react with one atom of oxygen to form two " compound atoms "
of steam. In that case, every atom of oxygen must be split into half to make
two " compound atoms " of steam. This contradicts the fundamental postulate
of the atomic theory expressed in John Dalton's aphorism : " Thou knowest no
man can split an atom," meaning that atoms are assumed to be indivisible in
chemical reactions. 2 Similar contradictions are encountered in nearly every case
of combination between gases, hence J. Dalton regarded J. L. Gay Lussac's law
as untenable : Equal volumes of homogeneous gases, under like conditions of tem-
perature and pressure, do not contain the satne number of atoms. There is such a marked
uniformity in the deportment of elementary and compound gases with respect to
variations of temperature and pressure, that it is not very probable any essential
difference will be found in their constitution.
References.
^ J. L. Gay Lussac and A. von Humboldt, Journ. Phys., 60. 129, 1805; J. L. Gay Lussac, Mem.
Soc. Arcueil, 2. 207, 1809 ; Alembic Club Reprints, 4, 1893 ; A. Scott, Phil. Trans., 184. 543, 1893.
2 W. C. Henry, Memoirs of the Life av^, Scientific Researches of John Dalton, London, 1854 ;
Alembic Club Reprints, 4, 1893.
§ 2. Amadeo Avogadro's Postulate
Advances in knowledge are not commonly made without the previous exercise of some
boldness and licence in guessing.— W. Whewell.
J. J. Berzelius i thought it strange that J. L. Gay Lussac should have contented
himself with having determined the combining ratios of gaseous substances, and
should make no attempt to extend his discovery. Clearly with J. Dalton the faculty
of speculation was predominant, and with J. L. Gay-Lussac experimentation.
An epoch-making memoir entitled, Essai d'une maniere de determiner les masses
relatives des molecules elementaires des corps, et les proportions selon lesquelles elles
entrent dans les combinaisons, was published in 1811 by Amadeo Avogadro,^ an
Italian physicist. In his memoir A. Avogadro said —
J'ai propose \ine hypothese pour expliquer le fait decouvert par M. Gay Lussac, que
les volumes des substances gazeuses qui se combinent entre elles, et des gaz composes qui
en r68\iltent, sont toujours dans les rapports tr^s simples entre eux.
He pointed out that the difficulty with Dalton's hypothesis can be avoided if we
distinguish clearly between elementary atoms and the small particles of a gas.
Assume that the small particles of a gas are aggregates of a definite number of
atoms ; then, using A. Avogadro's own words :
Les molecules constituantes d'un gaz simple quelconque, c'est-a-dire celles qui s'y
tiennent a une distance telle a ne pouvoir exercer leur action mutuelle, ne sont pas form^es
d'une seule molecule el6mentaire mais r6sultent d'un certain nombre de ces molecules
r6unies en une seule par attraction.
A. Avogadro called these aggregates molecules, in order to distinguish them from
the ultimate atom. His actual term was molecules constituantes or molecules
integrantes — the former term was used for molecules of elements, the latter for
COMBINATION BY VOLUME
173
molecules of compounds. The one term molecule (the diminutive form of the
Latin word jnoles, a mass) is now applied to both Avogadro's inolecules constitiiantes
and molecules integr antes. Each molecule of an elementary gas is supposed to
contain the same number and kind of elementary atoms. What J. Dalton called
atoms A. Avogadro called molecules elementaires. The word " atom " does not
occur in the latter's memoir. The modern meanings of the terms atom and
molecule were clearly stated by A. M. Ampere 3 in 1832, and by A. Gaudin in the
same year. Some years later these distinctions were emphasized by A. Laurent
(1846) and employed in his posthumous book Methode de chiinie (Paris, 1854).
A. M. Ampere used the term particle for an aggregate formed by the juxtaposition
of molecules. He said :
In the passage from liquid to the gaseous state, the molecules are separate from one
another ; and conversely, in passing from the gaseous to the liquid state, the molecules
are drawn together. In the passage from the liquid to the solid state, I think that two
or more molecules are drawn together. Mechanical forces alone can separate the particles ;
chemical forces are required to split the molecules.
For the sake of simplicity, assume that each molecule of hydrogen gas is com-
posed of two atoms of hydrogen, and make a similar assumption for oxygen gas ;
and assume with A. Avogadro that equal volumes of all gases, at the same
temperature and pressure contain the same number of molecules. This
postulate is now known as Avogadro's hypothesis. In A. Avogadro's own words :
L'hypothese qui se presente la premiere a 6gard et qui parait mSme la seule admissible,
est de supposer que le nombre des molecules integrantes dans la gaz quelconque est toujoura
le meme a volume 6gal, ou est toujours proportionnel aux volumes. . . .
Suppose that two volumes of hydrogen contain 2w molecules of hydrogen, then one
volume of oxygen will contain n molecules. These react to form 2w molecules of
steam — each molecule of steam contains two atoms of hydrogen and one atom of
oxygen. The idea can be more clearly illustrated by means of the subjoined
diagrams. Each square represents one volume of a gas. Each volume contains
n molecules. We do not know the numerical value of n, but, for the sake of
simpUcity, take n=4. It makes no difference to the final conclusion what
numerical value we assign to n. Then we have :
+
t ••!
Hence, although the atoms of oxygen cannot be split, yet a 2-atora molecule of
oxygen can be subdivided so that one atom of oxygen enters the composition of
each of two molecules of water. Again, with hydrogen and chlorine.
+
<30
-i-
8 <P
a, . t <SG>
%A.
Diagrams similar in principle to these were used by M. A. Gaudin about 1832 in
his Recherches sur la structure intinie des corps inorganiques dejinis. It must not be
supposed for one moment that what may be called Gaudin's diagrams are intended
as pictures of the actual molecules. They are to be regarded as aids to the under-
standing of how Avogadro's hypothesis has led chemists to conclude that the mole-
cules of gaseous elements are really compounded atoms, and how Avogadro's
hypothesis reconciles the observed volume relations during the combination of
gases with the atomic theory.
It has been assumed for the sake of simplicity, that the molecule of water con-
tains three atoms, and that each molecule of hydrogen and oxygen contams two
atoms. As a matter of fact, all we can infer from the observed facts is that the
molecule of oxygen is split into halves, and, in the absence of evidence to the contrary,
174 INORGANIC AND THEORETICAL CHEMISTRY
we assume for every substance the simplest molecular structure consistent with the
observed facts.
A. Avogadro extended J. Dalton's atomic hypothesis and adapted it particularly
to^gases. We owe to the former the conception of two orders of minute par-
ticles : (1) the atom or unit of chemical exchange ; and (2) the molecules are the
smallest particles oi an element or compound which exist free in a gasr
This definition of a molecule is usually extended into the less satisfactory definition :
A molecule is the smallest [particle of an element or compound which exists
in a free state ; otherwise expressed, the molecules of an element or compound
are particles so small that the specific properties of the substance depend
upon the particles remaining intact. Hence, if molecules be subdivided the parts
no longer have the specific properties of the original substance. If the molecules
of steam, H2O, be subdivided, two atoms of hydrogen and one atom of oxygen
would be formed per molecule ; the atoms unite in pairs to form molecules.
Diatomic molecules for gaseous chlorine, hydrogen, and oxygen at ordinary
temperatures furnish a satisfactory explanation of what we know to-day, but it
is possible that at some future date, the evidence will compel us to consider these
molecules to be tetra- or hexa-atomic. This will not materially afiect the principle
as indicated above. The molecule of mercury is supposed to be monatomic ; and
the molecule of sulphur, hexatomic.
In 1814, A. M. Ampere advocated views similar to those of A. Avogadro, but he
compHcated the latter's simple hypothesis by an unsuccessful attempt to apply
his conception of molecules to crystalUne solids. Avogadro considered Ampere's
extension unjustifiable. A. M. Ampere clearly emphasized the hypothetical nature
of A. Avogadro's conception in a letter to M. le Comte Berthollet in 1814, when he
said : If the consequences of the hypothesis be confirmed by further experiments,
and the hypothesis be in agreement with known facts, elle pourra acquerir un degre
de prohdbilite qui approchera de ce qu^on nomme en physique la certitude. Increasing
knowledge has made A. Avogadro's hypothesis more and more probable ; it has
been tested in hundreds of experiments, and never found wanting. The hypothesis
has done such good service in giving a rational explanation of many different
phenomena that it has been accepted as a fundamental truth. It gave chemists a
clear definition of the atom, a method of determining the relative weights of the
atoms, and of estimating the number of atoms in the molecule.
References.
1 J. J. Berzelius, Essai sur la thiorie des proportions chimiques et sur Vinfluence chimique de
V ehctricite, Paris, 14, 1819 ; A. N. Meldrum, Avogadro arid Dalton, Edinburgh, 14, 1904.
2 A. Avogadro, Jmirn. Phys., 73. 58, 1811; 78. 131, 1814; Alembic Club Reprints, 4.
1893 ; Mem. Accad. Torino, 26. 440, 1864 ; J. Guareschi, Amadeo Avogadro e la teoria molecolare,
Torino, 1901.
8 A. M. Ampere, Bibl. univ. Geneve, 49. 225, 1832; Ann. Chim. Phys., (1), 90. 43, 1814;
(2), 58. 432, 1835 ; M. A. Gaudin, ib., (2), 52. 113, 1833 ; Becherches sur le groupement des atomes
dans les moliculeset surles causes les plus intimes des formes cristallines, Paris, 1847 ; L^ architecture
du monde des atomes, Paris, 1873 ; E. Grimaux, Thiories et notations chimiques, Paris, 1884 ; E.
Erlenmeyer, Zeit. Chem., 6. 610, 1863.
§ 3. The Relative Weights of the Molecules
In order to bring into harmony all the branches of chemistry, we must have recourse
to the complete application of the theory of Avogadro in order to compare the weights and
the numbers of the molecules. — S. Cannizzaro.
John Dalton in 1807 raised the query : " Are there the same number of particles
of any elastic fluid in a given volume and under a given pressure ? " It is curious
that in answering " No," J. Dalton ^ abandoned an hypothesis which afterwards
COMBINATION BY VOLUME 175
proved to be one of the most fruitful suggestions in the development of chemistry,
for, under the name of Avogadro's hypothesis, it has correlated what appeared antago-
nistic and contradictory, and has harmonized what seemed discordant and confused ;
it made Dalton's atomic hypothesis a clear, intelUgible, and fertile theory. As
C. A. Wurtz said in his The Atomic Theory (London, 1880), had it not been for this
development, J. Dalton's hypothesis was in a fair way of being sentenced to steriUty
and oblivion. A fellow countryman of A. Avogadro, namely S. Cannizzaro, seems
to have seen, more clearly than any other, the importance of A. Avogadro's hypo-
thesis in putting J. Dalton's on a firm basis.
S. Cannizzaro' s ideas were first pubHshed in a letter to S. de Luca embodying
a Sketch of a Course of Chemical Philosophy ^'^ given in the Royal University of Geneva
in 1858. Before S. Cannizzaro published his paper, rank confusion prevailed in
chemical Hterature. The terms atomic weight, molecular weight, and combining
or equivalent weight were used and abused in every conceivable way. J. B. A.
Dumas lost faith in the atomic theory and wrote in despair :
Si j 'en etais le maitre j 'eff acerais le mot atome de la science, persuade qu'il va plxis loin
que I'experience : et jamais en chimie nous ne devons aller plus loin que I'exp^rience.
Avogadro's hypothesis was necessary for salvation ; it lay dormant in chemical
literature for nearly half a century ; S. Cannizzaro brought the awakening, and
showed chemists that the atom must be defined with reference to A. Avogadro's
molecule. After reading S. Cannizzaro's pamphlet, Lothar Meyer (1860) thus
describes his own conversion : " the scales fell from my eyes, my doubts disappeared,
and a feehng of tranquil security took their place." A. Avogadro's hypothesis
was thus made the basis of the current theory of chemistry.
By definition, the relative density of a gas is a number which represents how
much heavier any volume of the gas is than an equal volume of the standard gas —
generally hydrogen — measured at the same temperature and pressure — generally
at 0° and 760 mm. pressure. Thus, the relative density of steam is 8*95. This
means that any volume, say a litre of steam, is nearly nine times as heavy as the
same volume of hydrogen.
Strictly speaking, the density of a gas is the weight of 1 c.c. of the gas at 0° and 760 nmi.
The density of a gas is usually expressed in terms of a litre of the gas because the number
representing the weight of 1 c.c. would be inconveniently small. One litre of hydrogen
at n.p.t. weighs very nearly 0*0896 grm. " So important is this standard weight-unit,"
said A. W. Hofmann in hia Introduction to Modern Chemistry (London, 1865), " that a name
is needed to denote it." He suggested crith {Kpie-n, a barley com, or small weight) to
denote the weight of a litre of hydrogen at n.p.t. The weight of the same volume of
oxygen would then be 16 criths, of nitrogen 14 criths, etc. The term has now dropped
out of use, although for a time it served a useful purpose.
By Avogadro's hypothesis, equal volumes of gases, under like conditions of
temperature and pressure, contain the same number of molecules, consequently,
the specific gravity or relative density of a gas is proportional to its molecular
weight. Let n represent the number of molecules in a volume v of each of two
different gases, and if the molecules of each gas are all alike with the respective
molecular masses M^ and i/g, then the one gas will have a mass nMi and the other
a mass wMg. Let the densities of the respective gases be Di and D.^y then since
density denotes the mass of unit volume, D^-.D^ — nM-^v ; nM^jv ; that is,
Z>i : i>2=^i : ^2 or
Di Ml Ml Mg Q.
Drw^DTD, ' ' ' • ^^
or the relative densities of any two gases are proportional to their respective mole-
cular weights ; and the quotient of the molecular weight by the density is the
same for all gases. It is convenient to employ the term molecular volume for the
quotient obtained by dividing the molecular weight M of a gas by its relative density
176 INORGANIC AND THEORETICAL CHEMISTRY
D\ consequently, from the second of equation (1), the molecular volumes of all
gases are the same.
If we accept this deduction, it enables us to determine the molecular weights of
gases, once we have fixed an arbitrary standard for the density. Cannizzaro's
unit, hydrogen==2, is frequently taken as the standard, or else hydrogen unity,
that is, as S. Cannizzaro expressed it, " the quantity of hydrogen contained in a
molecule of hydrogen chloride " is taken as unity. The determination of the
molecular weight of a gas is thus reduced to a laboratory measurement — the
determination of the relative density of the gas. Methods for measuring vapour
densities are outlined later.
It has been shown within certain limitations, that the numerical values for the
molecular weight and relative density of a gas referred to the standard hydrogen , 2,
are the same. That is,
Molecular weight = Relative density (H2=2) . . • (2)
For example, the observed density of steam is 18 (H2=2), the molecular weight ot
steam is therefore 18 likewise. Again, if the relative density be referred to the
standard hydrogen unity, or oxygen 16, the relative density is half the molecular
weight ; or the molecular weight is twice the density.
Molecular weight =: 2 X Relative density (H=l) . . • (3)
For instance, the density of steam is 9 (H=l), the molecular weight is therefore
twice 9 or 18 as before. When the relative density is referred to oxygen 32, as is
common in recent years, it is virtually assumed that there is an imaginary gas whose
relative density is unity ; and to avoid the hypothesis implied in the term molecular
weight, the term molar weight is applied to the relative density of a gas referred
to oxygen 32.
If the relative density be determined, as is frequently the case, with reference to
the standard air unity, then, since the density of air with respect to hydrogen is
28-75 (H2=2) ; or with reference to oxygen 28-98 (02=32), it follows that
Molecular weight=28- 75 X Relative density (Air unity) . . (4)
For example, the relative density of cyanogen is 1-806 (air unity), the molecular
weight is therefore 1-806x28-75=51-92 (H2=2). This is in agreement with the
formula C2N2.
It is unfortunate that these different units are employed, even though all give the
same final result. It shows the necessity for clearly understanding the particular
meaning of terms employed before elaborating an argument. The method of deter-
mining the relative density of a gas by weighing a globe full of gas and then full
of air, led to the use of air as a standard of reference. Thus, J. L. Gay Lussac (1815)
found a 2J-litre globe weighed w-\-2'l^ grms. when filled with air, and w+4-946
grms. when filled with cyanogen ; consequently the relative density of cyanogen,
air unity, is 4-946/2-738=1-806. The custom of referring all gas densities to air
as a standard was gradually adopted. The system has been shown in recent years
to be faulty when very accurate results are required because there are undoubtedly
slight variations in the composition of air, and this causes the density of air — ^the
standard of reference — to vary in a corresponding manner.
If the specific gravity of a gas is to be referred to water as standard^ the relative
density, air unity, is multiplied by the weight of one c.c. of air, viz. 0-001293;
by 0-00008996 if hydrogen unity be the standard; and by 000004469 if oxygen be 32.
Thus, the relative density of carbon dioxide is 1-57 (air unity) ; 22 (hydrogen
unity) ; and 44 if oxygen 32 be the standard. Hence, also, the specific gravity
with respect to water as standard is 1-57x0-001293=0-00203.
It will be noted that if W denotes the weight of a Htre of a gas of molecular weight ilf ,
and D denotes the relative density, air lanity, >r = |M X 0*08996 ; D=M/2S15, and
therefore 100{W —D)=M, or the molecular weight of a gas, is nearly 100 times the difference
between the weight of a litre of the gas at n.p.t. and the relative density of the gan, air unity.
COMBINATION BY VOLUME 177
Returning to S. Cannizzaro's important paper, S. Cannizzaro gave the following
numbers, among others, for the densities of the different gases referred to hydrogen
taken as 2, or to a semi-molecule of hydrogen taken as unity :
Relative densities.
Hydrogen •••....., 2*0
Oxygen [ [ 32-0
Chlorine ......... 71.q
Nitrogen 28*0
Water vapour •••..... Ig'O
Hydrogen chloride . . . . , . . 3g.5
If, therefore, the molecules of hydrogen, oxygen, nitrogen, and chlorine contain
two atoms, the atomic weights of these gases will be half the respective molecular
weights. Hence, making a selection from S. Cannizzaro's tables :
Table I. — S. Cannizzaro's Table of Atomic Weights.
Element.
Relative density of gas.
Atomic Weight, or Density -r 2.
Hydrogen .
Oxygen
Chlorine
Nitrogen
2
32
71
28
1-0
16-0
35-6
140
In the case of compounds, if the molecule of hydrogen chloride contains an
atom of chlorine and an atom of hydrogen, the molecular weight will be 35*5+1
= 36' 5; and the molecule of water vapour containing two atoms of hydrogen and
one atom of oxygen — or, as A. Avogadro (1811) expresses it, une demi-molecule
d'oxygene avec une molecule ou, ce qui est la mefne chose, deux demi-molecules d'hydrogene
— will have a molecular weight of 16+2=18. Hence, given the molecular weight of
a compound gas, and the weights of the atoms of all but one of the elements, it is
possible to compute the weight of the atom or atoms of that element in the molecule
in question. The 7nodus operandi will be discussed in two later sections.
A. Avogadro explicitly guarded against the assumption that the number of
constituent atoms in the molecule of a gas must always be 2. There is really nothing
in the facts to justify the assumption that the atoms themselves are simple particles.
For all we know to the contrarj^, the atoms may be clusters of n particles. Indeed,
we shall soon review some cogent evidence which has led many chemists to abandon
Newton's solid impenetrable atoms, and to infer that Dalton's atoms are not
nature's irreducible minima. Even if this inference be valid, each cluster of
n particles which forms an atom has a definite weight — atomic weight — and enters
into and is expelled from chemical combination as if it were a simple particle. If an
atom be a cluster of particles, each cluster, so far as we can tell, has up to the present
time behaved in chemical reactions as if it were an individual particle. The actual
weight of a molecule is certainly not the molecular weight. When it is said that
the molecular weight of hydrogen chloride is 36' 5, this number simply means that
we have conventionally agreed to fix the molecular weight of, say, oxygen as 32
units, and that the molecular weight of hydrogen chloride is to that of oxygen as
36-5: 32. Consequently, like atomic weights, molecular weights denote ratios,
they are relative not absolute numbers.
To deduce Avogadro' s law from the relation between the relative densities and the
molecular weights of the gases. Let Mj and M^ denote the weights of the molecules
of two gases— A and B respectively ; further, let n^ and n^ respectively denote the
number of molecules in unit volumes of the two gases. The weights of unit volumes
{i.e. the densities) of the two gases will be ilf i??i and Mg^s- The observed fact is
that the molecular weights {M^ and M^) of the gases are proportional to the densities
(Mi^i and M^n^) of the gases ; or M^ni : ilf 2^2=^1 ' ^2» from which it follows
VOL. I. N
178 INORGANIC AND THEORETICAL CHEMISTRY
that in unit volumes of the two gases ni=n2. This is the symbolic way of stating
Avogadro's rule. Hence, it has been claimed that Avogadro's postulate can be
deduced from the relation between the molecular weights and the densities of two
gases. It is easy to be misled by the apparent precision and rigorous accuracy
conveyed to the mind by reasoning expressed in mathematical symbols, and to
assume that the conclusions of such reasoning are certainties. Some affirm, on
the strength of the simple demonstration just indicated, that " Avogadro's hypo-
thesis is true." The reasoning is perfectly sound, but what about the premises,
or statements upon which the reasoning is based ? Avogadro's method for the
determination of molecular weights tacitly assumes that the hypothesis is true.
Hence, if the mathematical demonstration be employed to prove that Avogadro's
hypothesis is true, the argument proceeds in a vicious circle. It is assumed in
the premises what is " proved " in the demonstration. A conclusion proved by
mathematics cannot be any more certain than the premises on which the reasoning
is based.
Refebbnobs.
1 J. Dalton, A New System of Chemical Philosophy, Manchester, 1808.
2 S. Caimizzaro, Nuovo Cimento, 7. 321, 1858; Journ. Chem. Soc., 25. 941, 1872 ; Ostwald's
Klassiker, 30, 1891 ; Alemhic Clvh Reprints, 18, 1858 ; E. von Meyer, Journ. prakt. Chem., (2),
83. 182, 1911 ; L. Meyer, Ostwald's Klassiker, 30, 1891 ; J. B. A. Dumas, Lerons sur la philosophic
chimique, Paris, 1836.
§ 4. The Formulse of Compounds
Avogadro's hypothesis affords a bridge by which we can pass from large volumes of
gases, which we can handle, to the minuter molecules, which individually are invisible and
intangible.— W. A. Shenstone.
Since S. Cannizzaro's time, an enormous number of molecular weights have been
determined by the vapour density method. If the molecule cannot be decomposed,
it must be assumed that it is composed of one kind of matter only. If the substance
is compound, it must be analysed so as to find the ratio, by weight, of its component
elements referred to the oxygen standard (16). For instance, suppose that the
analysis of a gaseous compound furnished : Nitrogen, 82* 35 per cent. ; hydrogen,
17'65 per cent. Using S. Cannizzaro's data, if hydrogen has an atomic weight
of unity and nitrogen 14, the compound has the equivalent of 17* 6 5-^1, or 17*65
hydrogen atoms for every 82* 35^14 nitrogen atoms; or 5'9 nitrogen for every
17*65 hydrogen atoms. By hypothesis we cannot have fractions of atoms. The
nearest whole numbers are 3 hydrogen atoms for one nitrogen atom. Since the
sum of the atoms in the compound must represent the molecular weight, it follows
that the molecular weight must be 3n-\-lin, that is, the molecular weight is 17x1 ;
17x2 ; 17x3; . . . ot I7n. The formula is NnHsn. We can get no further until
we know the molecular weight. If the vapour density of the compound (hydrogen
=2) be 17, the molecular weight is 17. Hence, 17=17w, or n=l. The compound
analysed can therefore be represented by the formula NH3.
Examples.- — (1) E. W. Morley (1895) found, in some careful experiments on the synthesis
of water: Hydrogen used, 3-7198 grms. ; oxygen used, 29'5335 grms. ; water formed,
33-2530 grms. That is, one part by weight of hydrogen combines with 7-94 parts by weight
of oxygen to produce 8-94 parts by weight of steam. A molecule of steam must contain n
atoms of hydrogen, because parts of an atom do not take part in chemical changes. Hence,
n parts by weight of hydrogen per 7-94« parts by weight of oxygen give a molecule of
steam of weight 8-94w. This all follows from the atomic theory. To apply Avogadro's
hypothesis, with Cannizzaro's standard, the density of the steam must be determined.
It lies between 16 and 20. It is difficult to determine the number exactly. If n = l, the
density of the steam molecule will be near 8-94. This does not agree with the observed
density 16 to 20. If n~2, the density of the steam will be 17*88 ; and if n = 3, the density
of steam will be 26'82. Hence, w=2. This means that each molecule of water vapour
I
COMBINATION BY VOLUME 179
contains 2 atoms of hydrogen, atomic weight 1, and one atom of oxygen, atomic weight
15-88 ; or if we make our imit oxygen = 16, the atomic weight of hydrogen will be 1008.
(2) Two different compounds have the same ultimate composition, namely : carbon
92-31 per cent., hydrogen 769 per cent., but the one has a relative density 26, and the
other a relative density 78 (H=2). What is the formula of each compoimd ? There are
92-31^12=7-7 carbon atoms per 7-7-^1=7-7 hydrogen atoms; but we cannot have
fractions of atoms, hence dividing by 77 we get the ratio 1:1. That is, the formula of
the compoimd is CnHn- The molecular weights of this series of compoimds is (12 + l)n
or 13«. If w = 2, the molecular weight will be 26. Hence, one of the compounds is CjHj,
and the other is CgHg.
In calculating formulae for substances which cannot be vaporized, and one of
the methods to be described later cannot be applied, it is usual to assume that the
molecule has the simplest formula. In that case the formula is said to be empirical.
Some prefer to use the term formula weight in place of molecular weight when
the actual molecular weight has not been determined. The formula weight, like
the molecular weight of a compound, is the sum of the atomic weights of the
elements represented in the known or assumed formula of the compound.
Examples.- — (1) 10 grams of pm-e tin when oxidized in air gave 12-7 grams of oxide.
What is the formula of tin oxide ? The atomic weight of tin is 119, and of oxygen 16.
Hence, the ratio : Tin : oxygen = 10 ^119 : 2-7^16=0-084 : 0-17 = 1 : 2. The formula is
therefore written SnOg, although there is nothing to show why it is not Sn204 ; SngOg ;
. . . Snn02n-
(2) A sample of crystallized sodium carbonate furnished on analysis 37-2 per cent,
of NaXOg, and 62-8 per cent, of HgO. What is the formula of the compound ? The ratio
NaaCOg: H2O = 37-2^106 : 62-8-M8=0-35: 3-49 = 1 : 10. Hence, the formula is taken
as NagCOj.lOHgO, although there is nothing to show why it is not some multiple of this,
say, iwNaaCOg.lOwHaO.
(3) A. Jones (1892) analysed a sample of electric calamine, and found : Silica, SiOj,
25-33; zinc oxide, ZnO, 67-15; and water, HjO, 7*47 per cent. Show that this
corresponds verj'^ nearly w?th the formula Zn2Si04.H20.
(4) W. F. Hillebrand and W. H. Melville (1892) analysed some crystals obtained by the
action of sulphuric acid on uranium oxide, and found : UO2, 53-99 ; SO 3, 36*95 ; HgO,
14-13 per cent. Show that the molecular ratio of these three constituents is 1 : 2 : 3-94,
and that this corresponds with the formula 1102(803)2, 4H2O or 11(804)2. 4H2O.
(5) G. Femekes (1902) analysed a salt obtained by treating a solution of mercuric
chloride with potassium ferrocyanide, and found : Potassium, 15-82 per cent. ; mercury,
40-63 ; iron, 11-45 ; and cyanogen, CgNg, 31-78. Show that the simplest empirical formula
for the compound is K2HgFe(CN)8.
§ 5. The Relative Weights of the Atoms
Atoms are so inconceivably little that their aggregates are alone the ostensible subject
of experiments.- — S. Brown.
It has abeady been stated that the conceptions molecular weight and atomic
weight are quite independent of our theories about the nature of atoms and mole-
cules ; nor are the conceptions much affected by the actual weights of the atoms
and molecules because the terms under consideration are definite expressions of
Avogadro's hypothesis coupled with observed facts. It might therefore have been
misleading to head this paragraph : Weighing the Atoim. There are reasons for
supposing that the molecular weight of some compounds in the liquid or sohd con-
dition is a multiple of the molecular weight of the same substance in the gaseous
condition. The molecule of steam approximately corresponds with the formula
H2O ; but in liquid water there are reasons for supposing the molecule is either
(H20)3 or (H20)4, that is, the formula for hquid water is not HgO, for it contains
molecules corresponding with H4O2, HeOa, or H8O4. . _ ^ , , .1 . •
Refer back to the difficulty in fixing the atomic weight'of carbon from the ratio
of the weights of carbon and oxygen in the two oxides of carbon which we encountered
in applying J. Dalton's atomic theory. Suppose that we do not know the atomic
180
INOKGANIC AND THEORETICAL CHEMISTRY
weight of carbon, but that we do know the composition of a number of volatile
carbon compounds as well as their relative densities or molecular weights, Table II.
Table II. — ^Molecular Weights oe Some Carbon Compounds.
Volatile compound of carbon.
Composition by weight.
Molecular
weight.
1 Amount of carbon
1 per molecule.
Carbon monoxide
Carbon 12 ; oxygen 16
28
12
Carbon dioxide .
Carbon 12 ; oxygen 32
44
i 12
Methane ....
Carbon 12 ; hydrogen 4
16
1 12
Ethylene ....
Carbon 24 ; hydrogen 4
28
.12X2 = 24
Propylene ....
Carbon 36 ; hydrogen 6
42
12X3 = 36
Carbon disulphide
Carbon 12 ; sulphur 64
76
1 12
The smallest weight of carbon in a molecule of any of its known compounds is
12, and consequently this number is assumed to be the atomic weight of carbon.
The atomic weights of a great number of the elements have been determined in a
similar manner.
The determination of atomic weights. — According to J. Sebelin's Beitrluje
zur Geschichte der Atotngewichte (Braunschweig, 1884), when J. J. Berzelius was
asked how he was able to get such excellent analyses, analyses which have been the
admiration of generations of chemists, he answered :
Try to find that method of analysis in which the accuracy of the result is least dependent
upon the skill of the operating chemist ; and when this method has been selected, consider
what unavoidable conditions are present which are likely to affect the result with errors ;
and then ascertain whether the errors will increase or diminish the result. Then make
another determination in which the opposite effects can alone be produced. If the two
results are the same, the determination was correct.
The actual method used in finding the atomic weight of an element really
requires :
(1) An-exact analysis of a series of compounds containing the given element ; and
consequently the compounds investigated must be such as lend themselves
to exact analysis, and which can be prepared in a highly purified condition.
(2) It is an advantage if the compound be volatile without decomposition, so
that its vapour density can be determined. There are several other
methods of computing the molecular and hence also the atomic weights of
the different elements ; and in several, the compound need not be
volatile. Fortunately, atoms and molecules possess other qualities besides
mass, which are dependent upon their atomic weights and which can be
readily measured. Some of these will be described later.
(3) The smallest proportion of the element under investigation contained in all
the compounds whose molecular weights are known is finally selected as
the atomic weight of the given element.
J. A. Wanklyn (1894) ^ once claimed to have discovered a series of hydrocarbons,
one member of which contained carbon 102 parts by weight, and hydrogen 17 parts,
and had a vapour density of nearly 116 (hydrogen 2). Assuming the atomic weight
of carbon is 12, and of hydrogen 1, these numbers give formula C8.5H17. If this
statement had been corroborated, and we were quite sure that Wanklyn's hydro-
carbons were not mixtures, it would be necessary to make the atomic weight of
carbon = 6, and write the formula of the compound in question C17H17, and this
in spite of the fact that thousands of compounds of carbon are known, and all agree
with the number 12 for the atomic weight of carbon. The formula of carbon
monoxide— CO— would then be written CgO, etc.— but J. A. Wanklyn's claim has
never been established.
These remarks emphasize the importance of examining as large a number of
volatile compounds as possible when fixing the atomic weight of an element. The
COMBINATION BY VOLUME 181
importance of this principle was recognized as early as 1859, for F. A. Kekule then
wrote :
It is of exceptional importance for chemists to determine the relative masses of particles
which are not subdivided in chemical reactions. In order to determine the atomic or
molecular weights of the elements and their compounds with some degree of probability
It IS necessary to investigate a very great number of compounds and a very great number
of chemical reactions.
If only a small number of compounds be examined, there is always a possibility,
and perhaps a probability, that the actual minimum weight does not occur amongst
the set of compounds taken. It follows, therefore, that the atomic weight of an
element is the least amount of that element— relative to the standard
oxygen, 16— which is present in any molecule of all its known volatile
compounds. The value so obtained is the maximum possible value ; the real value
may afterwards prove to be a submultiple of this. The atomic weight must be
equal to a whole multiple or submultiple of its combining weight. Owing to the
fact that the molecular weights of so many volatile compounds of carbon are known,
it is not very probable that the atomic weight of carbon is less than. 12.
References.
1 J. A. Wanklyn, Chem. News., 70. 89, 147, 1894 ; F. A. Kekule, Liehig's Ann., 106. 129, 1858.
§ 6. Methods for Measuring the Vapour Densities of Gases, and of Volatile
Liquids and Solids
The history of science shows that even during that phase of her progress in which she
devotes herself to improving the accuracy of the numerical measurement of quantities
with which she has long been familiar, she is preparing the materials for the subjugation
of new regions, which would have remained unknown if she had been contented with the
rough methods of her early pioneer.— J. C. Maxwell.
When determinations of molecular weights are made to decide between quantities
widely different, minor corrections, necessary for exact values, are not required.
For instance, if chemical analysis showed that the molecular weight of a compound
is some multiple of 20, then a molecular weight of 83, by vapour density methods,
indicates that 4x20=80 is the molecular weight of the substance. With ordinary
vapour density determinations, therefore, the weight of 22 '4 litres of the gas or
vapour at 0° and 760 mm. is to be computed from measurements with hydrogen = 2
or oxygen = 32 as standards of reference. No new principle is involved. If an
intermediate value between two possible values for the molecular weight of a sub-
stance is consistently obtained, there is a disturbance — ^possibly association or
dissociation — which must be investigated more closely.
These remarks do not apply when the molecular weights of gases are estimated
from their densities in order to serve as a control for the atomic weights. The
densities are then determined with as great an accuracy as possible. In the fourth
century B.C., Aristotle made an unsuccessful attempt to determine the weight of
air contained in a bladder ; and, about 1632, G. Galilei established the fact that air
has weight ; R. Descartes (1638) said that G. GaUlei's primitive method of weighmg
air rCest pas mauvaise. Robert Boyle, in his Hydrostatics (Oxford, 1666), gives the
specific gravity of air 0-00125— with water unity as standard. The air of tartar, which
consists of a mixture of carburetted hydrogen and carbonic acid gases, was weighed
in a bladder by S. Hales,i and he compared the weight so obtained with that of
the same bladder filled with air ; F. Hauksbee determined the specific gravity of the
mixture of carbon dioxide and nitrogen obtained by passing air over red-hot iron.
The specific gravity of these mixed gases was so near that of air that the ex-
perimenters, by their methods, did not establish a difference. J. Mayow supposed
182
INORGANIC AND THEORETICAL CHEMISTRY
but did not prove that the nitro-igneous constituent of air was heavier than the resi-
dual air from which it was separated. H. Cavendish, however, in his Exferiinents
on Factitious Airs, in 1766, first estabhshed the difference in the specific gravities
of air, carbon dioxide, and hydrogen, and this has been cited as the first conclusive
proof of a pluraUty of elastic fluids. J. Priestley tried to weigh the different
kinds of air in glass flasks by the displacement of water in a pneumatic trough,
but the drops of water which adhered to the inside of the flask introduced too
many errors. J. Priestley then used a bladder, and added that although the
determination " cannot be done with precision in a bladder, as used by Mr.
Cavendish, because the degree of distension cannot be measured with much accuracy,
yet the circumstance is more than counterbalanced by being able to change the
air, with compressing the bladder, without wetting it." J. Priestley found the
bladder filled with —
Phlogisticated air weighed
Nitrous air
Common air .
Dephlogisticated air
dwts.
7
7
7
7
grams.
15
16
17
19
The early chemists apparently thought the determination of the density of a
gas to be so simple an operation that details would be redundant ; and they con-
sidered it was necessary merely to weigh a bladder or a flask first evacuated, and then
filled with the required gas. Towards the end of 1779, F. Fontana 2 devised a much
better method of measuring the specific gravities of different gases.
The stoppered globe A, Fig. 1, of known capacity is unscrewed from the gas stoppered
receiver B, exhausted, weighed, and again screwed to the receiver ; meanwhile, the
receiver is filled over a mercury pneumatic trough with the gas
under investigation. The stopcocks are opened, the cylinder
B depressed in the mercury until the surface of the mercury
is the same inside and outside the cylinder. The stopcocks are
then closed, the difference of the two weighings is taken to
represent the weight of the gas in the globe. This result, divided
by the capacity of the vessel expressed in cubic inches, gives the
weight of a cubic inch of the gas in question.
J. B. Biot and F. J. Arago (1806) determined the density of
undried gases by means of a globe between 5 and 6 litres
capacity. The results were corrected for the air displaced
by the globe ; the residual air in the evacuated flask ; the
cubical expansion of glass ; and the hygroscopic moisture
in the gas. The results were reduced to normal tempera-
ture and pressure, to sea-level, and to a latitude of 45°.
For a long time J. B. Biot and F. J. Arago's measurements
were considered to be a model for the work of others.
J. J. Berzelius and P. L. Dulong (1820) and J. B. A. Dumas and J. B. J. D.
Boussingault (1841) followed J. B. Biot and F. J. Arago's method, but they dried
the gases.
A new era was inaugurated by H. V. Regnault in 1847. He introduced many
vital improvements in J. B. Biot and F. J. Arago's procedure — chiefly in the use of
a counterpoise balloon, and in the filling and exhausting of the globes while they
were surrounded by a bath of melting ice. Modern work follows closely on the
lines marked out by H. V. Regnault. Every known precaution which will conduce
to the accuracy of the result is taken : (1) Attention is paid to the extreme purifica-
tion of the gases to be measured ; (2) the difference in the buoyancy in air of the
weight and of the substance to be weighed is eliminated by reducing the weighings
to the vacuum standard ; (3) Lord Rayleigh's correction (1893) for the difference
in the volume of the evacuated and filled balloon holding the gas is applied ; (4) the
expansion of the glass with variations of temperature is considered ; (5) corrections
are made for the deviations of the gas under investigation from Boyle's and Charles'
Fig. 1. — Fontana's Ap-
paratus for Measuring
the Density of Gases.
COMBINATION BY VOLUME 183
laws ; (7) an allowance is made for a slight condensation of gas on the inner walls
of the measuring vessel ; etc. In measuring the relative density of a substance
which is gaseous at ordinary temperatures, three methods are available :
A. Weighing a known volume of the gas. The balloon method was worked out by
H. V. Kegnault (about 1847), and it has been much used in more recent work, where
the general tendency has been to reduce the size of the balloons. H. V. Regnault
worked with balloons about 10 litres capacity ; E. W. Morley (1896) 3 with balloons
8-21 litres capacity; A. Leduc (1897), 23 litres ; Lord Rayleigh (1888-95),
1-8 litres ; P. A. Guye and C. Davilla (1905) used globes of capacity 0-38 to about
0-82 litre for nitric oxide ; E. P. Perman and J. H. Davis (1906), 0'5Htre ; and R. W.
Gray (1905), 0-267 litre. The determinations made with small balloons are quite
as concordant among themselves as those made with balloons of larger volume.
In this method the glass globe of volume v is counterpoised on the balance by a second
tare balloon of approximately the same volume so as to eliminate corrections necessary
for the buoyancy of the air. By repeated exhaustions and re-fillings, the balloon is filled
with the gas under investigation. The temperature and pressure are respectively 6 and p.
Let w denote the difference between the weights of the full and empty balloon. The volume
Vq of the gas at 0° and 760 mm. pressure is first calculated in the ordinary manner^:
( p \( 273 \ 0-3592t;«
From Avogadro's hypothesis the molecular weight of a gas represents the weight
of 22*3 litres of a gas if hydrogen = 2 be taken as the standard. Consequently,
if w grams of a gas occupy v^ c.c. at 0° and 760 mm. pressure, 22,300 c.c. will weigh
22,300«^-^^i grms., and this represents the molecular weight, or the uncorrected
relative density of the gas, hydrogen = 2. For a high degree of accuracy, it is of
course necessary to include correction terms as indicated above.
Examples. — (1) 585 c.c. of carbon dioxide measured at 18° and 756 mm. pressure
weighed 1 '076 gram. What is the molecular weight of the gas ? 685 c.c. of gas become,
at 0° and 760 mm., 546-1 c.c. Hence, the molecular weight is 22,300 X l-076-f546-l=43'9.
(2) H. V. Regnault (1845) filled a 10-litre globe with air at a pressure of 761-19 mm.
at the temperature of melting ice. In addition to the tare balloon 1487 grms. were required
to balance the globe. The globe was then exhausted to a pressure 88-43 mm., and
14-141 grms. were now required to restore equilibrium. The globe was then filled with
dry oxygen at 0° and 750*22 mm. pressure, 0-172 grm. was needed in addition to the tare
to balance the globe. The globe was then exhausted to 4-59 nun. pressure and weighed,
again 14-033 grms. were required. The globe lost 12-654 grms. of air at 761-19 — 8-43
= 752-76 mm. pressure and 0°. This corresponds with 12-776 grms. of air at 760 mm.
Similarly with the oxygen : 14-1281 grms. at 760 mm. and 0°. The weights refer to equal
volumes, and therefore the relative density of the oxygen (air unity) is 14-1281^12-776
= 1-10563.
B. Measuring the volume of a known weight of the gas. — The volume occupied by a
known weight of gas is measured in a suitable voluminometer, and the gas required to
fill the balloon is weighed in another vessel either («) by finding the loss of weight
due to the escape of gas from the generating apparatus, or {h) by absorbing the gas
in suitable apparatus. In the former case, given the temperature and pressure of
the confined gas, the capacity of the balloon, and the loss of weight in the vessel
from which the balloon was filled, the density follows directly. This method was
used by E. W. Morley (1896) for hydrogen, and by A. Jaquerod and A. Pintza (1904)
for sulphur dioxide. In a variation of this procedure, the measuring vessel is filled
with the purified gas and its temperature maintained at 0°, while the pressure
(approximately 760 mm.) is determined. The gas is then absorbed in a suitable
apparatus previously evacuated and connected with the voluminometer by a tightly
fitting joint. The weight of the absorbed gas completes the required data. This
method was used by P. A. Guye and A. Pintza (1904-5) for nitrous oxide, carbon
dioxide, and ammonia ; by E. P. Perman and J. H. Davis (1906) for ammonia ;
and by R. W. Gray and E. P. Burt (1909) for hydrogen chloride.
0. Measuring the buoyancy of the gas in atinosfheres at a known pressure.—
A good analytical balance will indicate 00001 grm. when carrying a load of 100 grms. ;
184 INORGANIC AND THEORETICAL CHEMISTRY
the balances used for assaying will indicate O'OOOOl grm. with a load of 10 grms.
The sensibility of instruments for detecting variations of mass has been so sharpened
that the latest form of micro-balance will carry a maximum load of 5x10"^ grms.,
and is sensitive to 3*3x 10"^ grm. The weighing of minute masses is called micro-
weighing.* Probably the first micro-balance was made by E. Warburg and
T. Ihmori in 1886. The beam of this balance was made of thin quartz rods to which
were cemented razor knife-edges ; the deflections of the beam were read from a
mirror and scale without taring the weights. The sensitiveness of this balance
was about the same as the assay balances. In 1906, W. Nernst and E. H. Riesenfeld
devised a torsion micro-balance in which a quartz fibre was cemented to the prongs
of a vertical brass fork ; and a thin glass rod likewise fixed horizontally to the quartz
fibre. One end of the rod is intended to serve as a pointer on a silvered scale, and the
other carries a tiny pan. The load causes a slight torsion of the quartz fibre. The
maximum load is 2 mgrm., and the lower limit of sensibility is 5xlO~^ grms. In
another type, the principle of Archimedes is applied, and a gas manometer takes
the place of a set of weights. This apparatus was improved by B. D. Steele and
K. Grant (1909), and W. Ramsay and R. W. Gray (1911), so that a weight 00000001
grm. can be accurately determined. In some of the improved forms a still greater
sensibility has been attained. In this way, the density of less than 0"75 cubic
millimetres of the emanation from radium has been determined, and the corre-
sponding molecular weight computed from the result. This is a triumph of manipu-
lative skill.
If air at the same temperature and pressvire as the ambient atmosphere be confined in
a quartz bulb, it will apparently weigh nothing, but if the outside air be reduced in density,
the air inside the quartz bulb will appear to have a positive weight. Given the pressure
of the ambient air, the weight of the confined air can be readily calculated as indicated
in the subjoined example. The beam of the instrument is a framework of thin rods of
silica arranged to swing on a central knife-edge resting on a central support. A scale-pan
or bucket and a sealed air bulb of known volume are suspended from the framework by
quartz threads, and coiinterpoised by a weight. All is enclosed in an air-tight metal case
fitted with a mercury gauge. A mirror reflects a beam of light from the window to a
scale a few feet away. The tube containing the gas \mder investigation is placed in
the bucket, and the pressure noted at which the beam is in equilibrium. This is indicated
by the spot of reflected light. The bulb containing the gas is broken, and all the glass
splinters are placed in the bucket. The gas is removed by evacuating the metal case a
few times. The pressure of the air again required to bring the spot of light to equilibrium
is noted. Suppose, by way of example, that the pressure of the air required to bring a tube
of xenon gas in the equilibrium position be 70 mm. ; and similarly the empty tube, 52-9 mm.
The difference, 17*11 mm., corresponds with a weight 608 millionths of a milligram. A
correction is required for differences in the weight of the glass vessel at pressures of 70 mm.
and 52*9 mm. It is 15 millionths mgrm. Again, the effect of the reduced pressure on the
buoyancy of the glass bulb and the silver counterj^oise is different. By substituting a
counterpoise of silica the difference was found to be 91 millionths of a mgrm. Hence the
weight of the gas in question is 608 — 91 + 15 = 532 millionths of a milligram.
The micro-balance has been used by F. W. Aston (1914) to compare the densities
of two gases. The gas to be investigated, density Z), is admitted to the balance
case, and the pressure p determined at which the balance beam is in a given position.
The corresponding pressure pi for a gas of known density Dj, say, oxygen is then
determined. The densities of the two gases D and Di are inversely proportional
to the pressure p and pi, or the density D of the required gas is PiDi/p.
The vapour density of solids and liquids which can be vaporized without
decomposition can be obtained by the following methods :
A. Weighing a known volume of the vapour.— In J. B. A. Dumas' process (1826) ^
the substance is vaporized in a weighed glass bulb at atmospheric pressure. The
bulb is then sealed up, and the weight of the vapour determined. The capacity of
the bulb is then measured. From the resulting data, the vapour density of the
gas follows directly.
Example. — ^Tbe following data were obtained by H. E. Roscoe (1878) for vanadium tetra-
chloride : Weight of globe filled with air (9°, 760 mm.), 24-4722 grams ; weight of sealed
COMBINATION BY VOLUME 185
globe (9°, 7G0 mm.), 25'0102 grams ; temperature of bath when sealing the globe, 215" ;
barometer when sealing the globe, 762 mm. ; and the weight of bulb full of water, 194 grams.
The globe held less, 24-4722 = 169*5 grams of water at 9°. This represents very nearly
169-5 c.c. of water, or the capacity of the globe is 1695 c.c. The apparent weight of the
substance at 9° is 25-0102 —24-4722 =0-538 gram. The empty globe was buoyed up, during
weighing, by its own bulk of air at 9° and 762 mm., and since 1 c.c. of air weighs 0-001293
grams, 169*5 c.c. of air at 9° and 762 mm. weigh at (0-001293 X 169-5 x 273 X 762)-i-(760
X 282) =0-213 gram. This added to 0-538 gram, gives 0-751 gram, the weight of the vapour
in the globe at the time of sealing. The 0-751 gram of vapour occupied 169-5 c.c. at 215**
and 762 mm. pressure, or 95*10 c.c. at 0° and' 760 mm. pressure. Hence, 22,300 c.c. of
vapour at normal temperature and pressure weigh 176*1 grams. This number also repre-
sents the molecular weight of vanadium chloride.
Vessels made of porcelain have enabled H. St. C. Deville and L. Troost (1858),
H. E. Koscoe (1878), and others to determine vapour densities by this process at
temperatures far exceeding those at which even hard glass softens. The objection
to Dumas' process is the amount of material which has to be vaporized in order
to drive out the air from the bulb. This waste is avoided in the two succeeding
methods — Hofmann's and Meyer's processes. By using porcelain or platinum
vessels, Dumas' process has been employed for bodies volatilizing at high
temperatures.
B. Measuring the volume of a known weight of the vapour. — J. L. Gay Lussac
(1811) showed that the vapour density of a substance can be determined by
measuring the volume of a known weight of the vapour in such a way that the
volatile substance is confined in a small vessel of known capacity by means of
mercury or any other substance which boils at a high enough temperature— 6.(7.
Wood's fusible alloy. When the vessel has been heated the bath is removed. After
cooling, the volume of the vapour at the highest temperature of the bath can be cal-
culated from the weight of, sa.j, mercury remaining in the vessel. J. L. Gay Lussac's
process was perhaps the oldest method used for measuring vapour densities. He
placed a known weight of the substance under investigation in a graduated
glass tube, about 40 cm. long, and filled with mercury. The tube dipped in
mercury and was surrounded by a hot jacket so as to vaporize the substance.
The temperature and volume of the confined vapour were measured. In A. W.
Hofmann's process (1868) the measuring tube is over 760 mm. in length.
Example. — -The following data were obtained for carbon tetrachloride, CCI4 : Weight
of liquid in bulb, 0-3380 grm. ; the volume of vapour, 109-8 c.c. ; the temperature of vapour,
99-5° ; the barometer, 7469 mm. ; and the height of mercury in tube, 283*4 mm. The
pressure of the vapour is the barometric height less the weight of the column of mercury
in the Hofmann's tube, that is, 746-9 — 283-4^463-5 mm. Hence, 0-3380 gram of vapour
at 99-5° and 463-5 mm. pressure occupy 109-8 c.c, and 49-09 c.c. at 0° and 760 mm. Hence,
22,300 c.c. of the vapour at normal temperature and pressure weigh 153-6 grams, and this
number represents the molecular weight of carbon tetrachloride.
A. W. Hofmann's process is useful when only a small amount of the substance
is available for a determination ; and for a substance which decomposes when
heated at a temperature in the vicinity of its boiling point at ordinary atmospheric
pressures. In V. and C. Meyer's process (1877) « the apparatus is simphfied by
measuring the volume of air displaced by a given weight of the substance
vaporized in a suitable vessel.
If the substance be vaporized too slowly, vapour will be carried forward with
the expelled air, and be condensed, thu» reducing the volume of air (or gas) measured
in the gas burette. It is considered that the vaporization vessel should be at least
30° above the boihng point of the substance in order to secure rapid vaporization.
If the gas be collected over water instead of over mercury and is filled with ordinary
moist air instead of with dry air, a correction for the pressure of aqueous vapour may
be applied. If air contains a per cent, of moisture ; and / denotes the pressure of
aqueous vapour at the room temperature ; and p, the barometric pressure, the
actual pressure of the confined gas is taken to be p-(l— r^o«)/ This refinement is
usually ignored.
186 INOKGANIC AND THEORETICAL CHEMISTRY
Example." — -The vapoiir density of water was determined, and the following data were
obtained. Xylene, boiling at about 138°, was used in the hot jacket E. It was found that
the weight of the water in the stoppered tube was 0*0102 grm. ; the temperature of the
gas in the biirette, 16*5° ; the barometer, 7038 mm. ; and the volume of gas, 16'6 c.c.
The 16-6 c.c. of vapour at 16-5° and 7038 mm. becomes 14*496 c.c. at 0° and 760 mm.
This is the volmne of 0*0102 gram of vapour. Hence, 22,300 c.c. of the vapour will weigh
15*7 grams. This number represents the molecular weight of water vapour.
V. Meyer's apparatus has been modified in various directions without altering
the fundamental principle. J. S. Lumsden (1903) proposed a modification in
which the increase of pressure was measured while the volume of the apparatus
was kept constant. Glass vessels are suited for this determination only at com-
paratively low temperatures ; vessels made of hard porcelain have been used by
J. Mensching and V. Meyer (1886) for temperatures up to about 1500° ; platinum,
platinum-iridium alloy, vitreous siUca by J. Dewar and A. Scott (1879), L. F.
Nilson and 0. Pettersson (1886), and by J. Mensching and V. Meyer (1886) ; and
vessels of iridium fined inside and outside with a magnesian cement, and heated
in an electric furnace, enabled W. Nernst (1903) and H. von Wartenberg (1908) to
measure vapour densities at temperatures as high as 1800° and even 2000°.
References.
1 S. Hales, Vegetable Staticks, London, 185, 1727 ; F. Hauksbee, Phil. Trans., 25. 2409, 1707 ;
J. Mayow, De parte aerea igneaque spiritus nitro, Oxford, 1669 ; H. Cavendish, Phil. Trans., 55.
141, 1766 ; J. Priestley, Experiments and Observations on the Different Kinds oj Air, London,
2. 93, 1790.
2 L. Cavallo, A Treatise on the Nature and Properties of Air, London, 422, 1781 ; J. B. Biot
and F. J. Arago, Mem. Acad., 301, 1806; J. J. Berzelius and P. L. Dulong, Ann. Chim. Phys.,
(2), 15. 386, 1820 ; J. B. A. Dumas and J. B. J. D. Boussingault, ib., (3), 3. 267, 1841 ; Liebig's
Ann., 40. 230, 1841 ; H. V. Regnault, Mem. Acad., 21. 25, 1841 ; Ann. Chim. Phys., (3), 5. 52,
1842 ; P. A. Guye and C. DaviUa, Mem. Soc. Phys. Geneve, 35. 615, 1908 ; E. P. Perman and
J. H. Davis, Journ. Chem. Soc., 90. 743, 1906 ; R. W. Gray, Journ. Chem. Soc, 87. 1601, 1905.
3 Lord Rayleigh, Chem. News., 58. 52, 1888 ; Phil. Trans., 196. A, 205, 1901 ; 198. A, 417, 1902 ;
Zeit. phys. Chem., 37. 713, 1901 ; 41. 71, 1902 ; 52. 705, 1905 ; Proc. Roy. Soc., 73. 153, 1904 ;
E. W. Morley, Amer. Journ. Science, (3), 41. 220, 276, 1891 ; A. Leduc, Campt. Rend., 123. 743,
1896 ; 125. 297, 571, 646, 1897 : 126. 413, 1898 ; Ann. Chim. Phys., (7), 15. 5, 1898 ; Journ. Phys.,
(3), 7. 5, 189, 1 898 ; P. A. Guye and A. Pintza, Mem. Soc. Phys. Geneve, 35. 594, 1908 ; A. Jaquerod
and A. Pintza, ib., 35. 589, 1908 ; E. P. Perman and J. H. Davis, Journ. Chem. Soc., 90. 743,
1906 ; R. W. Grav and F. P. Burt, ib., 96. 1633, 1909 ; W. Nernst, Zeit. Electrochem., 10. 629, 1904.
* E. Warburg and T. Ihmori, Wied. Ann., 27. 481, 1886 ; 31. 100, 1887 ; H. Petterson, Ein
neue Microunge und ihre Anwendung, Stockholm, 1914; F. W. Aston, Proc. Roy. Soc., 89. A,
439, 1914 ; W. Ramsay and R. W. Gray, ib., 84. A, 53, 1911 ; B. D. Steele and K. Grant, ib., 82.
A, 580, 1909; J. Kramer, Chem. Ztg., 41. 773, 1917; W. Nernst and H. Riesenfeld, Ber., 36,
2086, 1903 ; W. Nernst, Golt Nachr., 2, 1902.
* J. B.A.Bum&s, Ann. Chim. Phys., {2), 33. 341, 1826; C. W. Balke and E. F. Smith, Jowrn.
Chetn. Soc, 94. 1043, 1908 ; H. St. C. Deville and L. Troost, C&tnpt. Rend., 46. 239, 1858 ; Liebig'e,
Ann., 113. 42, 1860; H. E. Roscoe, Proc Roy. Soc, 27. 246, 1878; Ber., 11. 1196, 1878;
A. W. Hofmann, Ber., 1. 198, 1867; J. L. Gay Lussac, Ann. Chim. Phys., (1), 80. 118, 1811 ;
E. Ernyei, Zeit. anorg. Chem.., 25. 313, 1900.
« V. Meyer and C. Meyer, Ber., 12. 2204, 1879 ; V. Meyer, ib., 9. 1216, 1876 ; 11. 2068,
1878 ; A. Combes, Journ. Chem. Soc, 56. 571, 1889 ; J. S. Lumsden, ib., 83. 342, 1903 ; L. M.
Dennis and H. Isham, Journ. Amer. Chem. Soc, 29. 18, 1907 ; J. Mensching and V. Meyer, Zeit.
phys. Chem., 1. 145, 1887 ; J. Dewar and A. Scott, Proc Roy. Soc. Edin., 14. 410, 1887 ; L. F.
Nilson and 0. Pettersson, Journ. prakt. Chem., (2), 33. 1, 1886; Zeit. anal. Chem., 27. 197, 1888;
Ber., 17. 987, 1884; W. Nernst, Zeit. Electroch., 10. 629, 1904; H. von Wartenberg, Ber., 39.
381, 1908.
§ 7. The Struggle o! Avogadro's llypothesis for Recognition
The first attempt at generalization seldom succeeds ; speculation anticipates experience,
for the results of observation accumulate but slowly.- — J. J. Berzelius (1830).
A. Avogadro's hypothesis had a long struggle for recognition in spite of the fact
that his memoir was followed three years later by A. M. Ampere's note addressed
to M. le Comte Berthollet and entitled, Sur la determination des proportions dans
lesquelle.^ les corps se combinent, d'apres le nomhre et la disposition respective des
molecules dont leurs particles integrantes sont composees (1814), advocating similar
COMBINATION BY VOLUME 187
views. Half a century elapsed before the hypothesis was generally accepted.
Among the many reasons for its failure was the fact that comparatively few sub-
stances which could be vaporized were then known, and hence the molecular weights
of but few compounds had been determined with precision. At that time, an
accurate knowledge of the weights of the elements was considered to be the most
pressing subject of investigation. J. Dalton's atomic theory had just been born,
and accurate data were also needed before that theory could be utiUzed. Referring
to J. Dalton's theory J. J. Berzelius said :
I recognized that if the newly arisen light was to be spread, it would be necessary to
ascertain with the utmost accuracy the atomic weights of all elementary substances. . . .
Without such work, no light would follow the dawn.
Something more than the mere accumulation of experimental data was necessary
to find a method for determining the number of atoms in a molecule, in order that
the atomic weight of the constituent elements could be obtained. Dalton
pointed out that in fixing the atomic weight of oxygen with respect to hydrogen
unity, he assumed that water is a binary compound of one atom of hydrogen and one
of oxygen. If water be really a ternary compound containing two atoms of hydrogen
and one of oxygen, it will be necessary to double the atomic weight of oxygen
determined on the former assumption ; and if water contains two atoms of oxygen
and one of hydrogen, the atomic weight of the oxygen would have to be halved.
Similarly with other compounds. The uncertainties in the application of J. Dalton's
atomic theory are due to the arbitrary nature of the assumption of the number of atoms
in a molecule.^
W. H. WoUaston (1814).— In 1810, T. Thomson gave a list of the weights
of various acids and bases which neutralized one another, and showed that these
numbers were independent of the hypothesis of Dalton. W. H. WoUaston, in his
paper A synoptic scale of chemical equivalents (1814), proposed to substitute the
term equivalent in place of Dalton's atom. He claimed that his numbers were
not warped by the uncertainties of the atomic theory, and that for practical pur-
poses it is not necessary to know hypothetical atomic weights when equivalent
weights are known. WoUaston thus proposed to do for the elements what J. B.
Richter (1791-1802) had done for the acids and bases, and he accordingly used the
term equivalent proposed by H. Cavendish in 1788. Starting with oxygen 10
as the unit of reference, he found the equivalent of hydrogen to be TS — meaning
that 1'3 parts of hydrogen unite with 10 parts of oxygen to form water. In this
sense, equivalent weights are identical with combining weights. WoUaston,
however, got into difficulties in deahng with substances like carbon with two com-
bining weights, for he was obliged to assume that in carbon dioxide two equivalents
of oxygen were united with one of carbon so that equivalents and combining weights
were no longer the same. Since 7" 5 parts of carbon unite with 20 parts of oxy^gen
by weight to form carbon dioxide, W. H. WoUaston said the formula of the compound
is CO2 ; and because 75 parts of carbon unite with 10 parts of oxygen by weight
to form carbon monoxide the formula is CO. W. H. WoUaston might just as
arbitrarily have said 375 of carbon unite with 10 parts by weight of oxygen to form
carbon dioxide, and the formula is accordingly CO ; and in carbon monoxide 7*5
parts of carbon unite with 10 parts of oxygen by weight, and therefore the formula
is C2O. Hence, W. H. WoUaston's equivalents leave the difficulty with J. Dalton s
atoms just as it was. Ignoring this uncertainty, the former computed the
equivalents of 12 elements, and 45 compounds from various analyses. A. Ladenburg
(1869) 2 is severe on W. H. WoUaston, for he says that WoUaston's views involved
a retrograde step, for W. H. WoUaston believed that he was deahng with un-
ambiguous conceptions free from aU hypothesis ; and in introducing the term
equivalent, he confused the conceptions of the equivalent and the atom, and
rendered a vigorous struggle necessary before the two concepts could be ciaritied.
J. J. Berzelius (1810-26).— J. J. BerzeUus seems to have regarded the investigation
of the laws of combining proportions to be one of the most important objects
188
INORGANIC AND THEORETICAL CHEMISTRY
of his life's work, and in a memorable work, Essai sur la theorie des frofortions
chimiques et sur Vinjiuence chimique de relectricite, published in Sweden in 1814, and
at Paris in 1819, he developed his conceptions of the atomic theory, and collected
together the results of his arduous work on the combining proportions of the
elements which he had published between 1810 and 1812. In his Lehrbuch der
Cheynie (Dresden, 1825), J. J. Berzelius remarked that it did not matter much
whether the particles which combine chemically be called atoms, molecules, chemical
equivalents, or any other terms, but he preferred to use the term atom. In spite
of Dalton's demonstration that Gay Lussac's law does not mean that equal
volumes of elemental gases contain the same number of atoms, J. J. Berzelius used
this erroneous Volumtheorie as a guide in determining the numerical values of the
atomic weights of the elements which were gaseous at ordinary temperatures. He
said, one volume of an elementary substance in the gaseous state corresponds with
one atom, and he called the smallest particles Volumatome. This erroneous theory
gave him satisfactory results in deducing the composition of water, and of the two
carbon oxides ; and consequently also of the atomic weights of oxygen, hydrogen,
and carbon. J. J. Berzelius (1818) was not so fortunate with some of the metallic
oxides, particularly the sesqui-oxides. It is illogical, said he, to express the
composition of a series of oxides of an element A by the formulae A2O2, A2O4, A2O6 . . .,
and declared that in a series of compounds of the two elements one compound must
always be represented as containing a single atom of one of the elements, and
accordingly he wrote the formulae AG, AO2, AO3 ... If the simplest oxide were
A2O2 Berzelius said that the atoms of the elements A would be divisible mechanically.
J. Dalton saw the fallacy in this argument for fixing the number of atoms in a
molecule of the solid oxides, but, led on by the erroneous argument, Berzelius
(1818) wrote the formulae for ferrous oxide Fe02 instead of FeO ; and ferric oxide,
FeOa instead of Fe203, and thus obtained numbers for the atomic weights of iron,
chromium, etc., double the values of those now accepted. The chemical similarities
of iron with chromium led him to symbolize chromic oxide CrOg by analogy with
FeOa ; similarly he formulated the metal oxides Zn02, Mn02 • • • by analogy with
FeOg. In 1826, however, he wrote —
Assuming that chromic oxide contains three atoms of oxygen ; and that chromic
anhydride contains six atoms, in forming neutral salts, chromic anhydride neutralizes an
amount of base containing one-third as much oxygen as it itself contains. By analogy
with sulphuric anhydride, and other anhydrides with three atoms of oxygen, it is most
probable that chromic anhydride contains three atoms of oxygen to one of chromium,
consequently chromic oxide will contain three atoms of oxygen to two of chromium, and
the formulae for chromic oxide will be CrjOg, and for chromic anhydride, CrOg. The
isomorphic law will then make ferric oxide, FcgOg, instead of FeOg ; and aluminium oxide,
AI2O3, instead of AIO3.
These considerations led J. J. Berzelius in 1826 to halve the atomic weights of the
metals published in his 1818 Table of Atomic Weights. The following table
represents atomic weights of a few elements selected from Berzelius' 1818 and
1828 Tables, and recalculated for the standard oxygen 16 instead of 100 used by
Berzelius. The modern values also are given by way of comparison.
Table III. — Berzelius' Atomic Weights.
Elements
1818.
1828.
1919.
Carbon
1212
12-25
12-0
Oxygen
16-0
160
16-0
Sulphur
32-3
32-24
3206
Mercury
406
202-86
200-6
Iron
109-1
54-36
55-84
Sodium
93-5
46-62
23-00
Silver
433-7
216-61
107-88
COMBINATION BY VOLUME 189
Berzelius never succeeded in deciding whether the binary compounds he analyzed
contained two or more atoms per molecule. Without any rules to guide him, said
J. B. A. Dumas (1832), but guided mainly by analogies, he fixed by intuition the
atomic weights of a number of elements which subsequent experience has only
tended to confirm. When all is said, however, this method is unsatisfactory, because
it is arbitrary, and liable to be capriciously modified by each worker. In addition
to his volume law, Berzelius also used Dulong and Petit's rule of specific heats, and
Mitscherlich's isomorphic law to assist him in fixing the atomic weights of the
elements.
J. B. A. Dumas (1826-37).— In 1826, J. B. A. Dumas 3 pubUshed an important
memoir Sur quelques points de la theorie atomistique :
The object of these researches is to replace by definite conceptions, the arbitrary data
on which nearly the whole of the atomic theory is based.
Dumas accepted the hypothesis of Avogadro as propounded by Ampere, namely,
that equal volumes of gases contain an equal number of particles, and in the case
of the simple gases, that these particles are subdivided during chemical reactions.
J. B. A. Dumas thus recognized the importance of measuring the relative densities
of gases and vapours, and he devised his well-known method for determining these
constants. In 1832, he had determined the relative vapour densities of mercury,
iodine, phosphorus, sulphur, arsenic, etc., and noted some irregularities with sulphur,
mercury, phosphorus, and arsenic. He spoke of un demi-atome in the same way
that Avogadro spoke of une demi-molecule. J. J. Berzelius appears to have been
obsessed by his dualistic theory {q.v.), in which he assumed that the ultimate
particles (molecules) of elementary substances cannot be split when they form a
binary compound ; so that he wrote the formula of hydrogen chloride H2CI2 ; or
else he misunderstood J. B. A. Dumas, owing to the confusion of the words atom
and molecule, and was led to say (1826) : ^ "It is usually supposed that an
hypothesis ought to be abandoned as soon as it leads to an absurd conclusion; "
and if Avogadro' s hypothesis involves the subdivision of an atom, it must be
condemned.
In his Legons sur les philosophie chimique (Paris, 1837), J. B. A. Dumas appUed
Avogadro's volume law — equal volumes, an equal number of atoms — to explain the
formation of hydrogen chloride, HCl, and nitric oxide, NO, from the elementary
gases, and he showed that the physical atoms must be spHt during the reaction.
Hence, said J. B. A. Dumas, la chimie coupait les atonies que la physique ne pouvait
pas couper, so that Avogadro's molecules are Dumas' physical atoms. Here
again there was some confusion owing to the unfortunate use of the word atom
in place of Avogadro's molecule. J. B, A. Dumas then went on to show that
while Avogadro's volume law gives satisfactory values for the atomic weights
of oxygen, nitrogen, chlorine, bromine, and iodine, difficulties are encountered
with phosphorus, arsenic, mercury, and sulphur. For instance, ammonia is formed
by the union of three volumes of hydrogen and one volume of nitrogen ; and
phosphine, a similar gas, is presumably formed in a similar manner so that phosphine
should be produced by replacing the nitrogen of ammonia by an equal volume of
phosphorus gas. Consequently, it was argued that the density of phosphorus
vapour ought to be 31*4 (hydrogen unity— in the original, oxygen 100 is the
standard of reference) ; experiment gives a number twice as great ! A similar
discrepancy was found with arsenic. This can only mean that equal volumes of
the vapours of nitrogen, phosphorus, and arsenic do not contain the same number
of atoms. Again, J. B. A. Dumas showed that about 200 parts of mercury unite
with 16 parts of oxygen to form mercuric oxide, and therefore the atomic weight of
mercury must be nearly 200 ; but judging from the density of mercury vapour,
the atomic weight of this element is nearly 100. Consequently, le chaleur diviserait
les particles du corps plus que V action chitnique. Equal volumes of gases sometimes
contain an equal and sometimes an unequal number of atoms, and therefore the
190 INORGANIC AND THEORETICAL CHEMISTRY
determination of the densities of vapours cannot be a trustworthy guide in evaluat-
ing the atomic weights of the elements. The facts seemed to be against Avogadro's
hypothesis and J. B. A. Dumas accordingly gave it up in despair. He then tried
an application of Dulong and Petit's rule, but here again he was disappointed with
the exceptions ; and finally, after trying Mitscherlich's isomorphic rule, he said,
tout considers, la theorie atomique serait une science purement conjecturale, si elle ne
s'appuyait pas sur Visomorphisine.
W. Prout (1833) and A. Gaudin (1835).— W. Prout, in his work Chemistry . . .
considered with reference to Natural Theology (London, 122, 1833), adopted the
hypothesis of Avogadro's, viz. " under the same pressure and temperature, all
bodies in a perfectly gaseous state contain an equal number of self-repulsive mole-
cules," to explain the volume relations of hydrogen, oxygen and water, and of hydro-
gen, chlorine, and hydrochloric acid. W. Prout's explanation is almost as clear
as if it had been written to-day. From the observed results, said he,
It follows irresistibly that every self -repulsive molecule of oxygen has been divided into
two, and consequently must have originally consisted of at least two elementary molecules,
somehow or other associated, so as to have formed one self-repulsive molecule.
M. A. Gaudin, in his papers Recherches sur la structure intime des corps inorganiques,
published in 1833, had previously pointed out that J. B. A. Dumas' difficulty with
mercury and phosphorus could be explained by assuming that the mercury molecule
is monatomic, and that of phosphorus tetratomic ; evidently J. J. BerzeHus did not
like this mode of evading the discrepancy observed by J. B. A. Dumas, and con-
sidered it to be nur ein Spiel der Phantasie, although M. A. Gaudin's suggestion is
now generally accepted ; so also is Gaudin's happy use of the terms mono-, di-, tri-,
. . . atomic for indicating the number of atoms in a molecule.
Failures with Avogadro's hypothesis. — Towards the middle of the nineteenth
century, as a result of these failures to apply Avogadro's hypothesis, the atom was
abandoned by the majority of chemists as a discredited theory. In illustration,
J. B.A.Dumas said in 1837: Si fen etais lemaUre,feffacerais le motatome de la science.
WoUaston's equivalents were used, notably by L. GmeHn in his popular Handbuch
der theoretischen Chemie (Frankfurt-am-Main, 1817-9). In the early editions of this
book Gmelin used the term Mischungsgewichte — mixing weights — and in a later
edition (1843), the term stoichiometric numbers in place of equivalents. Gmelin
said, if an atom is the smallest quantity of a body which enters into combination,
the equivalents must represent atoms ; the atomic notation of Dalton is based on
hypothesis, equivalents are a reality. The inconsistencies involved in W. H.
WoUaston's equivalents were thus ignored. The whole subject at this time (1840-50)
was in a very confused state. In addition to the muddling of the terms atom,
molecule, and equivalent, there were tables of atomic weights, equivalents, H. V.
Kegnault's equivalents thermiques (1849) based upon Dulong and Petit's rule ;
H. Kose's (1857) and J. C. G. de lAa.T\giidi,G' ^equivalents isomorphiques (1855) based upon
Mitscherlich's rule ; and M. Faraday's electrochemical equivalents (1834). ^ Different
chemists used different standards for their equivalent and atomic weights. The
same chemical formula was used for different compounds, and different formulae
for the same compound — for instance, F. A. Kekule in his Lehrbuch der organischen
Chemie (Stuttgart, 1861) indicated nineteen different formulae which had been
proposed for acetic acid. Inorganic chemists thus failed to establish the conception
of an atom, but fortunately organic chemists had begun to see more clearly.
C. F. Gerhardt and A. Laurent (184^56).— In 1842, in a memoir entitled Re-
cherches sur la classification chimique des substances organiques, C. F. Gerhardt was
groping for a method of distinguishing between equivalent and atomic weights, and
he put forward some important views respecting the equivalents of certain elements
taking part in organic reactions. He showed that when an organic reaction gave
rise to water, carbon dioxide, carbon monoxide, or ammonia, the smallest amounts
produced are those represented by the formulae H4O2, C2O4, C2O2, S2O4, NH3
COMBINATION BY VOLUME 191
respectively, on the assumption that the equivalent or atomic weights are H=l,
0=8, 0=6, S=16, N=14:. Hence, the quantities indicated by the formulae must
represent an equal number of equivalents. It seems strange, said he, that no reaction
in organic chemistry can give rise to less than a single molecule of water, H4O2,
or carbon dioxide, C2O4 ; and that these quantities of gases occupy equal volumes.
Consequently, H4O2 and C2O4 represent either one or two equivalents ; the former
hypothesis fits the facts best, and therefore he argued that the equivalents of the
elements 0=8, C=6, and S=16 should be doubled so that the preceding formulae
can be written H2O, CO2, CO, SO2, and NH3 respectively. C. F. Gerhardt thus
obtained numbers for the equivalents of the elements, hydrogen, oxygen, carbon,
sulphur, and nitrogen in agreement with the atomic weights used by BerzeUus in
1826.
C. F. Gerhardt also advocated the adoption of a common standard for comparing
chemical formulae with one another, and he recommended the use of what is known
as the two-volume standard : those quantities by weight which occupy two volumes
when in the gaseous state and when the volume of atomic hydrogen is taken as unity.
Hence, Avogadro's hypothesis is sometimes called the Avogadro-Gerhardt law. In
a later part of his paper, Gerhardt showed that his notions of atomic weights, the
volume theory, and equivalents were not clear, because he stated that all these
concepts coincide. A. Laurent (1846), however, obtained a clear grasp of the
meanings to be attached to these terms, and he adopted Gerhardt's happy idea that
chemical formulae should represent comparable quantities ; he also adopted the
two-volume standard, but in doing so he was obliged to admit that there are some
exceptions — e.g. the vapour of ammonium chloride, and sulphuric acid — which
seemed to correspond with a four-volume standard. The names of Laurent and
Gerhardt are usually linked together ; it is, indeed, difficult to isolate the particular
contributions made by each because they published a great deal jointly, and, being
intimate friends, they probably discussed the whole subject together. A. Laurent's
posthumous Methods de chimie (Paris, 1854) and C. F. Gerhardt's Traite de chimie
organique (Paris, 1856), ^ did much to clear the conceptions of equivalents, atoms,
and molecules ; and their definitions of these entities, and most of their formulae
for organic and inorganic compounds are virtually in use to-day. Laurent repre-
sented the union of hydrogen and chlorine by the equation (HH)-|-(C1C1)
=(HC1)+(HC1) ; to-day we write, H2+Cl2=2HCl ; similarly the synthesis of
water was symbolized, (HH)+(HH)+(00)=(HH)0-f (HH)0 ; to-day we write,
2H2H-02=2H20. A. Gaudin (1832) employed special diagrams to symbolize these
reactions.
Several other systems of symbolizing chemical operations have been proposed. A. C.
Brown (1867), for example, used Greek letters to represent different chemical action, thus
(f> represented the replacement of hydrogen in a molecule by the radicle CHg.and if o denotes
a molecule of ammonia, NH3, the symbol <f)a represented the substitution of one hydrogen
atom in ammonia to form CH3NH2. In view of the great variety of chemical processes
and compoimds, such a system would be more cumbrous and throw greater strains on the
memory than the present system. B. C. Brodie proposed a new notation in his Calculus
of Chemical Operations ( 1867), which he regarded as " a rigid expression of fact, independent
of the atomic hypothesis. B. C. Brodie's system, however, involved assumptions even
more drastic than the atomic theory, and the notation was so confusing that it died as soon
as it was born.
S. Caimizzaro (1857-8).— In 1857, S. Cannizzaro stated his belief :
There are no exceptions to the universal law that equal volumes of f «««j;°"*^*j^^
numbers of molecules, and that the apparent exceptions wiU disappear when more searching
experiments are made.
He showed that the apparent exceptions to C. F. Gerhardt's two-volume law.pointcd
out by A. Laurent, are not real, for the work of H. St. C. Dev.Ue (1857) has shown
that ammonium chloride and sulphuric acid are decomposed by heat, and theretore
the observed vapour densities are the densities of mixtures of the decomposition
192 INORGANIC AND THEORETICAL CHEMISTRY
products and not of homogeneous compounds to which Avogadro's hypothesis alone
refers. Similar conclusions were deduced independently and almost simultaneously
by H. Kopp (1858) ^ and F. A. Kekule (1858). Immediately afterwards, S. Canniz-
zaro pubUshed his celebrated Sunto di un corso di Jllosofia chimica fatto nella Reale
Universita di Genova (1858), which placed Avogadro's hypothesis at the foundation
of the system of chemistry which obtains to-day — witness, among other works,
W. Nernst's popular Theoretische Chemie vom StandjmnJcte der Avogadroschen Regel
und der Thermodynamik (Stuttgart, 1916). The atomic theory of the present-day
chemistry is the work of many minds. In the words of G. Chrystal (1885) :
Few scientific ideas spring up suddenly without previous trace or history ; a close
examination always shows that the sprite was in the air before the Prospero came to catch
him. . . . There are long periods in science in which great improvements were effected
which cannot be traced to any individual, but seem to have been due merely to the working
of the minds of scientific men generally upon the matter, one giving it this little turn, another
that, in the main, always for the better.
Befebekces.
1 C. Graebe, Journ. prakt. Chem., (2), 87. 145, 1913 ; W. H. WoUaston, Phil. Trans., 104. 1,
1814 ; T. Thomson, The Elements of Chemistry, Edinburgh, 1810 ; E. Hjelt, Berzelius—Liehig—
Dumas. Ihre Stellung zur Eadikaltheorie, 1832-1840, Stuttgart, 1908 ; S. Cannizzaro, Historische
Notizen und Betrachtungen iiber die Anwendung der Atomtheorie in der Chemie und uber die
Systeme der Konstitutionsformeln von Verhindungen, Stuttgart, 1913 ; A. N. Meldrum, Avogadro
and Dalton. The Standing in Chemistry of their Hypotheses , Edinburgh, 1904.
2 A. Ladenburg, Vortrage vber die Entwickelungsgeschichte der Chemie in den letzten 100 Jdhren,
» J B. A. Dumas, Ann. Chim Phys., (2), 49. 210, 1832 ; 50. 170, 1832.
4 J. B. A. Dumas, Ann. Chim, Phys., (2), 33. 337, 1826.
^ H. V. B-egnault, Cours elementaire de chimie, Paris, 1849 ; H. Rose, Pogg. Ann., 100. 270,
1867; J. C. G. de Marignac, Archiv. Sciences Geneve, 2. 89, 1858; M. Faraday, Phil. Trans.. 124.
77, 1834.
^ A. Laurent, Ann. Chim. Phys., (3), 18. 266, 1846 ; A. Gaudin, Eecherches sur la structure
intimes des corps inorganiques definis, Paris, 1832; Ann. Chim. Phys., (2), 52. 113, 1833; A. 0.
Brown, Laboratory, 1. 37, 1867.
' H. Kopp, Liebig's Ann., 105. 390, 1858 ; F. A. Kekule, ib., 106. 143, 1858.
§ 8. Deviations from Avogadro's Law
When a fact appears to be opposed to a whole train of deductions, it invariably proves
to be capable of bearing some other interpretation. — Sherlock Holmes.
It is sometimes said that a phenomenon " ought to take place," but it does not ;
the phenomenon is then said to be abnormal or anomalous. These terms are not
very happily chosen, and they are sometimes used rather carelessly ; they are not
intended to imply that nature is erratic, arbitrary, and lawless. The words simply
mean that in groping for the truth, an unexpected result has been obtained, which
once stood, or now stands, challenging investigators to show how the unexpected
should have been expected. In this sense it has been said that abnormal phenomena
do not occur in nature. Some of the most treasured generalizations in science have
been won by investigating the abnormal. This applies both in the laboratory and
in the study.
Abnormal vapour densities. — According to Avogadro's hypothesis, if the relative
density of hydrogen be taken as unity, the quotient M/D=2, where M denoted the
molecular weight, and D the relative density of the gas. Some puzzling exceptions
to this rule were encountered during the early application of the hypothesis, for
several substances do not conform to the ratio when molecular weights deduced
by the ordinary chemical methods are employed, and, in consequence, these sub-
stances were said to possess abnormal vapour densities. This led chemists to look
upon Avogadro's rule with suspicion, and there were some controversies as to
whether (i) substances with abnormal vapour densities really follow Avogadro's rule ;
COMBINATION BY VOLUME 193
or whether (ii) substances with an abnormally low vapour density are dissociated
into simpler molecules, and substances with an abnormally high vapour density
are associated into more complex molecules. J. B. A. Dumas (1836) ^ thought
that the abnormal vapour densities invahdated the hypothesis, while M. A. Gaudin
(1833) considered that the alleged failure was due to a pecuHarity in the molecules
of the gas, which, when taken into account, left the hypothesis quite valid. In-
dependent proofs of the validity of M. A. Gaudin's inference are discussed later on
when the particular substances are treated. As soon as M. A. Gaudin's interpre-
tation had been demonstrated experimentally, Avogadro's hypothesis won its way
into the heart of chemical science.
There are two possible deviations with compounds ; the ratio
Molecular weight , . , ., ^
— .^ — o — may be greater or less than 2
when the density of hydrogen is taken imity. In the case of elementary gases,
S. Cannizzaro (1858) showed that the atomic weight^ is equal to half the vapour
density of the gas, if hydrogen 2 be the unit, or to the vapour density itself, ^/Z)=l,
if hydrogen be unity. Here, again, there are two possible deviations :
Atomic weight , , , ^t. -^
— =r rr— ^ — may be greater or less than unity
when hydrogen unity is the standard of reference. The interpretation of the results
in the two cases are similar.
The molecules of the substance are decomposed or dissociated ; the molecules are
actually less complex than corresponds with the simple chemical formulae, and the
ratio MjD is greater than 2, or the ratio AjD is greater than unity. For example, the
vapour density of steam is 9 (H=l), the molecular weight 18, and the ratio MID=2;
at a very high temperature, there are reasons for supposing that the vapour density
would be 6, and the ratio M/D would appear to be 3. This corresponds with the
value of M/D on the assumption that the steam is dissociated into its elementary
molecules : two volumes of hydrogen, and one volume of oxygen, so that the density
of a mixture is involved and not that of a homogeneous substance as is required if
Avogadro's rule is to be applied. The density of such a mixture will be (24-16)-^3
=6 ; the assumed dissociation thus gives a number in agreement with observation.
If the observed density were 8, this would represent a mixture with 33j per cent,
of dissociated, and 66| per cent, of undissociated steam. The cases with phosphorus
pentachloride, PCI5 ; ammonium chloride, NH4CI ; sulphuric acid, H2SO4 ; mer-
curous chloride, HgCl ; nitrogen peroxide, N2O4 ; and hydrogen iodide, HI, are
discussed later. With elementary gases, J. B. A. Dumas (1832) found that mercury
vapour has a density of 100 corresponding with an atomic weight of 100, but the
atomic weight deduced by chemical methods is 200, consequently v4/Z)=2 instead
of 1. It is therefore assumed that the molecule of mercury vapour is monatomic
and MID=2, while A/D=l. The cases with iodine, the metal vapours, etc., are
discussed later.
The molecules of the substance are associated or condensed ; the molecules are
more complex than corresponds with the simple chemical formulae ; and the ratio
M/D is less than 2, or the ratio A/D less than unity. The molecular weight of acety-
lene, C2H2, is 26, the vapour density is 13, and the ratio M/D is normal. Benzene
has exactly the same chemical composition, and its vapour density is 39 (H unity) ;
if the molecular weights of the two gases be the same, the ratio M/D for benzene
would be 0-67, but if benzene be more complex than acetylene, say (C2H2)3 or C,,H<,,
the molecular weight of the complex molecule will be 78, and the ratio MJD becomes
normal. Hence, for this and other reasons, benzene is regarded as if it were a
product formed by the condensation of three molecules of acetylene. Phosphorus
trioxide and pentoxide, and other examples, are discussed later. With elemental
gases, J. B. A. Dumas (1832) found that the density of phosphorus vapour is 62*8,
VOL. I. ^
194
INOKGANIC AND THEORETICAL CHEMISTRY
and the atomic weight deduced by chemical methods, by analogy with nitrogen,
is 31 "4, so that the ratio A/D is one-half. This is taken to mean that the molecular
weight of phosphorus is not that equivalent to P2> ^^^ is rather equivalent to P4.
Sulphur and arsenic are discussed later.
The effect of changes in the molecular weight of a gas on the laws of Boyle and
Charles. — The gas equation,
must now be revised in order to allow for changes in the molecular weight of the gas
when it changes from one state to another. Remembering that the density D of a
gas is equal to the molecular weight M divided by the volume v, or M=Dv, the
gas equation can be written,
P _ n
TD ~ T^D^
provided M=Mi. Let M, 7), and v respectively denote the molecular weight,
density, and volume of the gas by one condition of temperature and pressure ; and
Ml, Di, and %,the same constants for another condition of temperature and pressure,
then, by substitution in a preceding equation, pvlMT=piVilMiTi. If the volume
Vi at some standard temperature Ti and pressure pi be taken, the numbers pi, Vj,
and Ti will always have one fixed value. Let R denote this constant value of
PivJTi. The gas equation then assumes one of the forms :
pv
MT
^£^^;oT,pv=^^RT',ov,pv = iRT
where * stands in place of the ratio of the molecular weights of the gas in the two
conditions, M/M^. If the molecules of the gas neither dissociate nor polymerize
when the conditions change, M=Mi ; or pv=RT because i—\. Again, if the gas
molecules polymerize or condense so that, say, two molecules combine together to
form one molecule, there will be only half as many molecules in a given space as
before : M=^Mi, and pv=\RT. If, however, the gas dissociates or decomposes
so that each molecule of the gas forms two molecules of another gas or gases, then
M=2Mi, and pv=2RT. Hence, the ordinary gas equation, pv=RT, is a special
case of the more general relation, pv=iRT, where the numerical value of i
indicates whether or not the gas keeps the same molecular concentration during
the change. K i=l, there is neither dissociation nor polymerization ; if i be
less than unity, the gas polymerizes ; and if i be greater than unity, the gas
dissociates when the conditions are changed.
The effect of deviations from Avogadro's hypothesis on Gay Lussac's law of
volumes. — The molecular volumes of many gases are not all the same, and they
thus exhibit small deviations from the law MID=2 (hydrogen unity). This is
shown for a few gases at 0° and 760 mm. in the following table :
Table IV.< — A Comparison of the Molecular Volumes of Some Gases.
Gases.
Molecular
Observed density
Molecular
weight M.
(0=16).
volumes M/D.
Oxygen, Oj .
32
16
2-000
Nitrogen, Ng .
28-02
14-00
2-001
Carbon monoxide, CO
28-00
14-01
2-000
Carbon dioxide, CO 2 .
44-00
22-15
1-988
Methane, CH4 .
1603
803
1-998
Ethane, CjHe .
30-05
15-20
1-980
Ethylene, C2H4
28-03
14-28
1-966
Acetylene, C2H2
2602
i
13-12
1-984
COMBINATION BY VOLUME 195
In calculations involving gaseous volumes, the errors due to the deviations of
the molecular volumes from the theoretical may be greater than the experimental
errors. Instead of writing the reaction, 200+02=2002, in the form, 2 Vols.
00+1 Vol. 02=2 Vols. CO2; it becomes necessary to write 2xO-994=l-988
volumes of carbon dioxide, and the equation becomes 2 Vols. CO+1 Vol. 02=1 '988
Vols. CO2. Similarly, with equations involving other discrepant gases. If the
partial pressure of the deviating gas be less about 25 per cent., the discrepancy
may be disregarded since the lower the partial pressure of the gas, the more
nearly does it behave like an ideal gas. Thus, the lower the pressure confining
carbon dioxide, CO2, at 20°, and of ethane, C2H6, at 0°, the more nearly do the
molecular volumes approach the value 2 for ideal gases.
Pressure
100
300
500
600
700
760 mm.
Carbon dioxide
. 1-998
1-996
1-994
1-992
1-990
1-988
Ethane
. 1-998
1-992
1-988
1-984
1-982
1-980
Correction of the ratio M/D for gases which deviate from Boyle's law.— It
follows from Avogadro's hypothesis : (i) The molecular volumes — i.e. the quotients
of the molecular weights M by the respective densities D — of all gases are the same,
so that 3Ii : Di=M2 : D^, and (ii) the molecular weights of all gases are pro-
portional to their densities, so that Mj : M^^D^ : D2. These deductions can be true
only for gases in which the pressure is not affected by intermolecular attractions
as is the case with gases which follow the simple gas laws. Densities calculated for
gases which do not conform with Boyle's law do not agree satisfactorily with obser-
vations unless the gases are attenuated or rarefied, thus showing that Avogadro's
hypothesis is not strictly accurate with gases under normal pressure. Similarly,
the experiments of H. V. Regnault 2 (1847) and others have shown that Boyle's
and Charles' laws approach exactitude only when the pressures are very small.
Gases approach the so-called ideal state when their pressures are reduced ; and, at
the limit, when the pressures are indefinitely small, Avogadro's hypothesis is strictly
valid. Otherwise expressed, the molecular volumes of all gases are exactly the
same only when the gases are extremely rarefied ; and the limiting value of the
ratio of the densities D^ and D^ of two gases will be equal to the ratio of their
molecular weights M^ and M^ only when the pressures of the respective gases
approach zero. The deviation of a gas from Boyle's relation Po%/pv=l, or
I—Po^qIpv^^O, can be symbolized :
^^l_Mo (1)
pv
where p^ and Vq respectively denote the atmospheric pressure and volume of the
gas at 0° ; and p and v the corresponding values at some small pressure. For the
so-called permanent gases, Kegnault's experiments show that the coefficient A
is very nearly constant between one and six atmospheres pressure. Consider two
gases under a very small pressure p ; let each be subjected to atmospheric pressure
Pq when the volumes become respectively Vi and v^ ; then, Vi=vp{l—Ai)/po ; and
V2=vp{l—A2}lj)Q ; and by division,
h _ Izi^i " (2)
v,-l-A, . • • •
This means that the molecular volumes of the two gases under atmospheric pressures
have the proportional values 1—Ai and 1— .42- Let Di and D2 denote the re-
spective densities of the gases under atmospheric pressures— temperature constant
—then, the ratio of the molecular weights Mi : M2 is equal to the ratio of the pro-
ducts of their molecular volumes by the corresponding densities ; that is, to the
ratio Di(l—Ai) : 2)2(1—^2) 5 or,
Mi_Di{i-zAi) (3)
M2~D2{1-'A2) • • • • V '
196
INORGANIC AND THEORETICAL CHEMISTRY
If A for the two gases be zero the expression reduces to that required by Avogadro's
rule. The densities employed in calculations with formula (3) are the weights in
grams of a normal litre of the respective gases ; the evaluation of the coefficients A
is a problem for the physicists. A number of values have been determined, but the
task is a difficult one, and is subject to some uncertainty since it involves an extra-
polation of the pv and y-curve. The following values for A] between atmospheric
and zero pressures are compiled from data by A. Leduc (1898), R. W. Gray and
F. P. Burt (1909), and A. Jaquerod and 0. Scheuer (1908) 3 :
Table V. — NumebicaIj Values of the Coefficent A.
Gas.
Ah
Gas.
Ah
Hydrogen ....
-0-00056 .
Nitrous oxide .
+0-00750
Oxygen
+0-00096
Hydrogen chloride
+0-00786
Nitrogen .
+0-00044
Sulphur dioxide
+002314
Carbon monoxide
+0-00060
Methane
+0-00175
Nitric oxide
+0-00114
Ethane .
+0-01194
Carbon dioxide .
+0-00678
Methyl chloride
+ 0-02468
Ammonia ....
+0-01504
Ether ....
+0-02587
If oxygen be the standard gas with ^2=32, D2=1'4290, and ^2=0*00096, it
follows that if the numerical values of the density D and the deviation a be known,
the
Molecular weight=22-5739D(l— ^)
The results computed by this method are in fair agreement with the values obtained
by chemical processes. For example, with oxygen, 32, as standard
Hydrogen. Nitrogen. Carbon monoxide. Nitric oxide. Methane.
M (Chemical) . . 2-015 28-019 28-009 30-006 16-039
M fPhysical) . . 2-016 28-020 28-000 30-010 16-032
These results are in close agreement. This physical method thus rivals in accuracy
the molecular weights of the permanent gases determined by chemical processes.
There are not so many complications with physical methods as are involved in
conductmg a series of chemical operations with pure substances. This physical
method is known as D. Berthelot's limiting density method of determining
molecular weights.^ The data required for the application of Berthelot's method
are (i) the densities, and (ii) the compressibility of the gas under investigation ;
and also (iii) the compressibility of the standard gas.
With the more easily liquefiable gases, the coefficient A changes rapidly with
changes of pressure, and consequently A cannot be assumed constant without
sensible error. It is therefore necessary to use values for the coefficients A deter-
mined for the variations of pressure near to those under which the density has to be
determined. The available data are not sufficiently exact to enable the method to
be used for accurate molecular weights of such gases, the coefficients A are usually
too high, and the molecular weights correspondingly low. For instance :
Carbon dioxide.
Nitrous oxide.
Ethylene.
Ammonia.
Sulphur dioxide.
M (Chemical
. 44-000
44-020
30-048
30-034
64-070
M (Physical)
. 44-013
44-003
30-037
30-018
64-063
where the comparison is not so favourable.
According to D. Berthelot,* the molecular weight 3f of a normal liquid is related
with its critical density Z)„ critical pressure pc, and critical temperature Tc by the
formula
M = 22-4
3-6 ' 273
J.
Pc
COMBINATION BY VOLUME 197
where 3" 6 represents the mean value of the actual to the theoretical density at the
critical temperature for normal or non-associated liquids. E. Mathias has also
shown that in accord with the law of rectilinear diameters, the critical density of a
substance is related to the densities of the Uquid Di and of the saturated vapour Z>r
at a temperature T by the expression
n - I>i-J^v . or, ri ^ A
when the temperature does not exceed the boihng point of the liquid under atmo-
spheric pressure. Consequently, by substitution of the second of these equations
in that of Berthelot,^
Molecular weight = 11-4 ^ ^
The molecular weights of substances which are liquid at ordinary temperatures,
calculated by this expression, are often a little too high. For example —
Cya SO2 CCI4 CS2 NH3 HjO SnCl*
Calculated . . 50 GSl 152*3 73'4 19-2 251 252-4
Formula weight . . 52 64 1538 76 17 18 260
References.
1 J. B. A. Dumas, Lecons sur la philosophie chimique, Paris, 1836 ; Ann. Chim. Phya.^ (2)»
49. 210, 1832 ; (2), 50. 170, 1832 ; M. A. Gaudin, ih., (2), 52, 113, 1833 ; S. Cannizzaro, Nuovo
Cimento, 8. 71, 1858.
2 H. V. Regnault, Mem. Acad., 21. 329, 1847.
3 R. W. Gray and F. P. Burt, Journ. Chem. Soc., 95. 1633, 1909 ; A. Jaquerod and 0. Scheuer,
Mem. Soc. Phys. Nat. Geneve, 35. 659. 1908 ; A. Ledue, Ann. Chim. Phys., (7), 15, 6, 1898 ; (8),
19. 441, 1910.
4 D. Berthelot, Ccrmpt. Rend., 126. 954, 1898; Journ. Phys., (3), 8. 263, 1899; Zext.
EleUrochem., 34. 621, 1904 ; Lord Rayleigh, Phil. Trans., 198, 417, 1902 ; Proc. Roy. Soc., 73.
153, 1904 ; H. F. V. Little, Science Progress, 7. 504, 1913 ; G. Baume, Journ. Chim. Phys., 6.
52, 1908 ; P. A. Guye, ib., 6. 778, 1908 ; 17. 141, 1919. ^ ,^
5 D. Berthelot, Compt. Rend., 128. 006, 1899; C. M. Guldberg, Zeit. phys. Chem., 32. 116,
1900 ; E. Mathias, Le point critique des corps purs, Paris, 164, 1904.
§ 9. Radicals or Radicles
For the chemist, each molecular compound is proximately made up of less compound
atoms which are indivisible by forces which can divide their product, and these m turn can
be separated by chemical agents into simple atoms. — S. Bbown.
In 1815, J. L. Gay Lussac,i after studying the properties of hydrocyanic acid,
reported cyanogen (CNjg, to be "a remarkable example, and at present, a unique
example, of a body which, although a compound, plays the part of a smgle body
in its combinations with hydrogen and the metals." Indeed, if chemists did not
know how to resolve cyanogen into its constituent elements, this compound would
very probably be classed as an element, and further, it would probably be assigned
a place in the halogen family of elements to be studied later. Since Gay Lussac s
discovery a great number of similar groups of what might be called pseudo-elements
have been found. For convenience, they are commonly called radicals ov, following
the custom of the London Chemical Society, radicles. There have been periodic
discussions on the spelUng of the term— radicle or radical. The latter is
taken to be historically correct, and the former etymologically correct.- Ihe word
198 INORGANIC AND THEORETICAL CHEMISTRY
radical was previously employed by G. de Morveau (1787) and by A. L. Lavoisier 3
with a different meaning, for with A. L. Lavoisier a radicle could be a simple or com-
pound body ; he says, le carhone est le radical de Vacide carhonique, and added that
vegetable acids contain le radical oxalique, tartarique, etc. The definition : a radicle
is a group of atoms which can enter into and be expelled from combination with-
out itself undergoing decomposition, is virtually that given by J. von Liebig in 1838.
Each radicle acts as an unchanging constant in a series of compounds ; and each can
be replaced by an equivalent element or elements. In very few cases has it been
possible to isolate the radicle, but the definition has nothing to say about the inde-
pendent existence of radicles. " Radicles," said A. Kekul^ (1858), " are not firmly
closed atomic groups, but they are merely aggregates of atoms placed near together
which do not separate in certain reactions, but fall apart in other reactions." For
convenience, the term radicle is sometimes applied to an atom in a compound
which can be replaced by another atom or radicle without a further change in the
nature of the compound ; in that case, the radicle is said to be a simple radicle, in
contrast with compound radicles, which are groups of atoms.
References .
1 J. L. Gay Lussac, Ann. Chim. Phys., (1), 95. 136, 1815.
2 Anonymous, Chem. News, 9. 143, 166, 191, 204, 1864; E. Divers, ib , 54. 36, 260, 1886;
J. SpiUer, ib., 54. 83, 1886; H. G. Madan, ib., 54. 71, 1886 ; Nature, 33. 535, 1886; J. F. Heyes,
ib., 33. 559, 1886.
3 J. von Liebig, Liebig's Ann., 25. 113, 1838 ; A. Kekule, ib., 106. 129, 1858 ; A. L. Lavoisier,
G. de Morveau, and A. F. de Fourcroy, Methode de nomenclature, Paris, 1787; A. L. Lavoisier,
Traite elementaire de chimie, Paris, 1789.
§ 10. The Atomic Weights of the Elements
Every chemical element is regarded as having a distinct and definite nature of its own,
which natm-e, moreover, determines all its activities.- — B. P. Browne.
The ratio between the atomic weights of oxygen and hydrogen is the base-line upon
which our entire system of atomic weights depends. — F. W. Clarke (1896).
What are the best representative values for the atomic weights of the elements ?
— The best available determinations of the value of the oxygen-hydrogen ratio give
numbers ranging between r005 and 1'008 when the standard reference is oxygen 16.
All measurements made by man are affected by unavoidable errors of experiment ;
and measurements of the numerical value of all constants differ within certain
Umits amongst themselves. It is convenient to select one representative value
from the set of different observations ranging between the limits I'OOS and 1*008.
The majority of chemists have agreed to let the International Committee of Atomic
Weights decide what are the best representative values for the atomic weights of
all the elements year by year. Hence, the generally accepted ratio for the atomic
weights of hydrogen and oxygen is 1*008 : 16. Every time new and more refined
methods of measurement are employed, a change— generally insignificantly small — ■
may be necessary. It must be recognized that the true atomic weights cannot be
altered by the votes of the majority of the members of the International Committee
of Atomic Weights.! There is an uncertain factor in the accepted values of the
atomic weights, as there is in all our judgments. Aristotle was no doubt right,
" Nothing can be positively known, and even this cannot be positively asserted."
This doctrine, however, if rigorously applied, would paralyze all action. Accord-
ingly, sound-minded people are accustomed to balance the evidence and then act.
A careful consideration of all the available evidence considerably reduces the risk
of error, and this method adopted by the Committee appears to be the most satis-
factory solution of the problem.
The atomic weights of the elements are indicated in the following table. The
numbers are those recommended by the International Committee on Atomic
COMBINATION BY VOLUME 199
Weights (1920). The atomic number, indicated in the same Table, will be
discussed later.
Table VI. — International Atomic Weights (1921). 0=16.
Atomic
number.
Symbol.
Atomic
weight.
Atomic
number.
Symbol.
Atomic
weight.
Aluminium
13
Al
271
Molybdenum
42
Mo
960
Antimony .
51
Sb
120-2
Neodymimn
60
Nd
144-3
Argon
18
A
39-9
Neon
10
Ne
20-2
Arsenic
33
As
74-96
Nickel
28
Ni
58-68
Barium
56
Ba
137-37
Niobi\ma (Colum-
Beryllium (Gluci-
bium
41
Nb
931
num)
4
Be
9-1
Niton (radium
Bismuth
83
Bi
208-0
emanation)
86
Nt
222-4
Boron
5
B
10-9
Nitrogen .
7
N
14-08
Bromine
35
Br
79-92
Osmium
76
Os
190-9
Cadmium .
48
Cd
112-40
Oxygen
8
0
1600
Caesium
65
Cs
132-81
Palladium .
46
Pd
106-7
Calcium
20
Ca
40-07
Phosphorus
16
P
31-04
Carbon
6
C
12-005
Platinum .
78
Pt
195-2
Cerium
58
Ce
140-25
Potassiimi .
19
K
39-10
Chlorine
17
CI
35-46
Praseodymiiun .
69
Pr
140-9
Chromium .
24
Cr
520
Radiiim
88
Ra
226-0
Cobalt
27
Co
58-97
Rhodium .
46
Rh
102-9
Columbium (Nio-
Rubidium .
37
Rb
85-45
bium)
41
Cb
93-1
Ruthenium
44
Ru
101-7
Copper
29
Cu
63-57
Samarium .
62
Sa
150-4
Dysprosium
66
Dy
162-5
Scandium .
21
Sc
441
Erbiima
68
Er
167-7
Selenium
34
Se
79-2
Europium .
63
Eu
152-0
Silicon
14
Si
28-3
Fluorine
9
F
19-0
Silver
47
H
107-88
Gadolinium
64
Gd
157-3
Sodium
11
Na
23-00
Gallium
31
Ga
70-1
Strontium .
38
Sr
87-63
Germanium
32
Ge
72-5
Sulphur
16
S
3206
Glucinum (Beryl-
Tantalum .
73
Ta
181-5
lium)
4
Gl
9-1
Tellurium .
62
Te
127-5
Gold .
79
Au
:197-2
Terbium
66
Tb
169-2
Heliimi
2
He
4-00
Thalliima .
81
Tl
2040
Holmium .
67
Ho
163-5
Thorivmi
90
Th
232-15
Hydrogen .
0-95
H
1-008
Thuliima
69
Tm
168-5
Indium
49
In
114-8
Tin . . .
60
Sn
118-7
Iodine
63
I
126-92
Titanium .
22
Ti
481
Iridium
77
Ir
193-1
T\mgsten .
74
W
184-0
Iron
26
Fe
55-84
Uranium
92
U
238-2
Krypton
36
Kr
82-92
Vanadium .
23
V
610
Lanthanum
57
La
139-0
Xenon
64
Xe
130-2
Lead .
82
Pb
207-20
Ytterbium (Neo-
Lithiima
3
Li
6-94
ytterbium)
70
Yb
173-5
Lutecium .
71
Lu
175-0
Yttrium
39
Yt
89-33
Magnesium
12
Mg
24-32
Zinc .
30
Zn
65-37
Manganese
25
Mn
54-93
Zirconium .
40
Zt
90-6
Mercury
80
Hg
200-6
For ordinary calculations involving the use of atomic weights, most of these
constants, excepting chlorine (35-5), copper (63-5), and zinc (65-5), are rounded
off to the nearest whole numbers. The elements just named are then assigned the
constants indicated in the brackets. The atomic weight table made by J. J.
Berzelius in 1826 has excited admiration on account of its accuracy, ^^'ith the
standard 0=16, most of J. J. Berzelius' numbers are remarkably close to those we
are using to-day. For instance, with the common elements :
o. H. N. ci.
Berzelius' atomic weights 16 1 14-15 35*47
To-day's numbers 16 1-008 14 35-46
s.
P.
Pb.
Cu.
32-2
31-4
207-4
63-4
32-07
31-0
207-1
63-6
200 INORGANIC AND THEORETICAL CHEMISTRY
This is a testimoDy to the accuracy of J. J. Berzelius' work and particularly so when
the state of the knowledge of analytical chemistry in Berzelius' time is borne in
mind.
Are atomic weights whole numbers ? — ^It must be added that although we
are compelled to take the numbers as we find them, yet, the experimental errors
involved in a complex operation are great, and these errors are sometimes so
obscured by a cloud of auxiliary calculations that they are not always easy to detect.
Consequently, G. D. Hinrichs (1893) suggests that the true atomic weight of an
element must be regarded as a limit to which the observed values approach as the
disturbing factors are eUminated.- It required a century of measurements on the
density of atmospheric nitrogen before the presence of 1 per cent, of argon was
detected therein. Accordingly, many chemists firmly believe that the rounded
numbers are the best representative values of the atomic weights, and that the
small deviations from the rounded numbers indicated in the International Table
represent real, if unrecognized, errors of experiment ; M. Rudolphi (1901) also
attributed the deviations of the atomic weights from whole numbers to the presence
of small quantities of unknown elements whose properties are closely allied to the
elements with which they are mixed.
Why is oxygen 16 taken as the standard in preference to hydrogen unity ?—
During the latter part of the nineteenth century, J. Dalton's (1803) standard
hydrogen unity, was used for the atomic weights instead of oxygen 16. Hydrogen
was selected as a standard for gas densities and atomic weights because it is the
lightest element known. In determining atomic weights, it will be observed that
one of them, say A, is arbitrarily fixed as a standard, and the atomic weights of the
other elements are fixed through the relations B=liA ; C^^^k^B ; D=]c^C ; . . .
where ^i, k2, k-^, . . . k^ are numerical ratios. Here, obviously, the numerical
ratios referred to the element A as standard are :
Hence, since each observed ratio k embodies unknown errors, the errors will accumu-
late most on that particular ratio which is least directly connected with the standard
of reference, A. Consequently, J. S. Stas (1860-65) pointed out, as J. J. Ber-
zeHus (1818) did before him, that the determination of the atomic weight of an
element should be connected with the standard as directly as possible. Very few
compounds of the metals with hydrogen are suitable for an atomic weight deter-
mination, while nearly all the elements form stable compounds with oxygen. Hence,
if hydrogen be the standard, it is necessary to find the exact relation between the
given element and oxygen, and then calculate what that relation would be on the
assumption that the relation between hydrogen and oxygen is known. C. W.
Blomstrand expressed similar ideas in his Die Chemie der Jetztzeit (Heidelberg,
1869) ; he said the atomic weights of practically all the elements are compared with
hydrogen through the intervention of oxygen. Hydrogen compounds — hydrides —
are comparatively rare ; oxygen compounds^ — oxides — are common. Hence, the
weight ratio between oxygen and hydrogen must be known with great accuracy
since a small error becomes cumulative and it becomes serious in elements with a
large atomic weight — e.g. with uranium, the experimental error is multiplied about
15 times. Every improved determination of the relation between hydrogen and
oxygen would then be followed by an alteration in the weight of every other element
whose value, with respect to hydrogen as a standard, has been determined by the
indirect process just indicated, for as J. J. Berzelius said in 1816, oxygen is a kind
of nucleus about which chemistry has grown. The determination of the exact
relation between hydrogen and oxygen appears to be more difficult than many
other determinations, and hence, the majority of chemists think it better to refer
the atomic weights of the elements to oxygen 16 as the standard instead of making
COMBINATION BY VOLUME 201
the atomic weights depend on the more or less uncertain relation H : O. Hydrogen
is a theoretical standard, oxygen is the real basis. The standard oxygen 16 is quite
arbitrary. G. D. Hinrichs (1893) proposed carbon (diamond) =12 as the standard
of reference. T. Thomson (1825) used oxygen 1 ; W. H. Wollaston (1814), oxygen
10 ; J. S. Stas (1860-65), oxygen 16 ; and J. J. Berzelius (1830) oxygen 100 as
standard. The latter number makes the atomic weights of many elements incon-
veniently large, and if the atomic weight of oxygen be any whole number less than
16, fractional atomic weights will be required. The use of the oxygen 0=16 unit
involved the least change in the number in vogue when hydrogen unity was the
standard.
This question of a standard is not of mere academic interest, because, in buying
and selling ores on the percentage amount of contained metal, a difference in the
atomic weight selected may involve appreciable differences in the estimated value
of the ore. For instance, if oxygen be taken 16, the corresponding atomic weight
of antimony is 1199, and of uranium 239'61 ; if hydrogen be taken as unity, these
values become respectively 118'9 and 2 37 "6 5— differences of one and two units.3
Refebences.
1 P. A. Guye, Journ. Chim. Phys., 14. 449, 1916 ; T. Renard, ih., 15. 541, 1917.
2 G. D. Hinrichs, The True Atomic Weights of the Chemical Elements and the Unity oj Matter,
St. Louis, 1894 ; E. W. Morley, Journ. Amer. Chem. Soc., 22. 57, 1900 ; J. S. Stas, Recherches sur
les rapports reciproques des poids atomiques, Bruxelles, 1860 ; Recherches sur les lois des proportions
chimiques, etc., Bruxelles, 1865 ; J. J. Berzelius, Ldrbok i Kemien, Upsala, 1818 ; W. H. Wollaston,
Phil. Trans., 104. ], 1814 ; T. Thomson, An Attempt to establish the First Principles of Chemistry
by Experiment, London, 1825; H. Collins, Gfiem. News, 119. 247, 1919 ; M. Rudolphi, Chem. Ztg.,
25. 1133, 1901.
3 H. Erdmann, Zeit. anorg. Chem., 27. 127, 1901 ; T. W. Richards, ib., 28. 355, 1901 ; B.
Brauner, Ber., 22. 1106, 1889 ; 24. 256, 1897 ; 26. 186, 1901 ; L. Meyer and K. Seubert, ib., 18.
1089, 1885 ; W. A. Noyes, ib., 24. 523, 1891 ; W. Ostwald, Lehrbuch der allgemeinen Chemie,
Leipzig, 1. 43, 1891.
§ 11. The Relation between the Molecular Weights and the Volumes of
Gases
The theory of molecules is an ideal conception placed by the mind like another Atlas
underneath a measureless world of facts to give them intelligible cohesion and hold them up
to view.— S. Brown.
The molecular weight of any gas is numerically equal to the weight of any
volume of the gas when the weight of an equal volume of hydrogen under the same
physical conditions of temperature and pressure is 2. Two grams of hydrogen,
taken as the standard, occupy 22*3 to 22*4 litres at normal temperature— 0°— and
normal pressure— 760 mm. of mercury. Hence, it follows directly from Avogadro's
hypothesis that the molecular weight of any gas, expressed in grams, occupies
approximately 22*3 litres at 0° and 760 mm. pressure. Consequently, to find the
molecular weight of a gaseous substance, weigh 22*3 litres of the gas at a convenient
temperature and pressure ; calculate the corresponding volume at 0° and 760 mm.
pressure, and calculate by proportion the weight of 22' 3 litres.
Example.— A litre of gas at 20° and 730 mm. weighs 1-764 grams, what is the molecular
weight of the gas ? By the method of calculation indicated in the next chapter, one litre
of a gas at 20° and 730 mm. pressure contracts to 894-5 c.c. at 760 mm. and 0°. Hence,
if 894-5 c.c. weigh 1-764 grams, 223 litres will weigh 43-97 grams. Hence the molecular
weight of the gas is nearly 44.
It must here be mentioned that the number 22' 3 is not quite right for all gases.
Many gaseous molecules have a slight attraction for one another, so that the mole-
cules are slightly more closely packed than is represented by Avogadro's hypothesis.
The greater the intermolecular attraction, the greater the weight of 22*3 litres, and
202 INORGANIC AND THEORETICAL CHEMISTRY
consequently, the less the volume of a molecular weight of the gas expressed in
grams. Thus, experiment shows :
Hydrogen. Oxygen. Nitrogen. Chlorine. Hydr^X" divide. (0°.'760 mm.). ^^-^J^'
22-40 22-39 22*45 22-01 22-22 22-26 22-39 22-55
The deviation from 22*3 can be neglected in ordinary chemical calculations.
The molecular weight of a compound not only tells us a weight, but it also tells
us that if the molecular weight be expressed in grams, the substance when gaseous
will occupy 22-3 litres at 0° and 760 mm. Further, the molecular weight of a gas,
expressed in kilograms, occupies, approximately, 22*3 cubic metres at 0° and 760
mm. pressure. By mere chance, the number of avoirdupois ounces in a kilogram is
35"26, which is very nearly the same as the number of cubic feet in a cubic metre
(35' 31) — J. W. Richards.i The difierence is only one-seventh of 1 per cent. Hence,
the molecular weight of any gas, expressed in avoirdupois ounces, occupies,
approximately, 22*3 cubic feet at 0° and 760 mm. pressure. These factors are
useful in calculations involving cubic feet, cubic metres, and Htres.
References.
1 J. W. Richards, Journ. Franklin Inst., 152. 109, 1901.
§ 12. Chemical Equations and Chemical Arithmetic
In his calculations, the chemist relies on the supposed numerical relations of the invisible,
intangible, immeasurable particles he calls atoms. These relations have been determined
by others in whom he has confidence, and the accuracy of these relations has to be accepted
on faith.. — ^H. C. Bolton.
The molecular weight of an element or compound is the sum of the atomic
weight of each of the atoms of the constituent elements. — Let a molecule be com-
posed of rii atoms of one element, n2 atoms of another, n^ atoms of a third, and so on ;
further, let Ai, A^, A^, . . , denote the atomic weights of the respective elements,
then the molecular weight of the compound will be Wi^i+^2^2+^3^3+ • • •
For example, with the approximate atomic weights, the molecular weight of
hydrogen, H2, is 2 ; of water, H2O, 18 ; of sulphuric acid, H2SO4, 98 ; and of ferrous
ammonium sulphate, reS04.(NH4)2S04.6H20, 392 — since the summation furnishes
56+32+4xl6+2(14+4)+32+4xl6+6(2+16) = 392.
The process or art of calculating the numerical relations of the elements and their
compounds is sometimes called stoichiometry — from the Greek a-roix^Ta, a fundamental
constituent ; furptw, I measure. The term appears to have been devised by J. B. Richter,
in his book, Anfangagrunde der Stochyometrie oder Messkunst chymischer Elementes (Brealau,
1792-3), or, The rudiments of stoichiometry or the numerical relations of the chemical elements,
for that branch of chemistry which deals with the numerical proportions in which sub-
stances combine. To-day the term is sometimes extended to comprise molecular and
atomic weight determinations and also the general measurable properties of solids, liquids,
and gases ; solutions and mixtures ; etc. — witness, S. Young, Stoichiometry (London, 1908).
When the initial and final products of a chemical reaction as well as the com-
position and proportions of the molecules concerned in the reaction are known, the
facts can usually be symbolized or abbreviated into a kind of shorthand expression
which takes the form of a chemical equation. There are some limitations which
will be described later.
The equation indicates the nature of the different substances concerned in the
reaction ; as well as the proportions of the different substances which occur in
the initial and final products of the reaction. — For instance, when mercury is
heated in air and mercuric oxide, HgO, is formed, the reaction can be represented
in symbols : 2Hg+02=2HgO. We here ignore the nitrogen of the air because,
so far as we can tell, it plays no direct part in the chemical reaction. Similarly,
COMBINATION BY VOLUME 203
when mercuric oxide is heated to a high temperature, it decomposes, forming
metallic mercury and oxygen. In symbols, 2HgO=2Hg+02. The symbol = or
-> is used instead of the words " produces " or "forms," and the symbol + is used
for " together with " on the right side of the = sign, and for " reacts with " on the
left side. The latter equation reads : *' Two molecules of mercuric oxide, on decom-
position, produce a molecule of oxygen and two molecules of monatomic mercury.'*
The number and kind of the atoms of the two sides of the equation must always
be the same (persistence of weight).
The eauation indicates the proportions by weight of the substances concerned
in the reaction. — The atomic weight of mercury is 200, and the atomic weight of
oxygen is 16, hence, the molecular weight of mercuric oxide is 216, and of oxygen
32. The latter equation can therefore be read : "432 grams (ozs. or tons) of mer-
curic oxide in decomposing form 32 grams (ozs. or tons) of oxygen gas and 400
grams (ozs. or tons) of metallic mercury." Hence, the chemical equation can be
employed in all kinds of arithmetical problems dealing with weights of substances
formed or produced.
Examples.- — (1) How much mercuric oxide is required to furnish 20 grams of oxygen
gas ? Write down the proper equation ; write 432 below the mercuric oxide, and 32 below
the oxygen. We are not concerned with the mercury in this problem. Since we read from
the equation : 32 grams of oxygen are furnished by 432 grams of mercuric oxide, one gram
of oxygen will be furnished by 432-^32 = 13-5 grams of mercuric oxide; and 20 grams
of oxygen will come from 20 X 13*5=270 grams of mercuric oxide.
(2) Show that 2f grams of oxygen and 27J grams of mercury can be obtained theo-
retically from 30 grams of mercuric oxide. Obviously, 432 grams of mercuric oxide will
give 32 grams of oxygen, therefore 30 grams of mercuric oxide will give 2| grams of oxygen.
The equation indicates the proportion by volume of the gases concerned in the
reaction. — We have seen in the preceding section that if we express
-, , , .... Volume at 0" and 760 mnx,
Molecular weight in per molecular weight.
Grams . . . 22*3 litres
Kilograms 22-3 cubic metres
Ozs. (avoir.) 22*3 cubic feet
Consequently, the idea conveyed by the equation, 2HgO=02+2Hg, can be
expressed in these words : '' 432 grams (kilograms or ozs.) of mercuric oxide will
furnish 32 grams (kilograms or ozs.) of oxygen, or 22' 3 litres (cub. metres or cub. ft.)
of oxygen gas at 0° and 760 mm. and 400 grams of mercury."
Examples.— (1) What volume of oxvgen will be obtained by heating 30 grams of
mercuric oxide ? 432 grams of mercuric oxide wiU furnish 30 x22•3-^432 = l•55 htres of
oxygen gas at 0° and 760 mm. pressure. x rvo j
(2) How much mercuric oxide will be needed for 10 cub. ft. of oxygen gas at 0 Mid
760 mm. pressure ? Here 22-3 cub. ft. of the gas come from 432 ozs. of mercuric oxide,
hence, 432 X 10-^22-3 = 193 ozs., or 12 lbs. 1 oz. of mercuric oxide are required.
It wiU be observed that in these examples it has been assumed that the reactions
go to an end. This is an idealized imaginary condition which rarely obtams in practice
where other factors— temperature, concentration, unequal mixing, etc —introduce
disturbances. In practice, there are nearly always some losses, and the actual
vield is X per cent, of that theoretically possible on the assumption that the ideaUzea
equation is the limit or goal of perfection. In order to make sure that a reaction
will proceed to an end, y per cent, excess of the initial products may be reqmred.
Each reaction, in this respect, has its own specific character For example^ tne
formation of nitric acid, HNO3, by heating sulphuric acid H2SO4, with^s^um
nitrate, NaNOg, is represented by the equation: 2NaN03+H2bU4--liiNU3
+Na2S04, where 170 parts of sodium nitrate apparently require 98 parts ot sul-
phuric acid to produce 126 parts of nitric acid. The manufacturer, l^o^e^^'^^^
found by a process of trial and failure that under his conditions, an excess of about
80 more parts of sulphuric acid are needed to convert the 170 parts of sodium
204 INORGANIC AND THEORETICAL CHEMISTRY
»
nitrate into nitric acid. The equation would then be more correctly written :
2NaNO3+l-817H2SO4=2HNO3+Na2SO4+0-817H;iSO4,wheretheexcess0-817H2SO4
on both sides of the equation does not cancel out when the reaction is applied
under industrial conditions. This, however, makes no difference to the general
principles of chemical arithmetic here discussed. If the limitations of the stoichio-
metrical rules be not appreciated by the industrial chemist, his work will be
considerably hampered. In general, the rigid appHcation of fixed (scientific)
principles, without a due appreciation of their limitations, is disastrous in the
application of scientific methods in industrial work where success is estimated, not
by the profoundness of a theory, but by the results achieved, or dividends secured.
§ 13. The Relation between Atomic and Combining Weights — Valency
Die Valenz nur ein Ausdruck des Gesetzes der multiplen Proportionen i^.- — C. W.
Blomstrand (1869).
Each atom carries into its combinations two things : first, its own proper energy ; and
second, the faculty of expending this energy in its own way, in attaching other atoms to
itself, not indiscriminately, but definite atoms and in definite numbers.- — C. A. Wubtz
(1869).
Observation shows that the relative combining weights of oxygen and hydrogen
are very nearly as 0 : H=8 : 1 ; and that the atomic weights of oxygen and hydrogen,
deduced from the atomic theory and Avogadro's hypothesis, are very nearly as
0 : H=16 : 1. In fine, the atomic weight of oxygen is twice its combining weight.
For carbon in carbon dioxide, the combining weight is 3, while the atomic weight
of carbon is 12, that is, the atomic weight of carbon is four times the combining
weight. In the case of hydrogen and chlorine, the atomic and combining weights
are the same. In A. W. Hofmann's Introduction to Modern Chemistry (London,
1865), it is emphasized that the atomic weight of an element represents the
minimum quantity of an element which can take part in forming a molecule of
a compound ; the equivalent, or combining weight of an element, represents the
minimum quantity of an element which is required to fix one atom of hydrogen
taken as a standard ; and the valency or valence {valens, worth), or the atom-fixing
power of an element, represents the number of times the combining or equivalent
weight is contained in the atomic weight. In illustration,
Hydrogen. Chlorine. Oxygen. Nitrogen. Carbon.
Atomic weight . . 1 35*5 16 14 12
Combining weight . 1 35"5 8 4*67 3
Valency ... 1 1 2 3 4
Consequently, as a first approximation,
Atomic weight
: Valency.
Combining weight
Elements, however, may have more than one equivalent or combining weight, and
since the atomic weight remains constant, an element may have more than one
valency. Consequently, an atom not only has the power of fixing an atom of
another element, but, under definite conditions, it has a definite number of such
powers.
Although valency is primarily a number or a numerical ratio, the term is also
used to express a general characteristic of the elements. The valency o£ an element
(or radicle) represents the general property of an atom (or radicle) to combine
with a certain definite number o! other atoms (or radicles). In order to avoid
confusing valency a number with valency a property, some restrict the use of the
term so that valency is reserved for the property, and valence for the number ; thus,
mercury is an element with a valency of one or two, and in mercuric chloride, HgCl2,
COMBINATION BY VOLUME 205
mercury has a valence of 2, and in mercurous chloride, HgCI, a valence of one.
This suggestion is good when there is any risk of confusion.
The meaning of valency can be represented another way. Numerous observa-
tions indicate that there is generally a limit to the number of atoms which can unite
with a given atom, so that the atoms of an element appear to differ from one another
with respect to the number of other atoms with which they habitually combine ;
valency may then be regarded as representing a habit of an element for combination ;
it has nothing to do with the force holding the atoms together. The valence of an
element is obtained by finding — directly or indirectly — how many atoms of hydrogen
can combine with or be replaced by an atom of the given element. The valence of
hydrogen is always taken as unity. Hence the definition : The valence of an element
is a number which expresses how many atoms of hydrogen, or of other atoms
equivalent to hydrogen, can unite with one atom of the element in question.
Strictly speaking, valency is only applicable to those gases and liquids whose molecular
weights have been determined ; and it is extended to solids by analogy with gases.
We do not know the molecular weights of solids, and we therefore do not know if
the valency concept can be extended to solids ; it may possibly require modification.
Chemical affinity and valency are both pecuUar but essentially different pro-
perties of the atom, and they must not be confounded. The terms, however, are
sometimes used synonymously, since valency could not be manifested between two
elements which have no affinity for one another. Affinity refers to the act of
chemical combination ; valency governs the form of chemical combination. The
intensity of the chemical energy displayed by hydrogen, oxygen, nitrogen, and
carbon, in the act of combining with chlorine, is very different — chlorine unites with
hydrogen with great avidity ; with carbon the action is so sluggish that it requires
a powerful stimulant ; while, the union of chlorine with oxygen and nitrogen, is so
difficult that it can only be effected indirectly, not directly. On the other hand,
however vigorous the act of combination, the hydrogen atom is so constituted that
it can unite with only one atom of chlorine, while carbon can unite with four,
nitrogen with three, and oxygen with two. If the energy of the combination of
chlorine with these four elements be represented by the amount of heat, evolved
(+)or absorbed (— ) during the combination, the chemical affinity is approximately :
HCl.
OClg.
NCI3.
CCI4.
Chemical affinity .
, +22-0
-8-9
-12-8
+5-2 xinite
Valency
1
2
2
4
Stable.
Unstable.
Very unstable.
Stable.
A. S. Couper (1858), one of the pioneers in clarifying our ideas about valency, dis-
tinguished the two concepts by calling the former affinity of kind, and the latter
affinity of degree. Affinity of kind, said' he, is the specific affinities manifested bythe
elements the one for the other ; affinity of degree is the grades or limits of combina-
tion which the elements display.
According to the law of multiple proportions, the states of saturation of the
elements chsmge per saltu7n ; so also according to the doctrine of valency the affinities
of the elements are exhausted by stages. The two conceptions are not identical.
According to the latter, each element has a capacity for saturation which is
definite for a given combination, but which varies from element to element. In
1858, S. Cannizzaro explained the difference by comparing the two series of chlorides :
HgCl and HgClg ; CuCl and CuClg ; etc., and he added that the law of multiple
proportions asserts that the quantities of an element contained in different molecules
must be whole multiples of one and the same quantity ; but this law cannot foresee
that one atom of the element is equivalent in one case to one atom of hydrogen and
in the second case to two atoms of hydrogen.
Nomenclature.— With hydrogen and chlorine, the atomic and conibiniug weights
are the same, and the valency is unity. These elements are accordingly said to be
univalent, or monads ; for similar reasons, oxygen is bivalent, or a dyad ; nitrogen
is tervalent, or a triad ; carbon is quadrivalent, or a tetrad ; and so on to octovalent
206 INORGANIC AND THEORETICAL CHEMISTRY
elements or octads. The valency of an element is frequently represented by
attaching the necessary numbers, in dashes or Roman numerals, to the top right-
hand corner of the symbol for the element, as suggested by W. Odling in 1855.
Thus, the symbols ff and CP respectively mean that hydrogen and chlorine are
univalent ; 0" means that oxygen is bivalent ; N^" means that nitrogen is ter-
valent ; and C^ that carbon is quadrivalent. By collecting together a few com-
pounds with their symbols the idea can be made clearer.
)ivalent.
Bivalent.
Tervalent.
Quadrivalent.
Qnlnquevalent.
Sexivalent.
ffCP
Ba°0°
Hg^N^
H/C^^
pvFgi
S^^Fe^
Na^Cl^
Mo^Clg'
Fe^Clgi
c^^cv
W^Brgi
U^^Fgi
K^r
Zn^CV
Mo^Clgi
Mo^^CV
Mo^Cls^
Mo^^Fg^
Some heptads and octads are known. Hence, the valency of all known atoms can
be represented by an integer ranging from 0, 1, 2, ... to 8. The elements
generally combine in such a way that an equal number of valencies are opposed to
one another.
No chemical compound is known to be formed by the union of the elements
of the argon family, the so-called inert or noble gases. So far as our knowledge
goes, these gases have therefore a zero-valency, and the elements appear to be
non-valent. Any element existing free in a monatomic condition is non-valent in
the sense that its atoms are not united with others by means of valency bonds ;
but the two cases differ in that the maximum valency of the latter is n units, while
that of the inert gases is zero.
A few examples of radicles of different valency may be quoted : Monad radicles
—OH, CN (generally written " Cy "), NO3, NH4 (sometimes written "Am"),
COOH, etc. Dyad radicles — SO4, SO3, CO3, SiOa, etc. Triad radicles — PO4,
FeCye, etc. Tetrad radicles — FeCyg, Si04, etc. There are some important hydro-
carbon radicles — CH3, called methyl ; C2H5, ethyl ; C3H7, propyl ; C4H9, hutyl ;
C5H11, amyl ; etc. The members of the group of hydrocarbon 'radicles with the
general formula C^Hyw+i, are called the allcyl radicles. The members of the group,
C„Hn-i radicles — CgHs, phenyl; C6H4.CH2, benzyl, etc. — are called the aryl
radicles. There are also many other uni- and poly-valent hydrocarbon radicles.
Structural, graphic, or constitutional formulae. — The valency of an element is
sometimes represented by attaching the necessary number of hyphens to the symbol
for the element. This enables the molecules of a substance to be represented by
a kind of graphic formula. The symbol for hydrogen will have one hyphen ; oxygen,
two ; nitrogen, three ; carbon, four ; etc. ; a bivalent oxygen atom may be repre-
sented 0", —0—, 0=, 0<C, etc. The hyphens are usually attached so that the
graphic formula occupies as little space as possible ; they are drawn in the most
convenient direction. The atoms of a molecule are then supposed to be joined
together by their valencies ; and this is represented diagrammatically by hyphens.
The symbol for hydrogen chloride then becomes H— CI ; potassium iodide, K— I ;
water, H— 0— H ; mercuric oxide, Hg=0 ; a molecule of hydrogen, H— H ; a
molecule of oxygen, 0=0 ; carbon dioxide, 0=C=0 ; and
H-N<« o<^:z^ «>c<^
Ammonia. Ferric oxide. Methane.
Accordingly, the terms bonds or links are sometimes employed as well as valencies.
Graphic formulae are also called structural or constitutional formulae. Structural
formulcB primarily assume that the chemical properties of a substance are determined
by the arrangement of the atoms in the molecules ; and if the molecules of two compounds
of the same chemical composition have their atoms differently arranged, the properties
of the two compounds will be different. Graphic formulae are sometimes very con-
venient for representing the composition of compounds, but the student would err
COMBINATION BY VOLUME 207
rather seriously if he supposed that the symbol given above for, say, methane
represents the way the atoms are actually grouped in the molecule of methane.
This would involve a leap far beyond our real knowledge, although the available
evidence is in favour of the view that the atoms have a definite arrangement in the
molecule, and, in some cases, the little knowledge we do possess can be better
summarized by a graphic formula than in any other way. The graphic formula
furnishes a clearer mental image of the curious way certain groups of atoms remain
clustered together through a complex series of chemical changes than if the reaction
were represented by ordinary symbols. The structural formula has a real and
important signification ; it should symbolize the chemical character of the molecule.
A graphic formula is thus a kind of dummy model illustrating the way a compound
is formed, how it decomposes, and the relations between one compound and another.
Indeed, chemists now investigate the position of a particular atom in a chain or
ring of atoms, and find it to be at the side, in the middle, or in some other position
relative to the remaining atoms. Without accepting C. F. Gerhardt's contention
(1856) that lesformules chimiques nesont pas destinees a representerV arrangement des
atomes,^ it must not be believed for one moment that the model simulates reality,
since, for one thing, the formulae are built on a plane two-dimensional surface,
whereas the molecule probably extends into three dimensions ; again, graphic
formulae make the molecule appear as a fixed rigid structure, whereas there is some
evidence indicating that the atoms within the molecule are in ceaseless rhythmic
motion. The remarkable work which has been done by the aid of structural formulae
will always justify their use in the past and present, whatever future generations
may think of them. The wonderful development of organic chemistry, said J. U.
Nef (1904), is a consequence of the simple valency concept.
The doctrine of valency has furnished the chemist with a basis for calculation,
and enabled him to deduce algebraically the existence of series of compounds previ-
ously unknown. It has been said that the theory of valency has enabled the chemist
to predict reactions of unknown compounds with other known compounds, and
enabled him to found a mechanics of the atoms which in another direction is as
wonderful as the mechanics of the astronomer which has enabled him to fix the
position and path of an in\asible planet from its effect on the movements of one
visible and known. Although the theoretical limitation seems valid in the majority
of cases, yet there are several compounds whose existence appears contrary
to the valency hypothesis — e.g. nitric oxide. However, where investigation is
guided by a wrong theory, only those things which are sought are likely to be found,
and the theoretical limitation may not have any real counterpart in nature. Hence,
A. Gr. V. Harcourt 2 could say :
A chemist who should depart from the general course, and set himself to prepare
substances whose existence is not indicated by theory, would perhaps obtain results of more
than usual interest.
Maximum and active valency. — Most elements have more than one valency.
Stannous oxide has a composition corresponding with SnO ; and stannic oxide,
with Sn02. In the former case, the tin is said to be bivalent ; and in the latter,
quadrivalent. There are thus two series of tin compounds — stannous and stannic.
Similarly with copper, iron, etc. There are also two carbon oxides, carbon monoxide,
CO, and carbon dioxide, CO2. If carbon monoxide could be written 0=C=C=0,
and there is nothing in the analysis by weight which prevents this, all might be
well ; but writing the formula in this manner would involve a contradiction of
Avogadro's hypothesis, since the vapour density of carbon monoxide corresponds
with the molecule CO, not C2O2. We cannot see the way clear to admit carbon
monoxide as an exception to Avogadro's hypothesis, for that would introduce
confusion into our system, and there would be no immediate prospect of restoring
order. Some get over the difficulty by assuming that two of the free valencies
in carbon monoxide mutually saturate one another, and write the graphic formula
208 INORGANIC AND THEORETICAL CHEMISTRY
0=C: ; others assume that oxygen is quadrivalent, and write the graphic formula
for carbon monoxide C=0 ; and for carbon dioxide, C<k, the two oxygen
atoms are supposed to be doubly linked to one another and to the carbon atom.
The question is therefore somewhat involved. The case of sulphur bivalent in
hydrogen sulphide, H— S— H ; quadrivalent in sulphur dioxide 0=S=0 ; and
sexivalent in sulphur trioxide 02=S=0, fits very well into this scheme. So do
the series of compounds represented by ethane, C2Hg ; ethylene, C2H4 ; and acetylene,
C2H2, which can be respectively represented by the graphic formulae :
H^C-C^H Hv^jj^jj^H H— teC— H
h/ \h H H
Ethane (with single bonds) Ethylene (with double bonds) Acetylene (with triple bonds)
provided it be assumed that the respective carbon atoms are joined by single, double,
and triple bonds. It may be added that the circumstantial evidence advanced by
organic chemistry strongly favours this assumption.
Since chlorine or fluorine forms combinations with the metals far more generally
than does hydrogen, it has been proposed to use chlorine or fluorine in place of
hydrogen as the standard of valency. The hydrogen and fluorine valencies, however,
are not always the same. . For instance :
Hydrides . . LiH CaHg (BH3)2 CH4 PH3 SHg IH —
Fluorides . . LiF CaFg BF3 CF4 PF5 SFg IF OsFg
The maximum valency of the hydrides is thus attained with the tetrads ; but with
fluorides, the maximum valency is reached with the octads. The preceding defini-
tion of valency is troublesome if applied to azomide,HN3, although it works all right
with ammonia, NH3.
F. A. Kekule (1866) ^ argued that valency is a fundamental property of the
atom which is just as constant and invariable as the atomic weight ; the equivalent
weight of an element may vary, the valency cannot. E. Frankland (1852) showed
that the elements of the nitrogen family are sometimes ter- and sometimes quinque-
valent. A controversy whether valency is fixed or variable was carried on about
1864 by F. A. Kekule, C. A. Wurtz, A. Naquet, H. Kolbe, and A. W. Williamson.
The controversy, after all, turned out to be nothing more than ein Streit um ein
Wort. If valency means maximum saturation capacity, this property is unchange-
able, but if valency means that this maximum power is always exerted, and that
every atom exerts a constant invariable valency, the doctrine est en desaccord
Jlagrant avec les fails. The discussion was then diverted to atomic and molecular
compounds (q.v.). Each element has a maximum valency towards certain other
elements. When an element appears to have a lower valency than its maximum
valency, the compound is said to be an unsaturated compound, in contrast with
a saturated compound in which the atoms are exercising their maximum valency.
In many unsaturated compounds, the valencies appear to diminish in pairs. The
pairs of dormant or sleeping valencieSy crypto-valencies {k pv-n-ro^, hidden), or latent or ^05-
sive valencies are supposed to be self -saturated. Hence W. Odling (1855) proposed to
call elements with an odd number of bonds j)erissads (Trtpto-o-o's, odd), and those with
an even number of bonds artiads (aprto?, even). It was also assumed that the sum
of the valencies of the atoms forming a molecule is always an even number.
As a matter of fact, the hypothesis of the self-saturation of the bonds in pairs
breaks down completely. The idea probably arose from the application of an in-
accurate hypothesis — started in 1864 by E. Erlenmcyer * — which is stated in some
of the older books on chemistry in words like these : *' All chemical evidence shows
that a body with unsatisfied bonds cannot exist by itself." All chemical evidence,
as we shall see, shows nothing of the kind. Mercury and many other elements,
COMBINATION BY VOLUME
209
when vaporized, give gases with one-atom molecules. The principle of self-satura-
tion breaks down when applied to the nitrogen oxides, say nitric oxide, N'°0".
The relative density (Avogadro's hypothesis) will not let us write N2O2, that is,
0==N— N=0. We are therefore confronted with what appears to be an odd
unsaturated valency in the molecule— N=0. Again, chlorine forms chlorine
monoxide, CI2O, and chlorine peroxide, CIO2 ; indium forms the three chlorides,
InCl, InCl2, InCls- The original form of the doctrine of valency is not tenable ;
elements cannot be classed as invariably uni-, bi-, ter-, quadri-, . . . valent, nor
as artiads and perissads, since some elements can have any of these valencies accord-
ing to circumstances. Chlorine, nitrogen, ruthenium, and manganese can be cited
as examples ; again, molybdenum forms a series of compounds with univalent
chlorine or fluorine — M0CI2, M0CI3, M0CI4, M0CI5, and MoFg ; and vanadium forms
VCI2, VCI3, VCI4, and VCI5. In view of facts like these, it is difficult to maintain
the thesis that the apparent inconstancy of the valency of an element is due to
the mutual saturation of pairs of valencies. Either a molecule can exist with free
valencies, or Kekul6's maximum valency hypothesis breaks down when confronted
with facts.
A great many ingenious hypotheses, more or less satisfactory, have been sug-
gested to explain the difficulties. At present we are compelled to frankly admit
with W. Lessen (1880) and A. Claus (1881) that the active valency of an element
is a variable habit of combination. An explanation of the meaning of valency is
thus left open. C. A. Wurtz (1864) distinguished between what he called atomicite
actuelle and atomicite virtiielle, and in order to distinguish between the greatest
valency an element is known to exhibit, and the valency which actually prevails
in a particular compound, the terms maximum or absolute valency and active
or actual or free valency may be respectively employed. So far as we can see, the
active valency of an element is dependent upon the properties of the atoms of the
other elements with which it is combined as well as on the prevailing physical and
chemical conditions to which the element is exposed. Thus sulphur is bivalent
towards hydrogen, but it can be sexi valent with fluorine ; antimony, arsenic, and
phosphorus are tervalent towards hydrogen, while phosphorus and antimony may
be quinquevalent towards chlorine ; arsenic is tervalent towards chlorine — and
there is some doubt if the pentachloride, ASCI5, has been made.
Werner's Nomenclature. — With a complex series of salts, instead of representing
the number of times the acidic radicle is contained in the molecule — e.g. CuCl,
copper monochloride ; CUCI2, copper dichloride ; CuO, copper monoxide ; PtCl4,
platinum tetrachloride, etc. — it is simpler, according to A. Werner,^ to represent
compounds with the same valency by names ending in the same suffix or letter.
Thus, if M represents an atom of a basic element, and X an atom of acidic univalent
element,
Table VII.— A. Werner's Nomenclature of Salts.
Com-
Valency.
Tennina-
tion.
Examples.
pound.
Werner's name.
Old name.
MX
MX 2
MX3
MX4
MX 5
MXfl
MX;
MX3
uni-
bi-
ter-
quadri-
quinqvie-
sexi-
septi-
octi-
a
0
i
e
an
on
in
en
CuCl — cupraohloride
CuCla— cuprochloride
M0CI3 — molybdenichloride
M0CI4 — molybdenechloride
MoCl 5— moly bdanchloride
MoFg^ — molybdonfluoride
CI2O y — chlorinoxide
OsO 4 — osmium enoxide
copper monochloride
copper dichloride
molybdenum trichloride
molybdenum tetrachloride
molybdenmn pentafluoride
molybdenum hexafluoride
chlorine heptoxide
osmium tetroxide
The suffixes have been chosen to make them differ as little as possible from
those already in existence. The only serious objection appears to arise with salts
VOL. I. ^
210 INORGANIC AND THEORETICAL CHEMISTRY
like univalent and bivalent copper, mercury, etc., of the type CuCl, CuCl2 ; and
HgCl, HgCl2, where cuprous becomes cupra-, and cupric, cupro- ; and mercurous
becomes mercura-, and mercuric, mercuro-.
The effect of external conditions on the valency of an element.— Active valency
has been compared with friction in so far as it appears to be called into play by
external causes which may vary from zero upwards, because the valency of an ele-
ment is determined by the physical and chemical conditions under which the element
is placed. For instance,
(1) Temperature. — The valency of an element generally diminishes with rise of
temperature, e.g. sulphur trioxide, SO3, when heated dissociates into sulphur dioxide,
SO2, and oxygen ; and carbon dioxide, CO2, into carbon monoxide, CO, and oxygen.
Copper oxide, CuO, at 1110° becomes cuprous oxide, CU2O ; and lead dioxide,
Pb02, at 615° yields lead monoxide, PbO.
(2) Pressure. — The valency of an element is often diminished with a decrease
of pressure. Pressure usually facilitates chemical action. By heating bismuth with
water at 280° under a pressure of 10,000 atm. the monoxide, BiO, is formed, but
at higher temperatures and less pressure the sesquioxide, Bi203, is produced ;
similarly antimony is said to form the monoxide, SbO, and aluminium the
monoxide, AlO, under conditions where the sesquioxides would normally be
produced. Carbon monoxide, CO, under a pressure of 600 atm. at 320° is partially
converted into the dioxide, CO2, and free carbon.
(3) Light or radiant energy. — Numerous physical and chemical changes are
induced by exposure to light, and the reactions may be accompanied by changes
in the valency of some of the elements concerned. Thus, by exposure to light
ferric oxalate, Ee2"^(C204)3", is reduced to ferrous oxalate, Fe^^C204 — in symbols :
Fe2(C204)3=2FeC2044-2C02 ; and an aqueous solution of mercuric chloride,
HgCl2, is reduced to mercurous chloride, HgCl, under similar conditions : 4HgCl2
+2H20=4HClH-02+4HgCl. Similar remarks, mutatis mutandis, apply to the
effect of other forms of radiant energy.
(4) Chemical reagents. — Changes in the valency of an element are usually induced
by oxidizing or reducing agents. Thus, ferrous chloride, FeCl2, is oxidized to ferric
chloride, FeCls, by the action of hypochlorous acid, HCIO ; the reaction is symbol-
ized : 2Fe"Cl2+HCl+HC10=2Fe"^C]3+H20 ; and ferric chloride is reduced
to ferrous chloride by the action of sulphur dioxide, 2Fe^"Cl3+S02+H20=2Fe"Cl2
+2HCI+SO3. At the same time, it will be noticed, the sulphur dioxide, 0=S=0
0 • '
is oxidized to sidphur trioxide, 0=S<^^ where quadrivalent sulphur probably
becomes sexivalent. Hence, oxidation usually involves an increase in the valency
of an element, and reduction a decrease.
References.
^ C, F. Gerhardt, Traiti de chimie organique, Paris, 1856.
2 A. G. V. Harcourt, B. A. Bep., 36, 1875 ; J. U. Nef, Journ. Amer. Chem. Soc, 26. 1549,
1904 ; F. A. Kekule, Liebig's Ann., 106. 129, 1858 ; S. Cannizzaro, Nuovo Cimento, 8. 71, 1858 ;
A. S. Ck)uper, Compt. Bend., 46. 1157, 1858 ; Phil. Mag., (4), 16. 104, 1858.
3 F. A. Kekule, Liebig's Ann., 104. 129, 1857 ; 106. 129, 1858 ; 117. 120, 1861 ; 137. 74,
1866 ; Zeit. Chem., 7. 689, 1864 ; E. Frankland, Phil. Trans., 142. 417, 1852 ; Liebig's Ann., 85.
329, 1853 ; C. A. Wurtz, Ann. Chim. Phya., (6), 43. 492, 1885 ; Compt. Bend., 43. 199, 1856 ;
The Atomic Theory, London, 1880; Lecons de philosophie chimique, Paris, 1864; H. Kolbe,
Liebig's Ann., 113. 293, 1860 ; 101. 257,' 1857 ; A. W. Williamson, Phil. Mag., (3), 37. 350,
1850; Journ. Chem. Soc, 4. 350, 1852 ; W. Odling, ib., 7. 1, 1855; A. S. Couper, Compt. Bend.,
46. 1157, 1858 ; Phil. Mag., (4), 16. 104, 1858 ; A. Naqiiet, Zeit. Chem., 7. 679, 1864.
4 E. Erlenmeyer, Zeit. Chem., 6. 65, 97, 609, 1863 ; 7. 1, 72, 628, 1864 ; W. Lossen, Lielig's
Ann., 204. 336, 1880 ; Ber., 20. 3306, 1887 ; 14, 760, 1881 ; A. Glaus, Ber., 14. 432, 1881 ; A.
Wurtz, LcQons de philosophie chimique, Paris, 1864,
^ A. Werner, Neuere Anschauungen auf dem Gebiete der anorganischen Chemie, Braunschweig,
13, 1905 ; London, 75, 1911 ; B. Brauner, Zeit. anorg. Chem., 32, 10, 1902.
COMBINATION BY VOLUME 211
§ 14. The Polarity o! Valency
The doctrine that the chemical forces by which the elements of bodies are held together
or separated, are identical with the polar forces of electricity is now entirely established in
the minds of the most profound and philosophical chemists of our time. — W. Whewell.
An agent exhibits polarity when it is characterized not only by a numerical
value, but also by a sign indicating the direction in which it will act. For example,
during the electrolysis of binary compounds some elements always accumulate at
one particular electrode ; the hydrogen, for instance, goes to the cathode, never to
the anode ; and conversely, the oxygen goes to the anode, not to the cathode. It
is therefore assumed that hydrogen carries a positive electrical charge, oxygen a
negative charge ; otherwise expressed, oxygen has a negative polarity, hydrogen
a positive polarity.
In 1881, in a paper On the modern development of Faraday's conception of elec-
tricity, H. von Helmholtz deduced from Faraday's work that during electrolysis
the same quantity of either positive or negative electricity (96,540 coulombs)
always accompanies each univalent atom, or each valency of a multivalent element,
so that the same quantity of electricity passing through an electrolyte always sets
free or transfers the same number of units of affinity (or valency) at each electrode.
Otherwise expressed, an w-valent atom or radicle carries n unit charges of electricity.
Electricity thus behaves as if it were divisible into definite elementary portions —
positive or negative — which behave as if there were atoms of electricity. Following
G. J. Stoney's proposal (1881), these unit or atomic changes of electricity are called
electrons.^ It may therefore be said that valency is a polar phenomenon, each
valency being associated with a positive or negative electron. The valency of a uni-
valent hydrogen atom carrying a positive charge can therefore be called a positive
valency, and each valency of a bivalent oxygen atom carrying two negative charges
a negative valency. Each positive valency can be represented by a + or • sign
attached to the symbol of the element, say H"" or H" ; and, in a similar manner,
each negative valency represented by a — or ' sign, say 0 or 0". These symbols
properly interpreted represent observed facts.
Again, during the electrolysis of certain compounds, some elements — arsenic,
antimony, boron, bromine, carbon, iodine, nitrogen, phosphorus, selenium, silicon,
sulphur, tellurium, etc. — act sometimes like hydrogen and sometimes like oxygen
in that with some compounds a given element may accumulate at the positive pole
and with other compounds at the negative pole. Otherwise expressed, the atoms
of these elements sometimes carry positive and sometimes negative charges, so that
in some compounds the atoms of these elements have positive valencies, and at other
times negative valencies. R. Abegg (1904) called these elements with a dual nature
amphoteric elements {aficfyL, both). Hence, a description of the valency of an element
in a particular compound should indicate whether the active valency is positive or
negative. In further illustration, the sulphur in hydrogen sulphide, H2S, has two
negative valencies ; and in sulphur trioxide, SO3, the same element has six positive
valencies, so that a change from sulphur with —2 valencies to sulphur with -}-6
valencies involves a change of eight units of electricity— the algebraic difference 8,
not the numerical difference 4 units. Similarly, in methane, CH4, the carbon atom
has four negative valencies ; in carbon tetrafluoride, CF4, the carbon atom has four
positive valencies ; so that the passage from the former to the latter again involves
a change of eight valency units. To avoid confusion with valency as a number, the
term polar number has been employed to represent the algebraic number of negative
charges which are lost, or positive charges gained by an atom of an element m the
formation of a given compound. The valency and polar number of nitrogen m
ammonia are 3 and -3 respectively ; the valency of nitrogen in ammonium chloride
is 5 and the polar number -3 {i.e. -4+1), as illustrated in the diagram. Fig. 4.
In nitrous acid, HO.N : 0, with oxygen negative, the polar number is +3, and the
valency 3 ; whereas in H-N=02, the polar number is still +3, but the valency
212
INORGANIC AND THEORETICAL CHEMISTRY
is 5. In potassium permanganate, KO.MnOa, the polar number of manganese is
+7, and the valency 7.^
D. I. Mendel^efE (lS71) assumed that the highest oxide (omitting the peroxide)
gives the maximum valency of an element,
and R. Abegg (1904) adopted practically
the same suggestion for finding the maxi-
mum positive valency of an element ; J. N.
Friend (1908) suggested that the fluoride
be employed for the same purpose. The
hydrides usually give the numerical values
of the negative valency of the non-
metals. J. N. Friend (1908) has com-
piled the following Table VIII showing the
positive and negative values of some amphoteric elements with respect to their
hydrides and fluorides :
NH4CI
Fig. 2. — Polar Numbers of Nitrogen (—3).
Table VIII. — Hydrides and Fluorides of Some Amphoteric Elements.
Negative valency.
Positive valency.
Total
Hydride.
Active
valency.
Fluoride.
Active
valency.
valency.
Antimony
Arsenic
Boron
Bromine .
Carbon
Iodine
Nitrogen .
Phosphorus
Seleniima .
Silicon
Sulphm* .
Tellm-ium
SbHg
AsHg
BH3
BrH
CH.
IH
NH3
SeHg
SiH^:
SH2 :
TeHg
3
3
3
1
4
1
3
3
2
4
2
2
SbF,
AsF^
BrF3
CF,
IF5
NOF
SeFe
SiF,
SFe
TeFe
5
5
3
3
4
5
3
5
6
4
6
6
8
8
fi
4
8
6
6
8
8
8
8
8
It is interesting to note that the majority of the known amphoteric elements
give 8 as the sum of the positive and negative valencies. K. Abegg (1904), indeed,
assumed that all elements are amphoteric and possess 8 positive and negative
valencies, but the observed facts with hydrogen, the alkali metals, and the inert gases
do not favour this generalization. The positive valencies of the alkali metals appear
to be so strong that they show little or no sign of their supposed negative valencies ;
and the negative valencies of fluorine are so strong that they show little or no sign
of positive valencies. E,. Abegg and G. Bodlander (1899) developed the hypothesis
that elements have a different A^alency according as they are united with electro-
positive or electronegative elements ; and that each element possesses the two
kinds of valency — positive and negative. The usually accepted valencies of the
non-metals are negative, and of the metals, positive ; R. Abegg and G. Bodlander
called these the normal valencies of the elements ; and the secondary valencies of
opposite polarity, active only under special conditions, were called contra- valencies.
The normal valencies are supposed to be the stronger. The sum of the normal and
contra-valencies, as indicated, is assumed to be 8, ranging over the different
families of elements :
Normal valencies
Contravalencies .
Polar number
Metals.
Na Mg Al
+1 +2 +3
—7 —6 —5
-6 -4 -2
Non-metals.
Si P S CI
-f4 —3 —2 — 1
—4 -f5 +6 +7
0 4-2 +4 +6
COMBINATION BY VOLUME 213
Thus chlorine is univalent, polar number —1 in hydrogen chloride, HCl, where it
is coupled with electropositive hydrogen ; but it has its maximum heptavalency,
polar number +7, when it is united with electronegative oxygen in chlorine hept-
oxide, CI2O7. In particular cases, neither all the normal nor all the contra-valencies
may be active. The contra-valencies in a particular family of elements increase in
activity as the atomic weights of the elements increase ; thus, in the halogen
family, fluorine (atomic weight 19) does not form a compound with oxygen, while
iodine (atomic weight 127) gives a stable oxide. All the normal valencies of an
element are supposed to be equivalent,* but if one be saturated, the remainder are
weakened. Consequently, the active valency of an element depends upon the
electrochemical character of the associated atoms — arsenic pentafluoride, A8F5,
for instance, is fairly stable (0. Euif and H. Graf, 1906), while arsenic pentachloride,
AsCls, is so very unstable that it is doubtful if it really has been prepared (C.
Baskerville and H. H. Bennett, 1902). The formation of the so-called molecular
compounds by the union of two or more molecules is attributed to the presence of
unsaturated, contra, or secondary valencies in at least one of the constituent atoms.
There are some modifications of this theory of valency. Most are agreed about this
interpretation of positive and negative valencies ; and the formation of double and associated
compounds is supposed to be due to the exercise of residual, contra or secondary valencies.
L. Spiegel (1902) assumed that elements possess secondary valencies which can be called
forth only in pairs of equal and opposite sign, so that when not externally saturated they
neutralize one another and impart no electro-chemical characters to the element. Spiegel
called these extra-valencies, neutral affinities. S. Arrhenius (1904) made a similar assumption
and called them electrical double valencies, and J. N. Friend (1908) used a similar hypothesis
and called the sleeping valencies, residual or laient valencies. I. Langmuir (1916) assimied
that the aggregation of molecules into liquid and solid masses is due to the exercise of the
secondary valencies, and thus the cohesion of solids and liquids is due to the exercise of an
attraction similar in kind to chemical affinity. The electron hypothesis will be described later.
References.
1 G. J. Stoney, Phil. Mag., (5), 11. 381, 1881 ; Proc. Dublin Soc., 3. 51, 1883; H. von Hebn-
holtz, Journ. Chem. Soc., 39.' 277, 1881.
2 R. Abegg, Zeit. anorg. Chem., 39. 330, 1904 ; R. Abegg and G. Bodlander, t6., 20. 453,
1899 ; L. Spiegel, ib., 29. 365, 1902 ; D. I. Mendeleeff, Journ. Russian Phys. Chem. Soc., 1. 1, 1869 ;
N. Morozoff, ib., 38. 481, 1906 ; J. N. Friend, Journ. Chem. Soc., 93. 260, 1908 ; W. Ramsay, ib.,
93. 778, 1908 ; 0. Ruff and H. Graf, Ber., 39. 67, 1906 ; S. Arrhenius, Theorien der Chemie,
Leipzig, 1906 ; C. Baskerville and H. H. Bennett, Journ. Amer. Chem. Soc., 24. 1070, 1902 ;
I. Langmuir, ib., 38. 1145, 2221, 1916 ; H. E. Armstrong, Phil. Mag., (5), 25. 21, 1888.
§ 15. The Association of Atoms in Three Dimensions
The arrangement of the atoms of a molecule in one plane is equally convenient in
diagrams, and improbable as a natural fact.— A. G. Vernon Harcourt (1875).
When our views are sufficiently extended as to enable us to reason with precision con-
cerning the proportions of elemental atoms, we shall find the arithmetical relation will not
be sufficient to explain their mutual action and we shall be obliged to acquire a geometrical
conception of their relative arrangement in all three dimensions of solid extension. . . .
When the number of particles (combined with one particle) exists in the proportion of
4:1, stable equilibrium may take place if the four particles are situated at the angles of
the four equilateral triangles composing a regular tetrahedron. ... It is perhaps too
much to hope that the geometrical arrangement of primary particles will ever be perfectly
known.— W. H. Wollaston (1808).
In order to explain why the atoms of diatomic molecules travel about in pairs,
it seems to be necessary to assume that the atoms exert an attraction on one another,
and that the position of the atoms in space must be conditioned by the attractive
forces. As Isaac Newton said in his OpticJcs (London, 1704) :
How the particles which touch only in a few points can stick together and that so firmly
as they do, without the assistance of something which causes them to be attracted or
pressed towards one another, is very difficult to conceive.
2U INORGANIC AND THEORETICAL CHEMISTRY
When two univalent atoms unite with one bivalent atom, it is natural to imagine
two points of contact, and two directions in which the bivalent atom exerts its
power of combination. This conception of direction appears to be almost necessary
in the case of carbon with its four valencies, and organic chemists have founded
upon this what is known as stereochemistry {a-rep^os, solid), or chemistry in three
dimensions, or chemistry in space, on lines dimly foreshadowed by W. H. Wollaston
in 1808, and A. M. Ampere in 1814. Since then, many chemists have thrown out
hints of a tridimensional arrangement of the atoms in a molecule — L. Pasteur (1861),
F. A. Kekule (1861), A. M. ButlerofE (1863), E. Paterno (1869), A. Gaudin (1873), etc.
Thus, in his celebrated lecture, Recherches sur la dissymetrie inoleculaire des produits
organiques (Paris, 1861), L. Pasteur asked : Are the atoms of (Z-tartaric acid grouped
on the spiral of a helix winding to the right, or placed at the summits of an irregular
tetrahedron, or disposed according to some other asymmetric grouping ? and
replied : We cannot answer these questions. It was not until the appearance of
J. H. van't Hoff's paper. On a system of atomic formulce in three dimension.'^, in
Holland, September, 1874 ; ^ and J. A. le Bel's Stir la relations qui existent entre les
formules atomique^, in France, November, 1874, that this idea was systematically
developed as a working hypothesis in organic chemistry. After demonstrating the
probability of the hypothesis that the carbon atom exerts its valencies in definite
directions in tridimensional space, it appeared highly probable that other elements
would be found to exhibit the same phenomenon, and thus arose a stereochemistry
Fig. 3. — Diagrammatic representation of the Tetrahedron Theory of Quadrivalent Carbon with
Single-, Double-, and Triple-linked Carbon Atoms.
of nitrogen, sulphur, silicon, selenium, tin, etc. The relative directions of the four
valencies of the carbon atom have been studied, and the attempt has been made to
find the effect of the displacement of these directions upon the properties of the re-
sulting compounds. It appears to be necessary to assume that the carbon atom is
a material body with a certain shape and size, because K. Auwers (1890) has shown
that in the case of two carbon atoms united by a double-bond, the linking forces
probably act in such a way as to make an angle with each other and not a straight
line joining the two points, because the existence of such forces acting from mere
point-centres is highly improbable. Without making any suggestion as to the
actual form of the tetrahedral arrangement of the valencies of the carbon atom —
whether the attractive forces are concentrated at the apices (J. Wislicenus, 1888),
or at the centres of the faces (A. Wunderlich, 1886) — organic chemists, following
Wollaston's suggestion, find it convenient to represent graphically the four valencies
of the carbon as acting in the direction of the line joining the centre with the
apices of a regular tetrahedron. According to this hypothesis, the constitution of
methane, CH4, will be that represented in the diagram. Fig. 5, where the circles
represent the relative positions of the hydrogen atoms with respect to the central
carbon atom ; similarly, for ethane, C2H6, with a pair of single-linked carbon atoms.
Fig. 5, acetylene, C2H2, with a pair of triple-hnked carbon atoms. Fig. 5 ; and benzene,
CqHq, with a chain of six carbon atoms alternately single- and double-linked so as to
form a closed chain or ring.
In the case of double- or triple-linked carbon atoms, are the lines assumed to
COMBINATION BY VOLUME 215
be normally directed from the centre of the tetrahedron, bent with or without
straining, or do the forces act rigidly in one fixed direction so that their com-
ponents alone act in a direction parallel with the line joining the centres of the two
tetrahedra ? If it be assumed, with A. Naumann (1890), that the two valencies
joining a pair of double-linked carbon atoms in, say, ethylene, C2H4, are directed
from the centre of a tetrahedron towards the apices, and if each of these forces be
resolved in two directions according to the parallelogram of forces, the sum of the
components of each of these forces acting in the direction of the line joining the
centres of the two tetrahedra, is effective in holding the two carbon atoms together.
If the force with two single-linked carbon atoms be taken as unity, the force holding
a pair of double-linked atoms will be 0'577 X 2, and between a pair of triple-linked
carbon atoms, 0'33x3. This is not in agreement with J. Thomsen's thermal data.
A. von Baeyer, in a paper Ueher Polyacetyleneverhindungen (1885), showed that
if the four valencies of carbon are directed from a centre towards the four corners
of a regular tetrahedron, the lines must make an angle of 109° 28' with one another ;
and he made the assumption that if the direction of the attraction be diverted,
there will be a corresponding strain ; the greater the divergence, the greater the
strain ; and the greater the strain, the less the stability of the resulting molecule.
The negative heat of formation of acetylene with its two carbon atoms connected
by a triple bond, and the great instability of the acetylene compounds, show that
the three linldng bonds of the two acetylene carbons may be under some such strain ;
otherwise it might be anticipated that a pair of triple-linked atoms would be more
stable than a pair of double-linked atoms, and the latter in turn more stable than a
pair of single-linked carbon atoms. J. Thomsen's study of the heats of formation
of the hydrocarbons (1882) shows that the breaking up of a double-bond requires
15*46 Cals. less thermal energy than a pair of single-bonds, and the breaking of a
triple -bond requires 43 '92 Cals. less thermal energy than is needed for three single-
bonds.
Consequently, A. von Baeyer's strain theory of valency — Spannungstheorie — •
assumes that the four valencies of the carbon atom normally act in the direction
of the lines joining the centre with the apices of a regular tetrahedron making angles
109° 28' with one another ; and if these directions be bent or diverted, the lines are
strained as if they were elastic wires, so that the greater the divergence the greater
the strain, and the less the stability of the molecule. It follows that if the carbon
atoms all lie in one plane, the angles of divergence with ethylene and with tri-,
tetra-, penta-, and hexamethylene, CnH-m, wiU ^^
CHg HgC^CHg
^^^ ^^*^ II2C CHg n/\nTr HoC/ NcHo
CHg HgC CH2 HgC CH2 HgC CH2 H2C CHg
(C2H4), 54° 44' (C3H6) 24" 44' (C4H8), 9« 44' (CsHiq), 0** 44', (CgHig), —5' 16'
Ethylene Trimethylene Tetramethylene Pentamethylene Hexamethylene
and generally, for a ring compound of this type containing n carbon atoms in the
ring, the angle of divergence will be 54° 44' less (w— 2) 90° -^-w. H. Sachse introduced
further developments.
This hypothesis explains how the members of the closed ring series increase in
stability up to a maximum with pentamethylene, which should be more stable than
all the other members of the series, for the higher members decrease in stability with
increasing complexity ; the theory also explains how organic compounds with open
chains have a greater tendency to form closed rings with five and six members than
closed rings of greater or less complexity. F. Stohmann and C. Kleber's measure-
ments (1892) of the energy required to break such rings and add two hydrogen
atoms are in approximate agreement with this deduction ; so also is I. Traube's
work (1899) on atomic volumes. There are, however, several series of compounds
whose behaviour does not fit in quite so well with the hypothesis. For instance,
216 INORGANIC AND THEORETICAL CHEMISTRY
it will be obvious that the strain theory itself cannot be a sufficient explanation of
ring formation because it does not take the influence of chemical affinity into account
— e.g. the influence of side-chains in facilitating the closing of the ring. H. N.
Stokes (1900) applied a similar hypothesis to the phosphimic acids in which the
phosphorus atoms form closed rings and the results were in general agreement with
the hypothesis.
References.
1 J. H. van't Hoff, La chimie dans Vespace, Rotterdam, 1875 ; Bull. Soc. Chim., 24. 295, 338,
1875; J. A. le Bel, ib., 22. 337, 1874; M. Berthelot, ib., 24. 338, 1875; F. W. Clarke, Amer.
Chemist, 6. 81, 1875 ; L. Pasteur, Recherches sur la dissymetrie moUculaire des produits organiques
naturels, Paris, 1861 ; Alembic Club Reprints, 14, 1897 ; F. A. Kekule, Liebig's Ann., 101. 200,
1857 ; A. M. Butleroff, Zeit. Chem., 4. 549, 1861 ; Lehrbuch der organischen Chemie, Leipzig,
1868 ; E. Patemo, Giorn. Scienze Nat. Palermo, 5, 1869 ; Gazz. Chim. Ital., 23. 35, 1893 ; A.
Gaudin, L^ architecture du monde des atomes, Paris, 1873 ; K. Auwers, Die Entwicklung der Stereo-
chemie, Heidelberg, 1890 ; A. Wunderlich, Configuration organischer Molekule, Leipzig, 1886 ;
J. Wislicenus, Ber., 21. 581, 1888 ; A. Naumann, ib., 23. 477, 1890 ; A. von Baeyer, ih., 18. 2277,
1885; A. G. V. Harcourt, B. A. Rep., 32. 1875 ; W. H. Wollaston, Phil. Trans., 98. 96, 1808 ;
• F. Stohmann and C. Kleber, Journ. prakt. Chem., (2), 45. 475, 1892 ; I. Traube, Ueher den Raum
der Atrnne, Stiittgart, 1899 ; N. N. Stokes, Bull. U.S. Geol. Sur., 167. 117, 1900 ; J. Thomsen,
Thermochemische Untersuchungen, Leipzig, 1882; H. Sachse, Zeit. phys. Chem., 10. 203, 1892.
§ 16. The Evolution of the Valency Concept
The doctrine of valency is no mere speculation or hypothesis evolved by the brilliant
fancy or imagination of one man ; it is the logical outcome of knowledge acquired step by
step. The conception has been one of slow growth, for it gradually incorporated itself
into science as the necessity arose for devising a suitable explanation for accumulated
observations. — E. P. Venable (1899).
While the mists which enveloped the concepts, molecule, atom, and equivalent,
were being dispelled by illuminating rays of A. Avogadro's hypothesis, many theories
to explain chemical composition were struggling for existence. In the resulting
controversies — chiefly among J. J. Berzelius, J. B. A. Dumas, J. von Liebig, A.
Laurent, and C. F. Gerhardt — the facts were interpreted by different hypotheses ;
and, as A. Ladenburg (1886) has shown, this was far more favourable for progress
than if a single theoretical opinion had come too prominently in front. The differ-
ences of opinion quickened interest and experiment, and gave chemistry a very
intimate knowledge of many classes of compounds, because the advocate of each
hypothesis tried to support his own views by evidence which could be obtained
only by a close study of the chemical characteristics of the compounds in dispute.
It is difficult for chemists to appreciate the labour involved in clarifying the concepts
which now appear so simple. As one writer has said :
In the glamour of recent discoveries and the attractiveness of what is new and startling,
the pioneer spade work of a bygone age is forgotten or undervalued, and A, Carrel adds : Almost
every step in scientific progress which appears to be due to the efforts of one individual is,
in reality, the result indirectly of the unknown scientific work of many others.
The radicle or radical theories. — In his Traite elementaire de cJdmie (Paris,
1793), A. L. Lavoisier supposed that chemical compounds were formed by the union
of two bodies, and stated his belief that the composition of organic bodies depended
upon the existence of complexes or radicles in union with oxygen. Adopting a
suggestion of Guyton de Morveau (1787), Lavoisier called that portion of a compound
which is combined with oxygen, la base or le radicle. In developing his celebrated
dualistic polar hypothesis, J. J. Berzelius (1817) extended Lavoisier's idea. The
dominant feature of Berzelius' hypothesis is that chemical compounds can all be
resolved with two distinct parts electrically different. When J. B. A. Dumas and
P. F. G. BouUay (1828) i announced their belief that ether, (C2H5)20, consisted of two
COMBINATION BY VOLUME 217
parts— water, HgO, and a basic radicle, C2H4, which was called at the suggestion of
J. J. Berzelius cetherine — the radicle etherine was thought to be always present in
what are now called ethyl compounds. For instance, alcohol, C2H6O," would have
been regarded as a binary compound, C2H4.H2O ; and ether, C4H10O, as 2C2H4.H2O.
J. J. Berzelius at first opposed this hypothesis, but he afterwards incorporated the
idea in his dualism. J. B. A. Dumas and P. F. G. Boullay's etherine hypothesis
was not generally accepted because it did not adapt itself to the many new organic
compounds soon afterwards discovered. The interesting feature about this
hypothesis is that it represents an attempt to find a similarity in the structure of a
series of chemical compounds which possess like fundamental properties by showing
that they are all derived from one common primitive stock or type.
In 1815, in his memoir Recherches sur V acide prussique, J. L.Gay Lussac2 announced
the discovery of the radicle cyanogen, C2N2 (Kvavo?, blue ; y^wdw, I produce) ;
and he showed that the group ON, or Cy, persists as a radicle through a whole series
of chemical compounds.
Cy.H Cy.Cl Cy.Br Cy.NHg CHaCO.Cy
Cyanogen hydride. Cyanogen chloride. Cyanogen bromide. Cyanamide. Acetylcyanide.
Again, in 1832, J. von Liebig and F. Wohler described a series of compounds of the
radicle benzoyl, CqHsCO, of benzoic acid, in a memoir entitled Untersuchungen uber
das Radikal der Benzoesdure, The benzoyl radicle persists in many chemical com-
pounds— among others
C6H5CO.H CgHsCO.OH CeHgCO.Cl CeHgCO.NHg CeHgCO.OCgHs
Benzoyl hydride. Benzoic acid. Benzoyl chloride. Benzamide. Ethyl benzoate
The recognition of these two radicles — cyanogen and benzoyl — led to the development
of what is now called the older radicle theory. On this hypothesis complex groups
or radicles were supposed to exist unalterable in organic compounds, and to play
the same role as elements do in inorganic compounds. According to J. von Liebig
(1837),
Cyanogen is a radicle (I) because it is a non-varying constituent in a series of compounds ;
(2) because in these latter it can be replaced by other simple substances ; and (3) because
in its compounds with a simple substance, the latter can be turned out and replaced by
equivalents of other simple substances.
Hence, while this hypothesis was in favour, organic chemistry was regarded by
J. B. A. Dumas and J. von Liebig (1837) as the Chemistry of Compound Radicles.
The purpose of organic chemistry was supposed to involve the investigation and
isolation of radicles as the more intimate components of organic compounds.
Cyanogen and benzoyl, said A. Ladenburg (1869), were the pillars of the radicle
theory, and this hypothesis received further support from the work described in
R. Bunsen's brilliant memoirs, Untersuchungen uber die Kakodylreihe (1839-43), in
which it was shown that the so-called Cadet' s fuming liquid — obtained by A. A. F.
Cadet in 1760 by distilling potassium acetate with arsenious oxide — contained the
oxide of a radicle, with the empirical formula, As(CH3)2, and which he called kakodyl
or cacodyl {KdKO)Sr]<i, ill-smelling). R. Bunsen succeeded in isolating the radicle
itself, and also in preparing various salts — the chloride, bromide, fluoride, sulphide,
etc. Modifications of the theory of radicles were discussed by J. J. Berzelius (1833)
and J. von Liebig (1834-38), and the principle of radicles was generally accepted
although it was not so much emphasized during the reign of the so-called type
theories.
The type theories.— In 1834, J. B. A. Dumas 3 found that the hydrogen of many
organic compounds could be replaced or substituted by chlorine in such a way
that for every volume of chlorine introduced into a compound, an equal volume of
hydrogen was lost ; and, shortly afterwards, J. B. Dumas found that when oxygen
displaces hydrogen, half a volume of oxygen takes the place of one volume of hydro-
gen. Otherwise expressed, while equal volumes of hydrogen and chlorine are
218 INORGANIC AND THEORETICAL CHEMISTRY
equivalent, these elements possess only one-half the substituting value of the same
volume of oxygen. A further study of substitution or metalepsis (/xcraXr^i/Ats,
exchange), led J. B. A. Dumas, in his memoir, Sur k constitution de quelques corps
organiques et sur la theorie des substitutions (1839), to the so-called substitution
theory. J. B. A. Dumas discovered two important facts in his investigation of the
action of chlorine on some organic compounds : (1) When a compound containing
hydrogen is exposed to the dehydrogenating action of chlorine, bromine, or iodine,
for each atom of hydrogen that it loses, it takes up an equivalent volume of chlorine,
bromine, etc. (2) If a compound contains water, it loses the hydrogen without
an equivalent substitution or replacement. The main assumption of the substitution
theory hangs on the doctrine that the structure and character of organic compounds
are not materially altered by the substitution of chlorine in place of hydrogen.
A. Laurent in a paper Theorie des comhinaisons organiques (1836), and later, in
his posthumous work, Methode de chimie (Paris, 1854), tried to reconcile the radicle
theory with these new facts discovered by J. B. A. Dumas. When the substitution
occurs equivalent by equivalent, the residual body exhibits certain analogies with
the original substance, for the substitution occurs without disturbing the structural
type — chlorine, for instance, may occupy the place left vacant by hydrogen. A.
Laurent argued that all organic compounds have definite forms or nuclei — radicaux
— and consist either of primary nuclei — radicaux fundamentaux — or of secondary or
derived nuclei — radicaux derives — in which the hydrogen atoms have been replaced
by others, or in which additional atoms have been taken up. This hypothesis was
called the nucleus theory ; it included the idea of substitution, and was based on
the radicle theory ; but it controverted the doctrine that radicles were unchange-
able, for the atoms of a radicle can be replaced by others ; it gave the first hint of
what is now known as " chemistry in space." The nucleus theory was specially
favoured by L. Gmelin in his celebrated Handhuch der Chemie (Heidelberg, 1843
et seq.), but it was not taken up by chemists generally.
In 1839, J. B. A. Dumas prepared trichloroacetic acid, CCI3.COOH, in which
three of the hydrogen atoms of acetic acid, CH3.COOH, are replaced by chlorine,
and the resulting compound retains the chief characteristics of the parent acid.
This led him, in his Memoire sur la hi des substitutions et la theorie des
types (1840), to extend Laurent's nucleus theory to what is now known as the
older theory of types, in which organic substances are supposed to be formed of
particles which may be replaced or displaced, so to speak, without destroying the
original substance. Compounds which have similar properties and a similar structure
were classed as belonging to one chemical type — e.g. acetic acid and the chloroacetic
acids. The relations between the members of a series of compounds belonging to
one chemical type thus recall those assumed by A. Laurent to subsist between the
original and the derived nuclei. J. B. A. Dumas also found it necessary to employ
what he called the rnechanical type to classify compounds which are related in struc-
ture but which manifest different chemical characteristics. Dumas rightly classed
acetic acid and alcohol under the same mechanical type, which included a number
of compounds which had little or no chemical relations with one another, though
they may be regarded as belonging to one natural family because they may be
derived by substitution one from the other — e.g. methane, CH4 ; formic acid,
H.CO.OH ; carbon tetrachloride, CCI3.CI. J. B. A. Dumas' mechanical type
resembled what H. V. Regnault (1838) had previously called the molecular type.
If a substance changes without losing its mechanical type, it follows the law of
substitution, but if it passes into another mechanical type, the law of substitution
is not maintained during the reaction. By this statement, J. B. A. Dumas admits
that his original idea of substitution is not always applicable, for an equivalent of
hydrogen is not always evolved when another is introduced into the compound ;
and a compound is not regarded as consisting of two parts, but is supposed to be a
uniform whole with its component parts related in an analogous fashion to the worlds
of a planetary system in which the atoms are held together by affinity instead of by
COMBINATION BY VOLUME 219
gravitation. Just as the stability of a planetary system depends not on the intrinsic
nature of the planetary units, but rather on their relative position with respect to
one another and to the sun, so J. B. A. Dumas supposed that the chemical properties
of a compound are primarily dependent on the arrangement and number of the
constituent atoms, and in a less degree on their chemical nature. Dumas thus
regarded the planetary molecule as the tyjpe of a series of compounds with a similar
structure ; and therefore, he opposed a unitary theory of chemical coynfosition in
place of J. J. Berzelias' dualism. Between 1838 and 1844, J. J. Berzelius vigorously
fought a losing fight in his Jahresherichten against these encroachments on his
dualistic views. C. F. Gerhardt having suggested that in compounds of an
organic base with an inorganic acid, the organic portion of the compound, termed
the copula, was supposed to unite by accouplement (copulation) with the inorganic
acid. J. J. Berzelius tried to explain the substitution products obtained by J. B. A.
Dumas by arbitrarily assuming that they were formed by the copulation or pairing
of imaginary copulae ; he explained the formation of trichloroacetic acid, for example,
by assuming it to be formed by the union of carbon chloride, C2CI6, with oxalic
acid, H2C2O4 ; and he assigned different rational formula? to trichloroacetic acid
and its parent — acetic acid. Berzelius' explanation broke down completely when
L. H. F. Meslen (1842) showed that chloroacetic acid could be reconverted to the
original acid by reduction with potassium amalgam. J. J. Berzelius obstinately
opposed the theory of t5rpes with his last breath, but he fought practically alone.
His one-time supporter, J. von Liebig, gave up the duaUstic hypothesis when it failed
to explain the newer facts.
In 1839, C. F. Gerhardt, in his memoir Sur la constitution des sels organiques a
acides complexes et leur rapports avec les sels ammoniacaux, rejected the radicle
theory and stated his belief that a compound must be regarded as a complex of atoms
bound each to all, and all to each ; but he could not help admitting that certain
groups of atoms do recur in chemical compounds. Accordingly, C. F. Gerhardt
attempted to reconcile his h5rpothesis with observation, by what he called his
theorie des residus — theory of residues — in which a group of atoms previously called
a compound radical was termed le reste — a residue ; unlike radicles, C. F. Gerhardt's
residues were not supposed to be present as such in a compound, for, said C. F.
Gerhardt, je prends V expression de radical dans le sens de rapport, et non dans celui
de corps isolahle ou isole. C. F. Gerhardt's molecule was une systeme unitaire — a
simple edifice and not a double building ; all assumptions of a binary structure
were excluded. He argued that the constitution of a compound can be deduced
only from its modes of formation and decomposition, and that according to the theory
of radicles several rational formulae and several radicles could be imagined in the
case of one substance formed in different ways — e.g. barium sulphate formed by
the reactions symbolized : (i) BaO+SOg ; (ii) BaOg+SOg ; and (iii) BaS-1-202. Con-
sequently, a chemical type is nothing more than a general system of reactions. Acetic
acid, water, and alcohol were classed in the same way because they undergo analogous
reactions — say, when they are deoxidized to form aldehyde, hydrogen, and ethyl
hydride, C2H5H, respectively. Gerhardt further supposed that when two substances
react with one another, an element in the one combines with an element in the other
to form one stable compound, and the residues also unite to form what he called
corps copules and later corps conjuges, meaning copulated or conjugated compounds.
Thus, the copula benzene, CgHg, unites with nitric acid, HNO3, to form water
HoO, and the copulated compound nitrobenzene, CeHg.NOa. In modernized
symbols :
NO2) CeHs) _H ) ^CgHsi
OH i H ! OH? NO2 f
Consequently, in a reaction between two substances, each molecule is split into
two parts, and the resulting residues unite in such a manner that a double exchange
takes place, and Gerhardt said : J'appelle radicaux ou residus les elements de tout
220 INORGANIC AND THEORETICAL CHEMISTRY
corps qui peuvent etre ainsi transportes dans un autre corps par Veffet d'une double
decomposition, ou qui y ont ete introduits par une semhlahle reaction. It is not very-
obvious why C. F. Gerhardt emphasized the distinction between his own type
formulae and those of J. B. A. Dumas. The former clearly supposes substitution
to be effected by replacing an element in a compound by an equivalent of another
element, or by the residues (radicles) of the reacting substances, and this is but a
restatement of the views of J. B. A. Dumas and A. Laurent. C. F. Gerhardt's
conception of radicles, said C. Schorlemmer (1879), soon supplanted the older views,
and its introduction into the theory of types led to the fusion of both theories.
The discovery of the organic ammonias by C. A. Wurtz (1849) and A. W. von
Hofmann (1850) * revealed the close relationship between the organic ammonia
bases and ammonia itself, and the hypothesis that the former were derivatives of
ammonia, NH3, produced by the substitution of hydrocarbon radicles in place of
hydrogen atoms :
NH3 (C2H5)NH2 (C2H6)2NH (C2H5)3N
furnished the only satisfactory explanation of the constitution of these compounds.
In this way, said C. A. Wurtz, the ammonia type was founded. Similarly, A. W.
Williamson's Theory of Mtherijication (1850), dealing with the substitution of hydro-
carbon radicles in place of the hydrogen atoms of water, established the water type.
A. W. Williamson demonstrated the close relationship between
Hjo ^2^5 \n CgHsi
'l^]o
\0
Water. Alcohol. Ether.
In harmony with a prior suggestion made by A, Laurent in 1846, A. W. Williamson
wrote : "I believe that throughout inorganic chemistry and for the beet known
organic compounds, one single type will be sufficient — it is that of water represented
as containing two atoms of. hydrogen to one of oxygen." Numerous nitrogen com-
pounds were then referred to C. A. Wurtz and A. W. von Hofmann 's ammonia type,
and many oxygen compounds were likewise referred to A. W. Williamson's water
type as termes de comparaison. C. F. Gerhardt in his Traite de chimie organique
(Paris, 1853-6) added hydrogen and (lydrogen chloride to the ammonia and water
types, and he attempted to classify all organic compounds by reference to the four
types : hydrogen, H2 ; hydrogen chloride, HCl ; water, H2O ; and ammonia, NH3.
In 1857, F. A. Kekule, in an important memoir Veber die sogenannten gepaarten
Verhindungen und die Theorie der 7nehratomigen Radicle, proposed to add methane,
CH4, to the list of primitive or simple types, and to remove hydrogen chloride from
the list because it is merely a special case of the hydrogen type. Thus arose the
newer theory of types which now assumed the forms
h} > |n
H)
H C
H^
Hydrogen type. Water type. Ammonia type. Methane type.
A. Laurent had suggested in 1846 that alcohol and ether as well as inorganic acids
and oxides could be regarded as derivatives of water. In 1848-9, T. S. Hunt
published several papers in which he showed that the composition of many oxyge-
nated compounds might be derived from water as a type, and he also referred the
formul£e of hydrocarbons to hydrogen as a type ; but T. S. Hunt's work had little
or no influence on the development of the theory of types since it was unknown to
those who were working in Europe on this subject.
A. W. Williamson introduced the idea of condensed types in 1850 ; dibasic acids
like sulphuric acids and oxalic were regarded as derived from two molecules of water,
and the acid radicle was supposed to replace one atom of hydrogen in each of the
two molecules of water. W. Odling, in his paper On the Constitution of Acids
COMBINATION BY VOLUME 221
and Salts (1855), developed the idea still further, and formulfie like these were
obtained :
h}0 type
5)0, type
|3|03type
C^HsOjo NO.Jo
^202)0 . S02 ^
''Xt}o. ly.
Acetic acid. Nitric acid.
Oxalic acid. Sulphuric acid.
Citric acid. Phosphoric acid.
F. A. Kekule (1857) also extended the type theory to include mixed types
supposed to be formed by the union of two simple or condensed types. For
example, chlorosulphuric acid, (H0)C1S02, can be referred to a mixed hydrogen
and water type ; and carbamic acid, HgN.COOH, was referred to the mixed
ammonia and water types :
h} so) h1- co^^
> H> h1« h}0
Mixed type. Chlorosulphuric acid. Mixed type. Carbamic acid.
The need for the introduction of condensed and mixed types showed the insufficiency
of the type theory, for as the number and complexity of organic compounds increase,
an indefinite number of types may be required. C. F. Gerhardt's type theory is
now considered but an interesting phase in the evolution of systematic chemistry.
The attempts to refer a large number of compounds to a limited number of types,
and the consequent need for viewing individual compounds from many different
points de vue, enabled chemists to see many analogies and contrasts previously
hidden, and to realize dimly the remarkable relations the atoms of a compound
bear each to each. It soon became evident that the theory of types represented
an artificial arbitrary system of classification ; even C. F. Gerhardt (1856) admitted
mes radicaux et mes types ne sont que des symholes, destines d concreter en quelquc sorte
certains rapports de composition et de transformation, and H. W. Kolbe (1843 et seq.)
seemed to get at the root of the matter when asked : *' Why are we to suppose
that nature has restricted herself to forming all bodies on the models of these four
types 1 Why on these models rather than on others ? The four types are nothing
but a vain artifice." He answered that " the grouping of organic compounds into
types verges on empty formalism, and is merely playing with formulae." He
sought to replace the purely formal types by others which he considered to be
related naturally with their derivatives. In a paper Ueber die chemische Konsti-
tution und Natur der organischen Radikale (1851), H. W. Kolbe built up a newer radicle
theory in which he eliminated those tenets which were not in harmony with fact ;
he showed, as J. J. Berzelius supposed, that in organic compounds there are definite
radicles which behave like the elements in inorganic compounds. The discovery
of the organometallic compounds — typified by zinc ethyl, Zn(C2H5)2 — by E. Frank-
land (1849), seems to exclude every doubt of the actual existence of compound
radicles ; and H. W. Kolbe (1850) electrolyzed aqueous solutions of the salts of the
fatty acids, and believed that he separated the constituent hydrocarbon radicles —
as a matter of fact, he obtained products of the union of two radicles. Thus, with
potassium acetate, CH3.CO.OK, he obtained gaseous carbon dioxide and ethane,
(CH3)2 or C2H6 ; the potassium, simultaneously obtained, undergoes a secondary
reaction with the solvent. The primary reaction in modern symbols is represented :
iCHg'.iCO.OlK CO2 , CHg J.
ICHgi-iCO.OlK CO2 CH3 "^
In conjunction with E. Frankland, H. W. Kolbe published a paper entitled Veber
den natUrlichen Zusammenhang der organischen mit den anorganischen Verbindungen,
die wissenschaftliche Grundlage zu einer naturgemassen Klassifikation der organischen
chemischen Korper (1859), in which it was shown by numerous examples that organic
222 INORGANIC AND THEORETICAL CHEMISTRY
compounds can be regarded as derivatives of inorganic compounds, and result
from the latter — in some cases directly — by wonderfully simple substitution pro-
cesses. Consequently, organic acids can be regarded as substitution derivatives of
carbonic acid, and consequently, H. Kolbe argued that carbonic acid is a natural
standard of reference for organic bodies because they are formed from this gas in
the vegetable kingdom. He said : The carbonic acid type must therefore exist
in the very nature of things, and it seems logical to refer all organic compounds
to this type, since they are all in fact derived from it.
For example — translating Kolbe's symbols into modem practice, and starting from
carbonic acid, (H0)2C0- — ^when a jhydroxyl group, HO, is replaced by a hydrogen, H, atom,
formic acid, H.CO.OH, is formed ; replacing OH by CHg furnishes acetic acid, CH3.CO.OH ;
replacing two OH-groups by two H-atoms yields formaldehyde, H CO.H. If an OH-group
be replaced by an H-atom, and an 0-atom by two H-atoms, methyl alcohol, CH3.OH,
results ; and if an OH-group is replaced by CHg, and an O-atom by two H-atoms, ethyl
alcohol, CH3.CH2.OH, is formed.
H. Kolbe is here perhaps a little inconsistent, for C. A. Wurtz, in his Histoire des
doctrines chimiques depuis Lavoisier jusqu'd nos jours (Paris, 1869), has pointed out
that water and ammonia are agents as indispensable as carbonic acid in the pro-
cesses of vegetable life. Kolbe's objections to C. F. Gerhardt's or F. A. Kekule's
types also apply to his own carbonic acid type.
The doctrine of valency. — In conformity with the general views of chemists
early in the nineteenth century, J. L. Gay Lussac,^ in his Recherches sur Vacideprussique
(1815), regarded salts as products of the union of an equivalent of an acid with an
equivalent of the base, but T. Graham's important Researches on the Arseniates,
Phosphates, and Modifications of Phosphoric Acid, published in 1833, showed that
phosphorus pentoxide can unite with one, two, and three equivalents of water to
form definite acids which can respectively unite with but one, two, and three
equivalents of the base to form definite salts with characteristic properties. Five
years later, J. von Liebig, in his memoir Ueber die Constitution der organischen
Sduren (1838), found other acids to behave in a similar manner, and he employed
the terms mono-, di-, tri-, and poly-basic acids to indicate the saturation" capacity
of the acids for the bases. The idea of basicity was further extended to organic
radicles ; and, in 1834, J. B. A. Dumas showed that an atom of hydrogen could
be replaced by an atom of chlorine, but only by the equivalent of half an atom of
oxygen, so that these quantities of chlorine and oxygen are equivalent to an atom
of hydrogen.
In his memorable paper On a New Series of Organic Compounds containing
Metals, published in 1852, E. Frankland applied the idea of equivalency or satura-
tion capacity to the elements. He showed that the power of the metals to combine
with oxygen is reduced when the metal is copulated with compound radicles in such
a way that, say, stannic ethyl oxide, (C2H5)2SnO, is to be regarded as stannic oxide,
Sn02, in which one oxygen atom is replaced by two ethyl radicles ; and stannic
ethide, Sn(C2H5)4, as stannic oxide with the two oxygen atoms replaced by four
ethyl radicles. E. Frankland then remarked :
When the formulae of inorganic chemical compoimds are considered, even a superficial
observer is impressed with the general symmetry of their construction. The compounds
of nitrogen, phosphorus, antimony, and arsenic especially exhibit the tendency of these
elements to form, com/pounds containing three or five atoms of other elements ; and it is in these
proportions that their affinities are best satisfied. . . . Without offering any hypothesis
regarding the cause of this symnietrical grouping of atoms, it is sufficiently evident, from
the examples just given, that such a tendency or law prevails, and that, no matter what the
character of the uniting atoms may be, the combining power of the attracting element is
always satisfied by the same number of these atoms.
Thus Frankland led chemists to see that within certain limits the atoms of the
elements possess definite saturation capacities ; and he proved that copulation
is a consequence of the saturation capacity of the elements. In 1877, Frankland
added that the hypothesis just outlined ' constitutes the basis of what has since
COMBINATION BY VOLUME 223
been called the doctrine of atomicity or the equivalence of the elements." The far-
reaching importance of the above quotation from Frankland was not realized until
some years afterwards. A vague inkling of the operation of some such law among
organic compounds was probably at the back of the minds of the founders of the
different theories of types, for in representing chemical transformations as the result
of substitutions of atoms or groups of atoms, the equivalency of the substituents
must have been tacitly assumed ; but they were prevented from realizing the
importance of the principle by laying too much stress on the position rather than on
the nature of the atoms concerned.
F. A. Kekule seems to have considered himself to have been the originator of
the doctrine of the valency, or, as he termed it, the atmnicity of the elements. As a
matter of fact, in 1854, two years after the publication of E. Frankland's paper,
F. A. Kekule did obtain a clearer vision of the doctrine, and, in 1857, he explained
the existence of primitive types — simple and mixed — by means of the valency of the
constituent elements. Soon afterwards, A. S. Couper, independently of F. A.
Kekule, published a paper, Sur une nouvelle theorie chimique (1858), in which he
deduced constitutional formulae for many compounds from the valency of the
elements, or rather, what he called affinity of degree of the elements as contrasted
with the ordinary manifestations of chemical affinity, or, as he called it, elective
affinity. A. S. Couper, for the first time, also represented the composition of com-
pounds by joining the symbols of the elements or compound radicles by means of
hyphens or linking bonds. In his Lehrhuch der organischen Chemie, oder der
Kohlenstoff-Verhindungen (Stuttgart, 1859), F. A. Kekule symbolized the valency
of an atom in graphic formulae by means of a diagram whose size represented the
valency as illustrated in the following examples :
Hydrogen chloride, HCl. Water, HgO. Sulphur dioxide SOg. Nitric acid, HNO3.
There was no intention, of course, to convey any idea of the relative dimensions of
the atoms. In 1865, A. C. Brown suggested a system in which the symbol of the
element was surrounded by a circle, with a number of radiating Hues corresponding
with the valency of the element. For instance,
©KB) ®:®
Hydrogen chloride, HCl. W^ater, HgO. Sulphur dioxide, SOg. Nitric acid, HNOg.
The grouping was not meant to indicate the physibal but rather the chemical position
of the atoms. E. Frankland adopted practically the same system in 1866, except
that he omitted the circles round the symbols of the elements, and this method of
pictorially representing the linking of the atoms of a molecule in definite order is
virtually that employed by A. S. Couper, and it has persisted up to the present day.
A. M. Butleroff (1861) followed up A. S. Couper's idea, and defined the structure of a
chemical compound to be the mode in which the atoms are mutually linked together
in the molecule. This does not afford any information of the position of the indi-
vidual atoms in space. The chemical characteristics of a compound, said Butleroff,
depend first upon the nature and relative quantity of its elementary constituents,
and then on its chemical structure.
F. A. Kekule (1857) classified the elements according to the replacing values of
their atoms. Hydrogen, chlorine, potassium, etc., were called monobasic o-r mon-
atomic elements ; oxygen and sulphur were dibasic or diatomic ; nitrogen, phos-
phorus, and arsenic were tribasic or triatomic ; and carbon was classed as a tetra-
basic or tetratomic element. There is an incongruity in the use of the terms
mono-, di-, . . . atomic, since similar terms are employed to represent the number
224 INORGANIC AND THEORETICAL CHEMISTRY
of atoms in a molecule ; and the confusion in the use of the terms mono-j
bi-y . . . basic atoms with J. von Liebig's polybasic acids, led E. Erienmeyer (1860)
to propose the terms ein-, zivei-^ drci-, and vier-iverthig which have come into use
in Germany ; the equivalent uni-y bi-, ter-, and quadri-valent, used by L. Meyer,
or mono-, di-, tri-, and tetra-valent, with W. Odling's alternative terms (1864) :
monad, dyad, triad, and tetrad are now in use. J. Wislicenus used the terms
monaffin, diaffin, triaffin, and tetraffin. In 1855, W. OdUng placed dashes beside
the symbol of the atom or radicle to express what he called the replaceable, or
representative, or substitution value of the atoms, and he recognized, as E. Frank-
land did in 1852, that an atom can have more than one replaceable value. Various
terms were used in place of valency during the clarification of the concept — e.g.
saturation cajpacity, combining capacity, atom-fixing power, affinity units, affinity
of degree, basicity, and atomicity. The two latter terms are objectionable.
A. W. Hofmann (1865) considered that atomicity is a barbarous term ; and is best
reserved to express the number of atoms in a molecule of an element ; the term
basicity is also best retained to express the number of stages in which the replace-
able hydrogen of an acid can be substituted by a metal. A. W. Hofmann did much
to spread a knowledge of the doctrine of valency. He employed the term quanti-
valence " to designate the particular atom-compensating power inherent in each of
the elements," and added " this power must by no means be confounded with the
specific intensity of the respective activities of the atoms." H. Wichelhaus ^
shortened A. W. Hofmann's quantivalence to valency (or valence) in 1868 : and
H. Wichelhaus' term is now in general use.
The doctrine of valency introduced by E. Frankland and amplified by F. A.
Kekule soon stilled the controversies which had been waged between the advocates
of the radicle and type theories. The nature of the problem was changed. Chemical
formulae were no longer employed to represent types of double decomposition,
but rather to show the relations which subsisted between the constituent atoms
of a molecule. The doctrine of valency enabled chemists to see, as in a glass darkly,
the intimate structure of the molecules by establishing the way in which the atoms
are bound together. Consequently, neither the type nor the radicle theory could claim
a victory, for the theory of composition based upon valency absorbed and assimilated
them both ; it showed that chemists had really admitted a water type because
there is a bivalent element oxygen ; an ammonia type because there is a tervalent
element nitrogen ; and a methane type because there is a quadrivalent element
carbon. As F. A. Kekule's mixed metaphor expressed it : " Both sides had been
striving towards the same goal by different paths ; each side thereupon profited
by the experience of the other, and with united forces sailed onward on the reunited
stream."
Refebences.
1 A. Ladenburg, Vortrdge iiber die Entwicklungsgeschichte der Chemie, Braunschweig, 1869 ;
A. L. Lavoisier, G. de Morveau, and A. F. de Fourcroy, Mithode de nomenclature, Paris, 1787 ;
J. B. A. Dumas and P. F. G. Boullay, Ann. Chim. Phijs., (2), 37. 15, 1828 : J. J. Berzelins,
Jahresh., 9. 286, 1830 ; 13. 190, 1834 ; Liebig's Ann., 3. 282, 1832.
2 A. Ladenburg, Vortrdge iiber die Entwicklungsgeschichte der Chemie, Braunschweig, 1869 ;
J. L. Gay Lussac, Ann. Chim. Phys., (1), 95. 136, 1815 ; F. Wohler and J. von Liebig, Liebig's
Ann., 3. 249, 1832 : J. von Liebig, Liebig's Ann., 25. 3, 1837 ; 9. 1, 1834 ; 11. 10, 1834 ; 19. 270,
1836 ; Pogg.Ann., 21. 533, 1831 ; Ann. Chim. Phys., (2), 37. 15, 1828; J. B. A. Dumas and J. von
Liebig, C(mpt. Rend., 5. 567, 1837 ; R. Bunsen, Liebig's Ann., 24. 271, 1837 ; 31. 175, 1839 ;
37. 1, 1841 ; 42. 14, 1842 ; 46. ], 1843 ; J. J. Berzelius, Pogg. Ann., 28. 626, 1833.
3 J. B. A. Dumas, Ann. Chim. Phys., (2), 56. 113, 140, 1834 ; (2), 73. 73, 1840 ; J. B. A. Dumas
and J. S. Stas, ib., (2), 73. 113, 1840; J. B. A. Dumas and E. M. Peligot, ib., (2), 74. 5, 1840;
J. B. A.-Dumas, C&mpt. Rend.,Q. 699, 1838 ; 7. 474, 1838 ; 8. 609, 1839 ; 10. 149, 1840 ; A. Laurent,
Ann. Chim. Phys., (2), 52. 275, 1833 ; (2), 59. 196, 1835 ; (2), 60. 220, 1835 ; (2), 61. 125, 1836
(2), 63. 27, 42, 207, 377, 1836 ; (3), 18. 266, 1846 ; Compt. Rend., 10. 409, 1840 ; H. V. Regnault:
Ann. Chim. Phys., (2), 59. 358, 1835 ; Liebig's Ann., 15. 60, 1835 ; 30. 139, 1839 ; J. von Liebig
Liebig's Ann., 31. 119, 1839 ; 32. 72, 1839 ; 33. 301, 1840 ; 50. 295, 1844 ; C. F. Gerhardt, Ann
Chim. Phys., (2), 72. 184, 1839; Precis de chimie organique, Paris, J 842; Journ. prakt. Chem
COMBINATION BY VOLUME 225
(1), 27. 439, 1842 ; (1), 30. 1, 1843 ; H. F. Meslen, Ann. Chim. Phys., (3), 10. 233, 1842 ; E.
Grimaux and C. Gerhardt, Charles Gerhardt, sa vie, son oeuvre, sa correspondence, Paris, 1900 ;
C. Schorlemmer, Bise and Development of Organic Chemistry, London, 1894.
* C. A. Wurtz, Compt. Berid., 28. 223, 1849 ; A. W. von Hofmann, Liebig's Ann., 74. 174,
1850 ; A. W. Williamson, Jotirn. Chem. Soc., 4. 106, 229, 1852 ; H. Kolbe, ib., 7. Ill, 1855 ;
W. Odling, ib.,7. 1, 1855; F. A. Kekule, Liebig's Ann., 104. 129, 1867; T. S. Hunt, Amer. Joum,
Science, (2), 6. 170, 1848; (2), 7. 175, 1849; (2), 8. 89, 1849; A. Laurent, Ann. Chim. Phys., (3),
17. 331, 1846; (3), 18. 266, 1846; H. Kolbe, Liebig's Ann., 45. 41, 1843; 54. 145, 1845; 69.
258, 1849 ; 75. 211, 1850 ; 76. 1, 1850 ; 113. 293, 1860 ; Handworterbuch der Chemie, Braun-
schweig, 6. 802, 1855 ; Veber die chemische Konstitution der organischen Kohlenwasser staff e,
Braunschweig, 1869 ; H. Kolbe and E. Frankland, Liebig's Ann., 65. 288, 1848 ; H. Kolbe, ib.,
101. 257, 1857.
s J. L. Gay Lussac, Ann. Chim. Phys., (1), 95. 136, 1815 ; T. Graham, Phil. Trans., 123. 253,
1833 ; J. von Liebig, Liebig's Ann., 26. 113, 1838 ; E. Frankland, Phil. Trans., 142. 417, 1852 ;
J. B. A. Dumas, Liebig's Ann., 32. 101, 1839 ; F. A. Kekul6, ib., 106. 129, 1858 ; 104. 133, 1857 ;
Ber., 23. 1265, 1890 ; A. S. Couper, Compt. Bend., 46. 1157, 1858 ; Phil. Mag., (4), 16. 104, 1858 ;
N. N. ButlerofE, Zeit. Chem,, 4. 549, 1861 ; E. Erlenmeyer, ib., 6. 65, 97, 609, 1863 ; 7. 1, 72, 628,
1864 ; W. Odling, Journ. Chem. Soc, 7. 1, 1855 ; A. C. Brown, t6., 18. 230, 1865 ; Proc, Boy. Soc.
Edin., 5. 429, 561, 1866.
^ H. Wichelhaus, Liebig' s Ann. Suppl., 6. 257, 1868 ; A. W. von Hofmann, Introduction to
Modem Chemistry, London, 169, 1865 ; L. Meyer, Die modernen Theorien der Chemie, Breslau, 67,
1864 ; E. Erlenmeyer, Zeit. Chem., 3. 540, 1860 ; J. Wislicenus, Liebig's Ann., 128. 2, 1863.
§ 17. Attempts to explain Valency
The general test of truth is evidence. — J. M. C. Duhamel.
The composition of all chemical compounds, says H. von Euler (1903), can be
regarded as a function of a valency force — Valenzkraft — which is probably of an
electric nature, and dependent on the temperature, pressure, and the nature of the
solvent. Numerous attempts have been made to invent some peculiarities in the
structure of the atoms which will explain that strange power manifesting itself
as valency. Even Lucretius attributed the differences in the behaviour of his atoms
to differences in their shape, size, and mode of motion. The subject has rather lent
itself to hypotheses established by the absence of a knowledge of contradictory
facts. A brief resume of the more striking forms of these hypotheses may act as a
danger beacon.
I. Differences in the valency of different elements have been explained by
supposing that an atom of an n-valcnt element is compounded of n units, each of which
is capable of attracting one other unit. A constant quantity of one element, said
E. Erlenmeyer (1862), i never binds itself to more or to less than a constant quantity
of another element-— this he called the law of constant affinivalencies. W. Odling
(1855) called these attracting units suh-atoms ; G. Ensrud (1907), Kernen or nuclei ;
L. Knorr (1894), Yalenzkorfer or valency bodies; E. Erlenmeyer (1867), affinivalencies ;
A. W. von Hofmann (1865), minimum atom-binding quantities of an element ; and
J. Wislicenus (1888), primitive atotns, which are located in certain parts of the atom
and from which they exert their influence. W. Lossen, in an important paper Ueber
die Vertheilung der Atome in der Molekul (1880), pointed out that this hypothesis
cannot be sound, for if a constant mass of, say, carbon binds itself to a constant
mass of oxygen in the molecule of carbon dioxide, CO2, the same mass of carbon is
bound to half the same constant mass of oxygen in carbon monoxide, CO. Hence,
the assumed constant mass must be variable. G. Ensrud (1907) supposed an atom
to be compounded of an enveloping shell of a substance of small density with a
nucleus of great density and eccentric shape. The envelopes of different atoms
repel one another, the nuclei attract one another in the direction along which valency
acts. An atom of an w-valent element has n nuclei. This hypothesis recalls J. F.
Redtenbacher's Das Dynamidensystem (Mannheim, 1857). Some of these hypo-
theses appear to have arisen by confusing the fractional parts of an atom with
fractional parts of its weight, and assuming that the former are equal to the
VOL. I. Q
226 INORGANIC AND THEORETICAL CHEMISTRY
latter. There is nothing to show that if the atom were divided up into a number
of attracting portions, each would be the same fractional part of the weight
of the atom. The modern electron hypothesis of valency is one form of this
hypothesis — vide Vol. III.
II. Other hypotheses assume that valency is an attracting force localized at certain
parts of the atom. The atoms are supposed to be joined together at these attracting
points ; in other words, some parts of the atom are less active than others. This
hypothesis has taken various forms. E. Erlenmeyer (1867) and A. Michaelis (1872)
suggested that the attractive forces are not exerted uniformly in all directions as is the
case with gravitation, but are specially strong in certain definite directions so that
a straight line joining two atoms directly bound together expresses the direction
of the mutually exerted force. A. Michaelis supposed an %-valent atom to have
n such directions, and, if it is bound by n—x bonds, to have these mutual actions
exerted in n—x such directions. A. C. Brown (1861-79) assumed that each atom
possesses two kinds of attractive forces — positive and negative — and the point
towards which these forces act was called a pole or active point. He made no as-
sumption as to the nature of the attractive or repellent forces. An /i-valent
element has n such positive and negative poles. When two atoms unite, the positive
pole of the one attracts the negative pole of the other, and vice versa. When a
bivalent atom combines with two univalent atoms, the forces emanating from the
bivalent atoms will be divided between its two poles in some proportion depending
on the forces of the two univalent atoms. In order to support the assumption that
valency is due to centres of attraction localized on the atom, subsidiary hypotheses
have to be invented. For instance, it has been assumed (i) that the atoms are bound
to one another through the attraction of electric or magnetic charges localized on
the atoms ; and also (ii) that the intensity of the attractive force is modified by the
shape of the atom.
(i) Electric charges localized on the atom. — The idea that the reacting units
are polarized, and carry definite electric charges, each charge representing
one valency, naturally grew from Davy's and Berzelius' electrochemical
hypothesis, and Faraday's work. There are many modified forms of the
hypothesis. For example, V. Meyer and E. Riecke (1888) assumed that the
carbon atom is surrounded by an aethereal envelope which, in the case of iso-
lated atoms, has a spherical shape like that supposed to be possessed by the
atoms themselves. The atom in the core carries the specific affinities ; the
sethereal envelope is the seat of the valencies. Each valency is determined
by the presence of two opposite electrical poles — called double or di-poles —
situated at the ends of a straight line which is small in comparison with the
diameter of the sethereal shell. The four valencies of carbon are represented
by four such di-poles each of which is able to move freely within the aethereal
shell, and to turn freely about its middle point. The carbon atom attaches
other atoms to its surface by the attractions of the di-poles. The modern form
of the electric charge hypothesis will be discussed later.
(ii) The shape of the atom. — J. H. van't Hoff, in his Ansichten iiher die
organische Chemie (Braunschweig, 1881), showed that the attractive forces
emanating from an atom will be uniform in all directions if the atom is spherical,
but if the shape be not spherical the intensity of the force, at short distances, will
be more concentrated in certain spots than in others. Thus, if the atom were
shaped like a regular tetrahedron, it would behave as if it were quadrivalent,
for the centres of the four bounding faces would represent maximal attractions.
Given the number of maximal points on the atom, it would be possible to deduce
the valency, and conversely. There will be as many maximal points as the
figure has sides. If the faces are unequally distant from the centre, the
maximal points may not all have the same value, so that, when the nature of
the united atoms also determines the attracting power, the number of effective
valencies of the attracting atom will be affected, and a change of valency will
COMBINATION BY VOLUME 227
be observed on comparing combinations of an element with other different
elements. J. Wislicenus (1888) has expressed a similar idea ; he said :
It is not impossible that the carbon atom more or less resembles — perhaps very
closely — the form of a regular tetrahedron ; and further, that the causes of
those attractions which are exhibited by the so-called units of affinity or bonds
are concentrated at the apices of this tetrahedral structure, so that where there
is least matter there is most force. These attractions are possibly analogous
to the electrical state of a metal tetrahedron charged with electricity.
If the atoms be also in rapid vibratory motion, only the parts where the
greatest attractions are exerted can retain their contacts, and therefore valency-
will be reduced by a rise of temperature, for a rise of temperature probably
augments the vibratory motions of the atoms.
III. Another set of hypotheses has assumed that valency is due to the need for
harmonizing the motions of the combining atoTns so as to form complexes whose parts
move in stable equilibrium. One form of this hypothesis is indicated later on.
According to L. Meyer (1884), the atoms in a molecule are not in a state of rest,
but they move rotationally about a centre of equiUbrium ; the orbits of similar
atoms in the molecules of the same substance are the same so that equivalent
atoms have the same paths, but the orbits of different atoms are greater, the greater
the valency of the atom. E. Molinari (1893) suggested a modification of this hypo-
thesis in a paper entitled Motochemistry (moto, motion). The valency of an atom
in a molecule is determined by the nature or energy of its oscillatory motion ; and
he claims that the constitution of compounds is dependent upon the intramolecular
movements rather than on the relative positions of the atoms in space. F. A.
Kekule (1872) considered that valency is determined by the relative number of
impacts which an atom receives from other atoms in unit time ; each of the uni-
valent atoms in a diatomic molecule impinges once, while the bivalent atoms
impinge twice in unit time. It is not very clear how this explains valency, and in
1878, F. A. Kekule said that " the nature of the motion of atoms, unknown at present,
may be imagined as an oscillatory one in such a way that the number of oscillations
executed in unit time exactly represents the valency of the atoms." F. M. Fla-
vitzky (1896), following N. N. Beketoff (1880),^ supposed that the atoms move
in curves which lie in planes parallel to one another ; the atoms of different elements
move in planes which are inclined at definite angles to one another ; the motion
of the atoms of one element can be completely counteracted by the motions of the
atoms of another element only when the two planes of motion are parallel ; other-
wise, according to the size of the angle between the planes of motion, an atom of
one element may require two, three, or more atoms of another element to balance it ;
and only those components come into action which are parallel to the plane of motion
of another atom. Accordingly, F. M. Flavitzky refers the valency of an element to
the difference in the angles between the planes of the orbits of the different rotating
atoms. J. H. van't Hoff, in his Die Lagerung der Atome im Raume (Braunschweig,
1894), argued against the hypothesis which ascribed isomeric phenomena to the
varied motions of the atoms because temperature presumably favours atomic
motions, and yet the phenomena become less and less complex as the temperature
rises, and constantly more complex as the temperature falls.
References.
1 E. Erlenmeyer, Liehig's Ann., 131. 124, 1864 ; Zeit. Chem., 6. 65, 97, 609, 1863 ; 7. 1, 72,
628, 1864 ; W. Odling, Journ. Chem. Soc, 7. 1, 1855 ; A. C. Brown, On the Theory of Chemical
Combination, Edinburgh, 1861 (1879) ; A. W. von Hofmann, Introduction to Modern Chemistry,
London, 1865; G. Ensrud, Zeit. phys. Chem., 58. 257, 1907; L. Knorr, Liebig's Ann., 219,
202, 1894; J. Wislicenus, Ber., 21. 681. 1888; W. Lessen, ib., 20. 3306, 1887: Liebig's
Ann., 204, 336, 1880; A. Michaelis, Ber., 5. 411, 1872; Liebig's Ann., 315. 58, 1901 ; H. Davy, Phtl.
Trans., 97. 1, 1807 ; J. J. Berzelius, Schweigger's Journ., 6. 119, 1812 ; Essai sur la thiorie des
proportions chimiques et sur V influence chimique et electricite, Paris, 1819 ; M. Faraday, Phil.
Trans., 124. 77, 1834 ; V. Meyer and E. Riecke, Ber., 21. 946, 1888.
228 INORGANIC AND THEORETICAL CHEMISTRY
2 N. N. Beketoff, Ber., 13. 2404, 1880; F. M. Flavitzky, Zeit. anorg. Chem., 19. 201, 1896;
E. Molinari, Joum. prakt Chem., (2), 48. 113, 1893 ; L. Meyer, Die modernen Theorien der Chemie
und ihre Bedeutung fiir die chemische Mechanik, Breslau, 1884 ; London, 1888 ; F. P. Venable,
Joum. Amer. Chem. Soc., 21. 192, 220, 1899 ; F. M. Flavitzky, Zeit. anorg. Chem., 11. 264, 1896.
§ 18. Atomic, Molecular, and Specific Volumes
Modem developments in crystallography indicate with ever increasing distinctness
that the chemical atom even when its individuality is shrouded by combination with other
different atoms, exhibits characteristics which are essentially its own, and which are
discernible in the compounds into which it enters. — W. J. Pope and W. Barlow (1907).
So far as the balance can indicate, the weight, and by inference the mass, of
an atom remains uniformly constant during all chemical changes ; but the evidence
is less clear with respect to the volume or space occupied by the atoms of an element
when it enters into chemical combination. A. le Royer and J. B. A. Dumas i opened
up the subject in 1821 with an attempt to determine the equivalent volumes of
the elements by dividing their atomic weights by their respective specific gravities ;
the quotients were called atomic volumes.
The atomic volume of an element is obtained by dividing the atomic weight by its
specific gravity ; similarly the molecular volume represents the moleciilar weight divided
by the specific gravity. Consequently, the atomic volume represents the space occupied
by the aggregates of atoms, including the interstitial spaces, whose weights are proportional
to the atomic weight ; otherwise expressed, the volume occupied by a quantity of the
element proportional to the atomic weight. The term equivalent volume was used before
the concept of the atom had been clarified by Avogadro's hypothesis. At the suggestion
of J. J. Berzelius, H. SchrSder employed the term tnolecular volume in place of equivalent
volume ; and H. Kopp's term specific volume had the same cormotation. It has
been urged that the term specific volume is objectionable because the specific gravity of
a body is the weight of unit volume, and the term specific volume by analogy suggests the
volume of unit weight. The terms atomic volume and molecular volume here employed
are defined by the ratio
Atomic weight . , . , Molecular weight ,, ,
-^ r^ .^ = Atomic volume ; -5 r^ ?- — =Molecular volume.
Specific gravity Specific gravity
Consequently, if the atomic or molecular weight be expressed in grams, the atomic or
moleciilar volume respectively denotes the number of cubic centimetres occupied by a
gram-atom or gram-molecule.
It follows from Avogadro'shypothesisthatallgaseshave the same molecular volume.
If the centres of gravity of the molecules of liquids were situated at the same average
distance apart — as they probably are with gases — a given volume of different
liquid would contain the same number of molecules ; and the molecular weights of
different liquids would be proportional to the specific gravities — as is also probably
the case with gases. Similar remarks apply to solids. With liquids and solids,
however, the molecules must be located at different distances apart because the
molecular weights of different liquids and solids are not proportional to their specific
gravities. The molecular volumes of liquids and sohds do not exhibit the same
uniformity as those of gases. This might have been predicted from the fact
that while the coefficients of thermal expansion and the compressibilities of the
different gases are approximately the same, each solid and each liquid has its own
characteristic constant.
The molecular volume of gases can be compared at an arbitrarily defined standard
temperature and pressure ; but since liquids are obviously not in the same molecular
condition, they are therefore not under comparable conditions at any one arbitrarily
defined temperature. Consequently, H. Schroder^ suggested that liquids would be
more nearly in the same comparable state at the temperatures at which their
vapour pressures are the same — e.g. at their boihng points under a standard pressure.
In the case of solids, the effect of temperature is not so marked as with hquids, and
in the first approximation, the specific gravity is taken at a convenient atmosphere
temperature — say 0°, 4°, 15°, etc. A. Horstmann, W. Lessen, and A. Bartoli
COMBINATION BY VOLUME 229
contend that (i) while the so-called atomic volume refers not only to the space filled
by the atom, but also to the space in which the atom oscillates, it is not likely,
a priori, that the molecules will be in the same state at 1°, the boiHng point of butane,
as they are at 317°, the boiling point of octadecane ; (ii) relations similar to those estab-
lished at the boiHng temperature are likewise manifest at, say, the arbitrary tem-
perature 0° ; and (iii) the boiling point cannot be a strictly comparable state because
it is affected by pressure to a different extent in the case of different liquids.
G. Tschermak, F. Krafft, and G. le Bas take the melting point as a comparable
state. In a valid corresponding state, the pressure, temperature, and volume should
be expressed in terms of their critical values, and T. E. Thorpe has emphasized the
fact that C. M. Guldberg has shown that the ratio of the critical temperature Tc
to the absolute boiling point Tj approximates to a constant. Consequently, the boiling
temperatures are approximately equal fractions of the critical temperatures. Con-
sequently, properties like the molecular volume which change but slowly with
temperature, are comparable at the ordinary boiling points. The results by the
different methods do not show any very decisive evidence in favour of any one
method, since relations which are revealed by the one may be obscured by the
other. I. Traube emphasized the disturbing effects of molecular association and
claimed that this can be eliminated by determining the molecidar volume in dilute
solution. The idea was applied many years previously by L. Playfair and J. P.
Joule, who argued that " solution in water is the obvious means of destroying the
cohesion of a body without at the same time altering its chemical properties."
From their observations on atomic volumes A. le Koyer and J. B. A. Dumas tried
to show that the atomic volumes are multiples of one and the same number and
thus form an arithmetical series, but more extended investigations proved this
tentative hypothesis was not in accordance with fact. At this period, the chemical
combination of gases in volumetric proportions was attracting much attention, and
attempts were made to show that solids likewise unite in definite volumetric pro-
portions. For example, W. Herapath ^ tried to prove that the atomic volume of
oxygen in a metal oxide bears a simple numerical relation to that of the metal with
which it is combined. The same problem was attacked by C. J. B. Karsten (1832)
and by P. F. G. BouUay (1840). Here again, more accurate observations falsified
the hypothesis. F. Ammermiiller (1840) concluded from his observations that the
molecular volumes of compounds containing the same elements in different propor-
tions are either the same, or else stand to one another in rational proportions. J. F.
Persoz (1839) showed that equivalent amounts of many compounds of analogous
composition have the same molecular volume, and he tried unsuccessfully to
establish A. le Royer and J. B. A. Dumas' arithmetical rule.
H. Kopp's first publication, Ueber die Voraushestimmung des specifischen Gexvichts
einiger Klassen chemischer Verhindungen, appeared in 1839 and his last publication
on the subject was made in 1889. The earlier papers are mainly occupied in collect-
ing material and in finding the best conditions for comparing the data. In 1844
H. Kopp tentatively concluded :
(1) Equal differences in composition correspond with equal differences in specific
volume. (2) Equivalent amounts of oxygen and hydrogen in liquid compounds occupy
nearly the same volume. (3) The specific volume of a compound is equal to the sum
of the specific volumes of its components. The same element almost invariably preserves
the same specific volume. Isomeric compoimds have the same specific volumes which
stand to one another in the same relation as the molecular weights of the compounds.
Variations in the chemical constitution of isomeric compounds are without effect on their
specific volume. (4) Comparisons of specific volumes of liquids are only valid at tempera-
tures at which the vapour pressures of the liquids are equal.
H. Kopp considered that these conclusions did not rest on a very firm experimental
basis, and he therefore made accurate determinations of the physical constants
required for testing them rigorously. The results of this work enabled him
to take a general survey of the subject in his memoir, Beitrdge zur Stijchiometrie
230 INORGANIC AND THEORETICAL CHEMISTRY
der physikalischen Eigenschaften chemischer Verhindungen, 1855. His main
conclusions were :
(1) The selection of the temperature of equal vapour pressure as a basis of comparison
seems to be warranted by the fact that regularities are thereby made evident which
otherwise are not apparent. (2) Differences of specific volume are proportional to differ-
ences in chemical composition. (3) Isomeric liquids of the same chemical type have equal
specific volumes. (4) The substitution of hydrogen for an equivalent amount of oxygen
only slightly affects the specific volume. (5) One atom of carbon can replace two atoms
of hydrogen without altering the specific volume.
The molecular volumes of the members of a homologous series of liquids which
difier in composition by CH2 increase nearly 22 units for each increment of CH2.
Thus, the molecular volume of formic acid, H.COOH, is 41'3 ; of acetic acid,
CH3.COOH, 63-6 ; and of propionic acid, C2H5.COOH, 856. Hence, the mole-
cular volume of the group CH2 is 22. Further, the replacement of one atom of
carbon by 2 atoms of hydrogen in a compound usually makes no marked change
in the molecular volume, and hence it is inferred that the atomic volume of carbon
is nearly equal to the molecular volume of H2. Since the molecular volume of
CH2 is 22, it follows that the atomic volume of carbon is 11. The difference,
22 — 11=11, thus represents the molecular volume of H2, and the atomic volume of
hydrogen is 5'5. Again, the molecular volume of water is 18"8 ; deduct 11, the
value of H2, and the atomic volume of oxygen 7*8 remains. The molecular volumes of
a large number of compounds can be calculated from the data so obtained, and
compared with those obtained by actual experiment. The results for many carbon
compounds are quite satisfactory. Thus, with alcohol, C2H5OH, the molecular
volume will be (2 Xll)+(6 x5-5)+7-8=62-8. The observed value is 62-2. Hence,
if a compound contains ni atoms of atomic volume Aj ; 7^2 atoms of atomic
volume A2I . . . , the
Molecular volume, 'y=%i^2-|-W2^2"i" • • •
H. Kopp here over-emphasized the additive character of this property, but he
did point out that the specific volume of a liquid is determined not only by its com-
position but also by its constitution, for he found that the relative position of the
oxygen atom in a molecule affected the specific volume. The atomic volumes of the
oxygen atoms in carbonylic and hydroxy lie oxygen are respectively 122 and 7 "8.
The idea will be clear by comparing methyl alcohol, CH3OH, with formaldehyde,
H.COH, and with formic acid, H.COOH—
H>C<OH ^-^<l 0=^<0H
Methyl alcohol Formaldehyde Formic aoid
(Hydroxylic oxygen). (Carbonylic oxygen). (Hydroxylic and carbonylic oxygen).
By applying similar methods to those described above, it is found that the atomic
volume of carbonylic oxygen is 12*2. The molecular volume of methyl alcohol is
accordingly 4x5-5+ll+7-8=40-8 ; of formaldehyde, 2x5-5+ll+12-2=:34-2 ;
and of formic acid, 12-2+7'8+2x5*5+ll=42-0. Consequently, it is inferred that
one and the same atom may have different atomic volumes according to
the conditions under which it is placed. In further illustration, sexivalent
sulphur has an atomic volume 120 ; quadrivalent sulphur, 22*6 ; and bivalent
sulphur, 28*3. Nitrogen in ammonia and related compounds has an atomic volume
23 ; in cyanogen compounds, 28 ; and in nitroxyl compounds, 33. Hence, the
molecular volume can sometimes he used (l) for estimating the molecular weight of a
liquid from its specific gravity and cojwposition ; and (2) it may reveal peculiarities
in the constitution of the molecule. For instance, it may be used to show whether
carbonyhc or hydroxylic oxygen is present.
Examples.- — ^(1) The observed molecular volume of acetic acid, C2H4O2, is 63'7. The
only molecular voliime compatible with this is 64, deduced on the assumption that the
compound contains one hydroxylic oxygen atom (7-8), and one carbonylic oxygen (12-2).
COMBINATION BY VOLUME 231
The formula for acetic acid is therefore written CHj— CO — OH. (2) The density of
phosgene, COClg, at its boiling point, is 1-415. What is the atomic volume of chlorine, on
the assumption that the atomic volume of oxygen is 12-2; and of carbon U-0 ? Ansr.
99/1-415 = 12-2 + ll'0 + 2a; ; x, the required atomic volume, is therefore 23-4.
A large number of solid and liquid compounds — over a thousand — have been
examined. The pioneer work was done by H. Kopp and extended by many other
workers.4 With solids the data which have been accumulated are even more difficult
to deal with, since the disturbing factors seem to be even more perplexing than is
the case with hquids. Although many additive regularities have been detected
ranging over a limited number of compounds, yet, almost every investigation has
emphasized the constitutive nature of this property, and narrowed the range of
the simple additive rule. Even the increment CH2 in a homologous series, when
determined at the boiling points, is not additive, for its effect becomes greater as
the series is ascended, but, as A. Horstmann showed, the effect is not so marked
when the comparison is made at equal temperatures, or, as F. Krafft has shown, at
the melting points. H. Kopp thought that isomerides of similar structure had the
same molecular volume, but P. Dobriner and R. Gartenmeister have shown that
the effect is related with the boiling points, for the lower boiling isomer has the
larger molecular volume ; J. C. Brown, A. Zander, T. E. Thorpe, and W. Stadel
showed that isomers with an *so-structure also have the larger molecular volume.
F. Neubeck showed that the molecular volume of the benzene derivatives is modified
according as the groups occupy the ortho-, para-, or meta-position. P. Walden and
T.Liebisch found that the race mic isomer of stereo-isomerides has the smaller mole-
cular volume, and I. Traube that the trans-isomer has the greater molecular volume.
H. L. Buff also showed that the atomic volume of an element varies according to
its degree of saturation, and that an unsaturated carbon atom has a larger atomic
volume than a saturated one. Hence, argued H. L. Buff, the atomic volume of
an element decreases as saturation proceeds. This was confirmed by R. Schiff and
W. Lossen, who found that on passing from a saturated carbon atom to one with the
double ethylene linkage, the molecular volume decreases about 8'5 units, and on
passing from the double ethylene to the triple acetylene linkage, the molecular volume
decreases about 6*5 units. In a homologous series of ethylene linkages, the effect
produced by each is rather less than the preceding one. A. Horstmann found that
" unsaturated compounds with closed chain formulae have considerably smaller
molecular volumes than those with open chain formulae and multiple hnkages of
the atoms." R. Willstatter showed that the contraction in the molecular volume
which accompanies the conversion of a normal chain hydrocarbon to the ring or
cycloid structure is larger than is caused in passing from a saturated to an un-
saturated compound. Consequently, molecular volumes are dependent upon
differences in the structure of the compound as well as on the nature of the atoms
in the molecule.
The difference between the molecular volumes of the MO oxides and the atomic
volume of M gives fairly constant values for the atomic volume of oxygen, but in
other cases very different values are obtained. Thus, the oxygen in cupric oxide
has an atomic volume 5"1, and in cuprous oxide 10*5. B. Brauner and J. I. Watts
have investigated the atomic volumes of the oxides, and found the results are in
accord with the periodic law, and conclude :
(1) In strong bases the oxygen has a negative value. (2) In the oxides of heavy metals
and metalloids the volume of the oxygen is positive. (3) The earth metals unite with
oxygen without any appreciable change of volume, and thus form a connecting link between
acids and bases. (4) The higher the specific volume of the element in the oxide, the less
positive or more negative is the specific volume of the oxygen. (5) The more negative the
value of the oxygen, the greater is the afifinity of the metal for the oxygen.
L. Play fair and J. P. Joule ^ noted that the molecular volumes of certain highly
hydrated salts— e.^. sodium decaquocarbonate, and the alkali dodecaquophosphates
and dodecaquoarsenates — are exactly equal to that of the water, considered as ice,
232 INOKGANIC AND THEOEETICAL CHEMISTRY
which they respectively contain, so that the molecules of the salt proper seem to exist
in the interstitial spaces of the water since they exert no apparent influence on the
bulk. The relation does not hold with salts less highly hydrated — e.g. borax,
sodium pyrophosphate, and aluminium sulphate — where the molecular volume is
the joint effect of the water considered as ice, and of the salt. R. Schiff also showed
that the members of certain classes of hydrated salts have practically the same
molecular volume — e.g. the alums have a molecular volume of about 277 ; the
double sulphates of the type M2'M"(S04)2.6H20 have a common molecular volume
of about 207 ; and the vitriols of the tjrpe M"S04.7H20, isomorphous or not, have
the same molecular volume 146. T. E. Thorpe and J. I. Watts have further shown
that the volumes occupied by the several molecules of water in polyhydrates vary
with the degree of hydration, for the molecular volumes of hydrated salts are not
usually equal to the sum of the molecular volumes of the anhydrous salt and of the
water (18"8). With the magnesium sulphates, for example,
MgSOi pliis ... 0 1 2 5 6 7 HgO
Molecular volume . 45-3 55-6 67-0 112-4 130-8 1464
The first molecule of water, the constitutional water or the water of halhydration of
T. Graham, here occupies a less volume than the remaining molecules. The second
molecule of water raised the molecular volume 11 "4 ; the next three molecules of
water raise the molecular volume an average of 11*8 ; the sixth molecule raises the
constant ]8*4, and the seventh, 15'6. T. E. Thorpe and J. I. Watts obtained
analogous results with the series of sulphates MSO4.WH2O, when n varied from 0 to 6.
This is in harmony with H. Kopp's general conclusion that the water molecules of
a hydrated salt contribute in different degrees to the total molecular volume, for in
salts containing a small number of water molecules (1 to 3), he found the average
molecular volume of the water is 12*4 ; in others containing a larger proportion
(2 to 7), the average molecular volume is 13"4 ; and in a third class, with the largest
proportion of water molecules (3 to 10), the average molecular volume is 15'3.
F. W. Clarke compared similarly the molecular volumes of a series of chlorides
MCI2.WH2O, when n varied from 2 to 6 ; and for a series of hydrated oxides —
B2O3.3H2O ; I2O5.H2O ; K2O.H2O ; CuO.HgO ; SrO.H20 ; BaO.HgO ; AI2O3.3H2O ;
Mn203.H20 ; Fe203.H20. In the former, the molecular volume of the water
varied from 12*5 to 15*0, and in the latter from 7*4 to 19-4. F. W. Clarke's
results emphasize the difference between water of crystallization and water of con-
stitution in that the chemical differences implied by these expressions are connected
with the relative magnitudes of the spaces occupied by chemically comparable
quantities of the hydrated salts. The contraction which occurs in the dilution of
sulphuric acid with water is indicated in Fig. 27, Cap. X.
The atomic volume of an element obtained by dividing atomic weight by its
specific gravity is not the same as the atomic volume deduced by H. Kopp from the
molecular volumes when the element is in combination. The two values are not
usually the same. For instance,
H. Kopp, atomic volume
Calculated from element
In 1831, T. Thomson compiled a table of atomic volumes of the metals, and noted
a correspondence in the atomic volume of the elements most nearly related with
one another. When the atomic volumes are plotted against the atomic weights,
L. Meyer 6 showed in 1869 that a periodic curve is obtained like Fig. 4 in Cap. VI,
where (1) the waves increase in amplitude as the atomic weights increase ; (2) the
elements of similar chemical properties occupy corresponding positions on the
waves ; (3) the more volatile and easily fusible elements occur on the crests or rising
portions of the curve, and the elements which fuse with difficulty are in the troughs
or on the descending portions of the curve. The curve was found by W. Borchers
to be more regular and the relations between the elements clearer if the equivalent
H.
C.
CI.
Br.
I.
0.
3-6
11
22-8
27-8
37-5
7-8-12-2
11
3-4
22-2
25-1
25-6
14-3
COMBINATION BY VOLUME 233
volume — atomic weight -f- maximum valency — be employed in place of the atomic
volumes.
The molecular volume of an element varies with the conditions under which the
molecules are placed. The atom is presumably always in oscillatory periodic
motion, and this motion gives rise to volume ; consequently, the molecular volume
is a relative measure of the space inhabited by the molecule ; it represents the
smallest space which the molecule requires for itself under the existing conditions.
Similar remarks apply to the atomic volume so that each atom can be regarded
as a material nucleus surrounded by an envelope, shell, or space — called the
sphere of action or sphere of influence into which no other atom or mole-
cule can penetrate. The sphere of influence is thus regarded as the effective
boundary surface of an atom. This is what is sometimes called the vibratory or
oscillatory volume of an atom, that is, the space within which the material nucleus
performs its oscillations. Such a space would have the quasi-rigidity characteristic
of a material nucleus rapidly revolving about a mean position. There is, however,
no need to make any assumptions as to the nature of the internal character of the
atomic nucleus with its encircling shell ; it is not even necessary to assume that the
complex is spherical. Under ordinary circumstances the complex can be regarded
as the atom itself, since the so-called sphere of influence is the actual boundary by
which we know and measure the behaviour of the atom. This is the concept of the
atomic volume as pictured by D. I. Mendeleeff (1889), 0. E. Meyer (1899), T. W.
Richards (1901), etc.
In J. D. van der Waals' equation {p-\-av~^){v—h)=RT, the term b represents
the volume occupied by the substance, i.e. the molecular volume at absolute zero,
since at this temperature v becomes equal to b and represents the volume occupied by
the substance of the atoms — ^the atom nucleus as it may be called — but it is said
to be four times the actual volume of the molecule. It is not practicable to
compare the values of b for different substances because of the lack of data ; but
from the theory of corresponding states, it may be shown that the critical volume Vg is
three times the value of 6, or Vc=Sb ; and the so-called critical coe£&cient, Tdpc,
or, the ratio of the critical temperature and critical pressure, is related to b by
the expression Tclpc=^hlR, where R is the gas constant ^v/273.
F. Exner ^ showed that, according to R. Clausius and 0. F. Mossotti, (jit^— l)/(jLt2+2)
is equal to the ratio of the volume actually occupied by matter to the apparent
volume of the substance, when /ju represents the refractive index for waves of infinite
wave length, and it is found that fi^ is equal to the dielectric constant. Consequently,
as P. A. Guye has shown, the product of (/x^— 1)/(jlc2+2) with the molecular volume
will be a measure of the space filled by matter in a gram-molecule of a substance.
Consequently, the magnitude 6 of J. D. van der Waals' equation, the critical volume,
and the critical coefficient may be represented as functions of the molecular refrac-
tion. I. Traube has shown that b is between 3'5 and 4 times as large as the molecular
refraction, MR, and P. A. Guye found therelation MR=18TclPc' From I. Traube's
result, it follows that atomic refraction can be employed as a measure of the material
nucleus of an atom composed of a material nucleus and an encircling shell or
sphere of influence. The sphere of influence represents a kind of shell about the
atom nucleus, and it is presumably that portion of the atom which is permeable to
light, and constitutes a dielectric medium which enables electromagnetic radiations
to be transmitted through a body at a speed which is characteristic of the particular
substance.
T. W. Richards 8 has shown that while it is assumed that the molecules of
a gas are particles moving independently at some distance apart, it is doubtful
if there are such interstitial spaces in liquids and solids. The impermeability of
glass to oxygen, nitrogen, and water for long periods does not lend support to the
view that there are empty spaces between the molecules ; and he limits the per-
meability which has been observed in rigid compact solids, to such substances as
can enter into the chemical structure of the soUds themselves. It is therefore
234
INORGANIC AND THEORETICAL CHEMISTRY
inferred that in solids and liquids the atoms formed of material nuclei with enveloping
shell are in close contact with one another at the boundary of their envelopes.
Such atoms are considered to be compressible and elastic ; they can contract and
expand, or vibrate among themselves even when their surfaces are closely packed
together ; and they are quite capable of sustaining and transmitting the vibratory
motions called heat. E. Griineisen's observations show that the compressibility
of a number of metals — copper, silver, aluminium, iron, and platinum — loses only
7 per cent, in cooling from the ordinary temperature down to that of liquid air, and
by extrapolation very little more diminution will occur in passing down to the
temperature of absolute zero, so that it is probable that metals are as compressible
at absolute zero as they are at ordinary temperatures. The value of 6 in J. D. van
der Waals' equation is fairly constant over a wide range of pressure, but it suddenly
begins to diminish when very high pressures have been attained. Again, for carbon
dioxide, the value of h increases as the temperature rises, thus, if a=000874, and
v=unity at 0° and 760 mm., then 6=51 c.c. per gram-molecule at 0°, and 64 c.c.
at 100°. For hydrogen likewise, h is 13'8 c.c. at 0° and 15' 1 c.c. at 100°. On the
other hand, H. K. Onnes has shown that the apparent volume h of helium atoms is
smaller at high than at low temperatures ; at 0°, h is 12 c.c, and at 100°, 10*4 c.c.
per 4 grms. of helium. This is not what would be anticipated, and T. W. Richards
1
Cs!
\
1
Rb
5
s
1
K
1
1
'
I
1
1
r
j
',
''
!
..J /
1
Li.
r
//
,-'
J J
1
X
(
1
t
He
s\j
V
Ni
\
Pd
)
/J/bmic u/e/^hts — >-
Fig. 4. — Compressibility and Atomic Volume Curves of the Elements.
makes the tentative suggestion that " the greater velocity of the colliding atoms
at the higher temperature has a greater compressing effect so that at high tem-
peratures the atoms seem to occupy less space than at lower temperatures."
T. W. Richards has shown that the compressibilities of the elements — i.e. the
relative contractions in volume per megabar (0'987 atm.) per sq. cm. — are closely
related to their atomic volumes, for the structure of the two curves is very similar as
indicated in Fig. 4, where the atomic volume curve is dotted, and the compressibility
curve is continuous. The greater the densities of the elements the less their
compressibility. The elements with the larger atomic volumes are the more com-
pressible and the more easily melted and volatilized. Consequently, the com-
pressibilities of different substances are not only dependent on the magnitude of
the applied pressure, but also on the internal pressure produced by the mutual
cohesive attraction between the particles. In gases, the cohesive pressure is small,
and accordingly the compressibility is large ; in solids and liquids the cohesive
pressure is large, and the compressibility is small. T. W. Richards' theory of
compressible atoms thus reveals the existence of internal cohesive and affinity
pressures holding the atoms and molecules together.
Cohesive pressure exerted by the cohesive forces which hind the molecules together. —
Cohesion manifests itself in various ways ; the most obvious is the mechanical
resistance which a body offers to the separation of one part of a substance from
another, and it appears more or less modified in such properties as ductility,
COMBINATION BY VOLUME 236
malleability, tenacity, hardness, surface tension, volatility, etc. The density of a
given substance is a manifestation of an internal pressure — the greater the density,
the greater the internal pressure. When a rise of temperature produces a marked
effect on the volume, it may be assumed that the internal pressure is less than when
a rise of temperature produces only a slight effect on the volume. Substances in
which the particles are held together by a high cohesive attraction are usually
difficult to volatilize ; they have a small atomic volume ; a relatively large density ;
a high surface tension ; high latent heat of evaporation ; and are least compressible.
Conversely, if the atomic volume be relatively large owing to a small cohesive
attraction, these substances will be most volatile and have the greatest com-
pressibility. In illustration, T. W. Kichards found that the three isomeric xylenes
agree well with these deductions :
o-Xylene. m-Xylene. p-Xylene.
Density (20°) . . . 0-8811 0*8658 0-8611
Boiling point . . . 144-0° 139*0° 136*2°
Compressibility (20°) . . 60 X 10-« 63*5 X 10-« 66*2 X 10-«
Surface tension (20°) . . 3*09 2*96 2*92 mgrm. per mm.
T. W. Richards also compared some properties of two isomeric butyric esters, and
found in each case :
Specific Compressi- comnrSmtv Coefficient of Boiling Surface ^^^^l
gravity. bility. 'p?Satm expansion. point. tension. ^^n.
Ethyl
butyrate 0*8785 76-9xlO-« 13-6 X 10-« 0-001247 120*8° 24-58 34-7
Ethyl iso-
butyrate 08710 90-8 X 10-« 15-0x10"^ 0*001294 109*8° 23*30 33*9
The denser substance has the less compressibility, the less decrease in com-
pressibility with an increase of pressure, the less coefficient of thermal expansion,
the higher boiling point, the greater surface tension, and the greater heat of
vaporization. All this is in accord with the assumption that a great cohesion
produces an internal pressure which is effective in reducing the molecular volume.
Hence it follows that not only is the atomic volume dependent upon the nature
and location of the different atoms in the molecule, but also on the cohesive attraction
of one molecule for another.
If the atomic volume be related with the cohesive pressure, and if the valency of an
element in a compound be related with the atomic volume, it might be anticipated that
th9 cohesion will be a function of the valencies of the combined elements. W. Sutherland
f oimd that the valencies of the elements in simple substances like sodium chloride influenced
the cohesion, but he was unable to establish a relationship for more complex bodies. A. P.
Mathews found empirically that the cohesion factor a of J. D. van der Waals' equation is
related with Uv, the number of valencies, and the molecular weight M by the expression
a=0-3l25xlO^^MI!v ; and from the known relations of a to the critical constants T^, v,,
and pc, it foUows that Mp^j:v=0-00^3Tc^ ; and also that Mi:v=4:-3xlO-^{VcTc)',
A. P. Mathews uses these expressions for computing the valencies of chlorine, oxygen,
sulphur, nitrogen, phosphorus, and the elements of the argon family.
Affinity fressure, produced by the chemical affinity or mutual attraction of adjacent
atoms. — The atomic volume of an element depends on the nature of the associated
atoms. In the middle of the eighteenth centur}% R. Kirwan sought to measure
the force of the attraction between the atoms in a chemical compound from the
diminution in volume which attended the union of two substances ; while H. Davy
and others have alluded to the increase in density of the product of the union of
two substances with a powerful affinity for one another. Thus, W. Miiller-Erzbach
(1881) said that in similarly constituted solids, those are the most stable which are
formed with the greatest contraction— e.^r. when lead replaces silver ; potassium,
sodium ; or, when chlorine replaces bromine or iodine, contraction occurs, and the
products of the replacement are the more stable. According to F. Ephraim and
P. Wagner, the molecular volume of a stable compound is smaller than the sum of
236
INORGANIC AND THEORETICAL CHEMISTRY
the volumes of its decomposition products, as shown by the schonites studied by
A. E. H. Tutton. The double alkali magnesium sulphates have smaller molecular
volumes than the copper or manganese salts, although the atomic volume of the
metal is greater with the first than with the other two. The stability of the salts
is more strictly parallel to their molecular volumes than to the atomic volumes of
the free metals. The percentage contraction on the atomic volume rather than the
actual contraction suffered by any particular atom is thus the important criterion
of the stability of a compound.
D. I. Mendeleeff has shown that the greater the affinity of the elements for one
another, the less the atomic volume of the resulting compound. Thus the contrac-
tion which occurs during the formation of potassium or sodium oxide is greater than
in the formation of stannic oxide. As V. Braun expressed it, the specific gravity of
solid chemical compounds is high in proportion to the intensity of the affinity which
unites their components, and G. S. Johnson inferred that the affinity of iodine in
potassium tri-iodide is small because combination occurs without contraction.
W. Miiller-Erzbach and I. Traube have emphasized the same idea. Other things
being equal, elements with the greatest densities have the least chemical
affinity and sufiei least change in their atomic volumes when they enter
into combination — these elements will be found distributed about the troughs
of the curve, Fig. 4. Conversely, elements with the smallest densities are
usually most energetic chemically and suffer the greatest contraction in
their atomic volumes when they form chemical compounds — these elements
will be found distributed about the peaks or crests of the curve. Fig. 4.
The fact that the less the density, or the greater the compressibility of an
element, the greater its contraction on combination, is best illustrated by
comparing the contractions occurring during the formation of similar compounds
of the elements having widely different compressibilities but similar affinities.
Strontium and lead are not very different in cohesive pressure as shown by the
closeness of their boiling points. When a gram-atom of strontium unites with a
gram-molecule of chlorine, there is a contraction of 32*6 c.c, and lead gives under
similar conditions a contraction of 20' 1 c.c. If affinity causes contraction this is
just what would be anticipated because the affinity of strontium for chlorine is
greater than that of lead ; this is confirmed by the respective heats of formation
772 and 346 kilojoules per gram-molecule. The compressibilities of the elements
of the alkah metals and the contractions which occur, in c.c. per gram-molecule,
when the corresponding chloride is formed, are as follows :
Table IX.
Element.
Compressibility.
Contraction
(C.c. per gram-molecule).
Lithium ....
Sodium ....
Potassium ....
Rubidium ....
Caesium ....
90xl0-«
15-6
31-7
40-0
61-0
17-6
21-5
33-1
36-8
53-6
Not all examples will give such unequivocal evidence of the effect of chemical
affinity in determining the atomic volumes of liquids and solids, because the effects
of chemical affinity will be modified or even overshadowed by the effects of cohesion.
Both must always be present, and it may be difficult to discriminate between the
two effects. Similarly, the heat of formation Q of a compound runs parallel with
the free energy, and may be regarded as proportional to the work done by the
affinity pressure between two elements. Similarly, when two elements unite, the
contraction A is evidence of the affinity uniting the elements. The contraction A
COMBINATION BY VOLUME
237
is the difference between the molecular volume of the compound, and the sum of
the atomic volumes of the component elements in a free state. The quotient of
the volume contraction. A, by the heat of formation, Q, will give a measure of
the average compressibility. In Table X, T. W. Richards showed that the com-
pressibility effect with the alkali halides must be the same in each member of the
series, and the values of A/Q for the different salts should fall in the same order
of magnitude as the compressibilities of the free alkali metals if the hypothesis
relating affinity pressure to compressibihty be correct. This is actually the
case.
Table X.-— Compressibilities and Affiiuity Pressures of the Alkali Halides.
Salt.
Compressibility
of metal.
Sum of
atomic
volumes.
Molecular
volume.
Contraction
Heat of
formation
Q.
LiCl .
NaCl .
KCl .
8-8x10 «
13-4
31-5
37-7
48-7
70-0
20-9
27-2
37-8
16-8 j 383
21-5 399
32-2 427
4-4
5-4
7-6
LiBr .
NaBr .
KBr .
8-8
13-4
31-5
38-8
49-2
70-5
25-2
34-2
44-2
130 334
150 j 359
26-3 398
3-9
4-2
6-6
Lil .
Nal .
KI
8-8
13-4
31-5
38-4
49-4
70-7
331
41-4
53-8
5-3 1 257
8-0 ' 289
16-9 1 335
!
21
2-8
51
In the light of T. W. Richards' hypothesis, we can also see that for H. Kopp's
rule to be vaHd the internal pressures of all compounds at their boiUng points should
be the same — subject to small variations due to differences in molecular complexity.
The intense intermolecular pressures under which the molecules exist modify the
boiling points, the surface tensions, the viscosities, etc. The effects produced by
cohesive and affinity pressures on atoms with an elastic compressible envelope,
as postulated by T. W. Richards, show that the volume occupied by an atom in the
free state cannot be the same as in the combined state, and that the volume of an
atom in combination will vary with the nature and orientation of the other atoms
with which it is combined.
H. Schroder 9 worked on the subject of atomic volumes simultaneously with
H. Kopp ; he accepted F. Ammermiiller's conclusion that equal volumes of the two
oxides of copper contain the same amounts of copper, and multiple amounts of oxygen,
and assumed that in the two compounds with the atomic volumes : CuaO =24*36 ;
and CuO =12-35 or Cu2O2=24*70, the quantities of copper are the same, and that
the volume of the copper is in each case the same. The volume occupied by the
oxygen in cuprous oxide then stands to that in cupric oxide as 1 : 2. H. Schroder
concluded with H. Kopp that the molecular volume of a compound is the sum of
the volumes of the component atoms. The former considered that the atomic
volume of a given element under similar structural conditions throughout all its
compounds is variable — the latter assumed that under these conditions the atomic
volume of an element is constant. The different atomic volumes which an element
can assume in different compounds were regarded by H. Schroder to be simple
multiples of a certain unit volume which he called the stere. The stere is not the
same for all elements, but it varies within comparatively narrow Umits. When two
elements are combined, one of them assumes the unit volume of the other, so that
the stere of one element dominates the volume of the compound, and the molecular
volume of the compound may be represented as a simple multiple of the stere of
one of the contained elements. For example, the stere of silver is 5-14, and the
Stere value.
Molecular volume.
Calculated. Observed,
. 5-14x6 =
30-8
30-8
. 5-14x5 =
25-7
25-8
. 5-14x6 =
30-8
30-8
. 5-14x8 =
411
41-8
238 INORGANIC AND THEORETICAL CHEMISTRY
atomic value is twice this, namely, 10'28, so that metallic silver occupies
2 steres.
Silver oxide
Silver chloride
Silver bromide
SQver iodide
Hence, the atomic volume of oxygen is J(6— 2)=2 silver steres ; of chlorine, 5—2=3
silver steres ; of bromine, 6—2=4 silver steres ; and of iodine, 8—2=6 silver steres.
The general conclusion is that the volumes of equivalents of difEerent elements
are approximately equal, or stand in some simple relation with one another. This
naturally raises the question whether there is any connection between the valency
of the atoms and its effect on the molecular volume. In a general way, an increase
in the valency of an atom is attended by an increase in the molecular volume,
although, as W. Stadel has shown, the molecular volume is influenced by all the
atoms in the molecule. G. le Bas i° compared the molecular volumes of eighteen
hydrocarbons of the paraffin series at their melting points, and found that the
quotient, obtained by dividing the molecular volume by the total number of valencies
of the carbon and hydrogen atoms present, is a constant — very nearly 2' 97. In
illustration, the molecular volume of dodecane, C12H26, is 2199 ; there are 12x4
carbon valencies, and 26 hydrogen valencies, or a total of 74 valencies ; consequently
219'9-h74=2'971. The constant 2*97 thus represents one unit of valency in these
compounds.
I. Traube ^i has investigated molecular volumes from a novel point of view.
He takes the specific gravity at ordinary temperatures — usually about 15° — and
he also allows for the association of the substance. I. Traube defines the molecular
solution volume, V^i of a substance in water by the relation
Molecular solution volume, V^=i -— ^- . — — —
Ug JJw
where M denotes the molecular weight of the dissolved substance, Dg and Dy, are the
specific gravities respectively of the solution and of water, and A denotes the number
of grams of water in which a gram-molecule of the substance is dissolved. If v
denotes the ordinary molecular volume, defined by MjD, the difference v~Vm
denotes the contraction which occurs in the process of solution, and is called the
molecular contraction, and it is found that if ionization and association effects are
eliminated, the molecular contraction has the constant value v— F^=135 c.c. per
gram-molecule.
If ionization occurs, the number of ions which in their action are equivalent to non-
ionized molecules must be taken into consideration. If a denotes the degree of ionization
of a solute decomposed into n ions, then, in place of 13 '5 c.c, the molecular contraction
= 13-5{l-t-(w — l)a}. If association occurs, the association factor must be considered.
The association factor is a number which represents how many times the molecular weight
of a substance is greater than corresponds with the simple gaseous molecule. In place of
13*5 c.c. the molecular contraction is 13-5/)8 c.c. per gram -molecule.
The molecular solution volume can be calculated from the volume constants
of the constituent elements according to H. Kopp's additive rule, and the intro-
duction of a correction factor. I. Traube found that this correction factor is a
constant 12*4 c.c, so that if n^, ^2, . . . denotes the respective number of atoms of
atomic volume Ai, A2, • . • ; and 2/71^=^2^1+^2^2+ • • •
Molecular solution volume, V^^=EnA-\-\2'^
The term SnA also includes a correction term for multiple bonds, etc.
By empirical calculation from the observed molecular volumes, I. Traube has computed
the solution volume constants in c.c. per gram-atom for different radicles. He obtains :
C, 9-9; H, 31; F, 65; CI, 13-2; Br, 17-7; I, 21-4; CN, 132; Na, 31; N"i, 16;
COMBINATION BY VOLUME 239
Nv 10-7; P"i, 17; pv 28-5; double bonds, -17; triple bonds, -3-4; Hydroxylic
oxygen (OH), 2-3 ; Hydrosulphylic sulphur (HS), 15-6. Oxygen atoms united to carbon
by a double bond, 5-5 ; sulphur atoms united to carbon by a double bond, 15-5; oxygen
atoms in a carbonyl group, or imited to a carbon atom with a hydroxyl group attached
to it, 0-4. The observed density of ether, (C2H5)20, is 0-7201 at 15° ; the molecular
weight is 74. Compare the observed and calculated molecular volume. The observed
value is 74/0-7201 = 102-7 ; the calculated value is (4 x9-9)+(10 x3-l) + 5-5+26'9=102-0.
If as before v denotes the molecular volume, and if there is no ionization,
V— F^=13'5/y4, then v=I!nV+12-4:-\-lS-5ip, where the association factor j3 is
usually nearly unity — e.g. with phosphorus trichloride, and carbon and sihcon tetra-
chlorides, j8=unity ; for benzene, j3=ri8 ; for toluene, 1-08, etc. — but with water
^=3'06 ; formic acid, 18 ; acetic acid, 1'56 ; methvl alcohol, 1'79 ; ethyl alcohol,
1-67, etc. When j8 is unity Vrn=EnA-\-\2-^ ; and'
Molecular volume, v=SnA-\-2b'^
meaning, according to I. Traube, that " in the formation of any molecule from its
atoms there is always a dilation ; the molecular dilation is the same or nearly the
same for all substances ; it is independent of the chemical nature of the substance
and can be only slightly modified by constitution ; and at 15°, the molecular solution
volume in aqueous solution is 12*4 c.c. per gram-molecule, and the molecular volume
25'9 c.c. per gram-molecule." Given the molecular volume it is possible to calculate
the association factor which may or may not agree with that deduced by other
methods.
I. Traube regards En A as the sum of the spaces occupied by the matter of the
atoms of a molecule. While the internal or nuclear volume of a molecule is the
space actually filled by the mass of the atom, the external volume is the nuclear
volume increased by the volume of a shell of combined sether. The external atomic
volume corresponds with the magnitude h of J. D. van der Waals' equation, and
is 3*5 to 4 times as large as the internal or nuclear volume. The difference Vrn—EnA
gives what I. Traube calls the molecular CO- volume— symbolized Cor. The
co-volume is a magnitude dependent on the temperature ; for 15°, the molecular
co-volume is 259 c.c, and at B°, the molecular co-volume is Cot.„ (1+0*003670),
or
the
Molecular co- volume, Co„=:24-5(l +0-003670)
very nearly. There is a close formal analogy between the temperature effect of the
CO- volume and the volume of gas. Since, for every newly formed gram-molecule
there is an expansion equal to the co-volume, and for every molecule which dis-
appears there is a corresponding contraction, I. Traube concluded : " In a reaction
between homogeneous liquids, the co-volumes of the initial and final products of
the reaction stand in a simple rational ratio " — this is J. A. C. Charles' law applied
to liquids. Since also I. Traube assumed that the molecular volume is the sum of
the true molecular volume and the co-volume, Avogadro's rule apphed to Hquids
becomes " with the same conditions of temperature and pressure, the co-volumes,
or the volumes in which the molecules move, are all equally great."
I. Traube's method can be employed for calculating the molecular volume, and also
the molecular weight M of an unknown substance of known specific gravity D. In
this case, since v=M/D, the chemical formula which gives the closest value to
M
^-UnA=26'9
is the desired chemical formula. Many examples will be found in H. Biltz's Die
Praxis der Molekulargewichtsbestimmung (Berlin, 1898 ; Easton, Pa., 1899).
The observed specific gravity of tetrachloroethane is 1-6258 (15°), the empirical formula
by analysis is CHCL. Hence, if the formvda CHClj obtains, the ratio M/D = 516 and
2nA=3d% or M/D -ZnA = 12-2 ; if the formula be C2H2CI,, M/Z) -27*7^ =103-2-78-8
240 INORGANIC AND THEORETICAL CHEMISTRY
=24-4; and if the formula be C3H3CI8, M/D-Zw^ =164-8- 118-2=36-6. Here then
24*4 approximates closest to the theoretical co-volume 26'9, and the formula is accordingly
C2H2CI4.
According to I. Traube, the atomic CO-volume is the difference between the
internal and external atomic volumes ; and it represents the volume of the com-
bined aether. I. Traube further postulates that the atomic co- volume is occupied by the
valency electrons, i.e. the electrons which endow the atom with valency ; for, unHke
the molecular co-volume, the atomic co-volume varies in size and is proportional to
the nuclear volume and the valency of the atom. I. Traube employed molecular
refraction as a measure of the nuclear volumes of the atoms in a compound and found
that the molecular refractive power, MR, of a saturated compound is proportional
to the total number of valencies, n, of the component atoms. The value of the
quotient MRjn for a large number of compounds deviates but little from the mean
0"787. I. Traube calls 0*787 the refraction stere — in illustration, the molecular
refraction of alcohol, C2H5OH, is 12-71, and n is 8+5+2+1=16 ; and 12-71-M6
=0'794:. The nuclear volumes of the atoms in a molecule are therefore proportional
to the valencies of the atoms.
W. C. Roberts-Austen 12 suggested that the remarkable influence of traces of
elements on masses of metals is proportional to the atomic volumes of the con-
taminant. He showed that the metals or metalloids near the troughs of L. Meyer's
periodic curve. Fig. 4, Cap. VI, do not diminish the tensile strength of gold ;
and that the metals which render gold fragile occupy high positions on the curve.
Hence he argues :
There is some relation between the influence exerted by the metallic and other im-
purities and either their atomic weights or their atomic volumes. It seems hardly probable
that it is due to atomic weight, because copper, with an atomic weight of 63-2, has nearly
the same infl\ience on the tenacity of pure gold as rhodium, with an atonlic weight of 104,
or as aluminium, the atomic weight of which is 27-0. It will be evident from the following
table, which embodies the results of the author's experiments, that metals which diminish
the tenacity and extensibility of gold have high atomic volumes, while those which increase
those properties have either the same atomic volume as gold, or a lower one. Fiu-ther,
silver has the same atomic volume as gold, 10 "2, and its presence in small quantities has
very little influence, one way or the other, on the tenacity or extensibility of gold.
It is suggested that the atoms with a small atomic volume can fill up interstitial
spaces which would otherwise remain void and this without disturbing the dis-
position of the other atoms, while atoms with a large atomic volume act prejudicially
by driving the atoms further asunder. The following experiment by E. Warburg
and F. Tegetmeier illustrates a porosity in solids which will permit the passage of
elements with a small atomic volume, but strain off those with a larger atomic
volume.
A cell with a glass partition with sodium amalgam about the anode and mercury about
the cathode was heated to between 100° and 200°- — when the glass became slightly con-
ducting. In about 30 hrs. an appreciable quantity of sodium had passed from the glass
into the mercury. The glass remained transparent, for the sodium lost by the glass was
replaced by that from the mercury amalgam. W. C. Roberts -Austen showed that in the
electrolysis of the glass, the passage of the sodium follows the ordinary law of electrolysis.
If lithium amalgam be used, the glass becomes opaque, and then lithium acciunulates in
the merciuy. The glass loses no potassium, but 7 '8 per cent, of sodium, and gains 4*3
per cent, of lithiiim. With potassium amalgam, the potassium does not replace the sodium
lost by the glass. It is suggested that the lithium atoms with an atomic volume 15*98 can
replace sodium atoms with an atomic volume 16*04, while potassium atoms with an atomic
volume 24 are too large to take the place of the smaller sodium atoms. The glass diaphragm
has thus been said to act as a mechanical sifter for the potassium atoms.
The simple relation between atomic volumes and tenacity is no doubt modified
when compounds are formed. F. Osmond also showed that elements with a smaller
atomic volume than iron retard the transformation of j3 to a iron, while elements
with a larger atomic volume than iron either have no influence upon the transition
temperature or else raise that temperature.
Several investigators have traced the influence of the atomic volume of a metal
COMBINATION BY VOLUME 241
on the mechanical properties ; for example, A. Wertheim and H. Tomlinson have
shown that there is a relation between the atomic volume and elasticity ; W. Suther-
land, between the atomic volume and rigidity ; R. A. Fessenden, between atomic
volume and cohesion ; and H. Crompton, between the latent heat of fusion and the
molecular volume of a compound. H. Crompton also showed that the molecular
heat of fusion L ; the absolute fusion temperatures T ; and the valencies n
of the elements, are so related that LjTn is a constant ; and that a similar
rule holds for compounds. From I. Traube's relation between valency and
atomic volume, it therefore follows that the latent heats of fusion are a func-
tion of the molecular volumes, or of the sum of the atomic volumes. Hence also
the latent heat of fusion must be a function of the sum of the valencies of the
atoms in a molecule.
W. Barlow and W. J. Pope 13 assume that each atom of a crystalline compound
occupies a definite space or sphere of influence, so that each atom has its own
polyhedral cell or domain throughout which its influence is predominant ; that
the space occupied by a substance is partitioned into atomic domains of definite
volume ; that the mode of arrangement of such atoms determines the shape
of the molecule ; that the atoms are held in place in stable equilibrium by
the balancing of interatomic attractive and repulsive forces ; that the atoms
are incompressible but deformable ; that the atoms are closely packed by
the squeezing together of the spheres so that the interstices are filled, while the
volumes of the polyhedral forms which the atoms assume remain constant ;
and that the homogeneous structure of a crystal is obtained by a symmetrical
arrangement of the atomic spheres of influence. It- is further shown that any
symmetrical and homogeneous aggregate of closely packed atomic spheres, can be
divided into space units which represent in composition and configuration the
chemical molecule ; and that the dimensions of each of these units will be in accord
with the crystal form of the compound. As Gr. D. Liveing (1891) remarked,
" The problem is then reduced to finding how to pack the greatest number
of equal spherical balls into a given space." When two or more arrangements,
equally closely packed, are possible, each may occur and polymorphism will
result.
W. Barlow and W. J. Pope assume that the relative volumes of the atoms of a
compound — the atomic domains — are proportional to their valencies, so that the
valency of an atom expresses the relative atomic volume ; or the valency of an atom
is proportional to the space occupied hy that atom in a crystalline compound — valency
volume. For example, in benzene, the volume of one carbon atom is four times
that of four hydrogen atoms because the valency of carbon is four, and of hydrogen
one. The atomic volume of the same element or radicle may be different in
different bodies. For example, A. E. H. Tutton has shown that the potassiimi and
caesium sulphates
Molecular volume.
Axial ratios, a : b : €=•
K2S04 .
94-91
0-5727 : 1 : 07418
CS2S04 .
84-58
0-5712: 1 : 0-7531
SO that the substitution of an atom of caesium for potassium lowers the molecular
volume while the configuration of the molecules as shown by the axial ratios a:h: c
remains virtually unchanged. The caesium atom enters the molecule in a definite
position, and therefore it does not seem possible for the configuration of the
molecule to remain constant unless the sulphuric acid group increases in the
same ratio. This shows that the absolute atomic volume of an element may
vary in different chemical bodies, though the valency volume, the ratio of the
volume of any constituent to the volume of the whole molecule, may remain
constant.
W. Barlow and W. J. Pope build up structures representing various chemical
molecules by closely packing deformable spheres in a homogeneous symmetrical
VOL. I. ^
242
INORGANIC AND THEORETICAL CHEMISTRY
arrangement, and they deduce the dimensions of the crystalline form of each sub-
stance. For example, they show that with equal spheres the dimensions of the
aggregate will have holohedral cubic or hexagonal symmetry, and the axial ratio
a : c with the latter must be either a:c=l: 08165 or a : c=l : 1*4142. Hence, in sub-
stances built up with atomic domains of equal size — elements, and binary compounds
of elements with, the same valency — the crystalline forms must be either cubic or
hexagonal. Of the forty elements whose crystals have been examined, half are
cubic, and one-third are hexagonal with axial ratios in accord with the theoretical
requirements. The remaining six are either pseudo-cubic or pseudo-hexagonal, and
their exceptional behaviour is explained by assuming that the atomic spheres are
packed in groups or aggregates so that some of the spheres are differently situated
in a close-packed homogeneous assemblage. Again, with the binary compounds of
the elements, J. W. Retgers found 88 per cent, of the known forms are either cubic
or hexagonal, and nearly all of these are composed of elements of equal valency.
For example :
Beryllium oxide, BeO
Zinc oxide, ZnO
a: c = l : 1-6305
1 : 1-6077
Zinc sulphide, ZnS .
Cadmium sulphide, CdS
1 : 1-8176
1 : 1-8109
In some cases in place of holohedral symmetry, these binary compounds exhibit
hemihedrism or tetartohedrism, a state of things which occurs with symmetrical
arrangements of two kinds of spheres of slightly different size.
The geometrical laws which govern the replacement of spheres in a closely
packed assemblage by others of (Efferent size run parallel with the chemical laws
which determine the relations between the valencies of the elements which can
replace one another in a chemical compoimd. It is said that all cases of substitution
can be reduced to one of two types : (i) those in which the sum of the valencies of
the substituting and substituted atoms or groups is constant — e.g. the substitution
of hydrogen by chlorine ; and (ii) those in which the sum of their valencies differ
— e.g. the substitution of hydrogen by methyl.
In substitutions of the first type, virtually no change occurs in the configuration
of the molecule. This is taken to be exemplified by B. Gossner's study of the
rhombic chloro- and bromo-ethanes — Table XI. :
Table XI.
CI3C.CCI8
ClsC.CBraCl .
ClaBrC.CBrCla
BrjCCBra .
HBrjCCBr, .
Molecalar
Valency |
volume.
volume. 1
1
113-34
!
14
.
116-72
14
,
12016
14
131-83
14 1
•
126-46
14
Axial ratios
a:b:e.
0-5677 : 1
0-5612 : 1
0-5646 : 1
0-5639 : 1
0-6650 : 1
0-3160
0-3171
0-3192
0-3142
0-3118
Equivalence parameters
x:y:z.
2-4260
2-4047
2-4090
2-4197
2-4294
4-2733 : 13503
4-2849 : 1-3587
4-2669 : 1-3620
4-2911 : 1-3483
4-2995 : 1*3406
The equivalence parameters x, y, z here represent the three dimensions of the
molecular volume reduced to a volume proportional to the valency volume. The
constancy of the values x, y, z is taken to mean that while the absolute volumes of
the molecules vary, the configuration of the molecules undergoes no change. This
means that the ratio of the volumes of substituting and substituted atoms to
the volume of the whole molecule is constant. If it were not so, the configuration
of the molecule would change. Hence, the relative volimies of hydrogen, bromine,
and chlorine are proportional to their valencies.
With substitutions of the second type, the dimensions of the molecule do change,
and the change is proportional to that of the valency of the atoms or radicles
concerned. For example, with the minerals, chondrodite, humite, clinohumite,
COMBINATION BY VOLUME
243
and forsterite, studied by S. L. Penfield and W. T. H. Howe— Table XII, T. V.
Barker emphasized the fact that prolectite has not been analyzed, and that
W. Barlow and W. J. Pope credit the mineral with the composition indicated in
the table solely because its axial ratio has the value there indicated.
Table XIT.
Forsterite, Mg2Si04 .
Prolectite, MgSi04.2Mg(F, OH) .
Chondrodite, Mg3(Si04)2.2Mg(F, OH)
Humite, Mg5(SiO4)3.2Mg(F,l0H) .
Clinohumite, Mg7(Si04)4.2Mg(F,OH)
Valency
volume
16
22
38
54
70
a:b:c.
0-9296 : 1 : 2-4492
1-0803 : 1 : 2-3877
1-0863: 1 : 3-1447
1-0802 : 1 : 4*4033
1-0803 : 1 : 5-6588
x'.y.z.
2-4492
2-3877
2-4294
2-4279
2-4347
2-2769: 2-8691
2-2102: 4*1689
2-2333 : 7*0199
2-2475: 98965
2-2769 : 12-8691
The minerals differ by the constant increment of the group Mg2Si04, which corre-
sponds with forsterite, which has the a : b ratio common to the whole series, while
the c : h ratio closely expresses the successive increments of the c-axis. This
indicates that possibly the structural units of chondrodite, humite, and clinohumite
are nothing more or less than the structural unit of prolectite on which have been
superposed one, two, and three structural units of forsterite. The Mg2Si04 group
enters the molecule in the z direction, since the values of x and y remain nearly
constant. Hence, say W. Barlow and W. J. Pope, the volume of the molecule ia
proportional to z ; but the ratio of the valency-volume w:z is nearly constant, so
that the volume is proportional to the valency ; consequently, it follows that the
axial ratios a:h : c, and the equivalence parameters x : y : z, can be predicted
for some member of a series when these values for other members are known.
T. V. Barker has shown that the close correspondence between fact and theory
is due to the closeness with which the hypothetical valency volumes satisfy the
conditions : volume of Mg2Si04 : volume of Mg(F,OH)2=2 : 1, and that the
particular volumes arbitrarily selected by W. Barlow and W. J. Pope present one
out of an infinite number of such solutions. Furthermore, when the minerals are
arranged in accord with chemical composition, the order is not that of the accepted
axial ratios, but forsterite, clinohumite, humite, chondrodite, and prolectite. There
is very little difference in composition between clinohumite and forsterite, even the
end-member prolectite differs from forsterite only by the amount of Mg(F,0H)2.
The group, indeed, is to be regarded as a series of varieties of forsterite with variable
amounts of Mg(F,0H)2, until further investigation has established the individuality
of the mineral species.
It has been noticed that water of crystallization causes the expansion of the
molecule mainly in one direction, showing that the water of crystallization probably
enters the crystal structure in layers. With double salts there is usually a simple
numerical relation between the valency values of the components which results
geometrically from the packing. Thus, with potash alum, K2S04.Al2(S04)3.24H20,
the valencies of the component parts are in the ratio 14 : 36 : 96, which is nearly
1:3:8.
In a close-packed assemblage of spheres, certain groups of spheres may be
situated so that they can be moved without affecting the arrangement of the others,
and other spheres can be fitted into the resulting cavities so as to reproduce the
original arrangement. Equivalent atoms or groups of atoms can replace others
of the same volume without change of structure since, by hypothesis, valency is
proportional to the relative volumes. Stable systems of related substitution
products lead to likeness in crystal form — isomorphism. Equivalent atoms, though
nearly equal, may differ slightly in volume, for A. E. H. Tutton, in his work on
244
INORGANIC AND THEORETICAL CHEMISTRY
the alkali sulphates and selenates, obtained the following results, indicated in
Table XHI :
Table XIII
Molecular
volume.
Valency
volume.
x:y:z.
Rb2s64 '.'.'.'.',
CS2SO4
KaSeO*. . . . .
RbaSe04
Cs,Se04
64-91
73-34
84-58
71-67
79-94
91-09
12
12
12
12
12
12
2-2109 : 2-1977 : 2-8463
2-2049 : 2-1899 : 2-8648
2-2003 : 2-1826 : 2-8777
2-2207 : 2-2083 : 28204
2-2147 : 2-1957 : 2-8412
2-2112:2-1900:2-8524
The difference in the values of x, y, z indicate slight differences in the configuration
of the molecules, corresponding with slight differences in the volumes of the atoms
of potassium, rubidium, and caesium. The atomic volumes of elements in the
same group of the periodic classification probably increase slightly with increasing
atomic weight. This principle gives an explanation of ; (1) The difference in the
crystal forms of the acids and many of their salts — e.g. H2SO4 and K2SO4 — by the
difference in the atomic volumes of hydrogen and potassium. (2) The isomorphism
of salts of different acids — ^plagioclase felspars, NaAlSiaOg and CaAl2Si208. Here
NaSi- and CaAl-groups have the same molecular volume and hence can replace
one another. (3) The isomorphism of complex bodies not observed in simple
bodies. (4) The variation in the form of the crystal when replacing atoms with the
same valency-volume but different in form — e.g. the substitution of a N02-group
by a methyl or CHs-group.
If like spheres are removed homogeneously from an assemblage of molecules
and replaced by larger spheres or groups of spheres the walls of the cavity will gape.
Close packing can be restored by inserting new spheres in the gap so produced, and
the volume of the gap will be equal to the difference in volume of replaced and
replacing spheres. If v be the volume of spheres replaced, the volume of replacing
spheres to produce close packing will be v ; t^+l+l or v+2 ; v+2+2 or v-\-i ;
and generally v-{-2n ; that is, the volume of replacing spheres differ by intervals
of 2. Chemically this means that the valencies of multivalent elements differ
constantly by intervals of 2. The change in valency by intervals of 2 involves no
variation in the actual volume of the atom. Thus, caesium can combine with iodine
to form Csl, Cslg, Cslg, Csly, and Cslg.
In general, if the volume m be replaced by a volume n-\-m, an additional
volume n must be added to conserve the marshalling — thus, if a univalent hydrogen
atom be replaced by a quadrivalent carbon atom, an additional three valencies
are required. Again, in replacing an H-atom by an oxygen atom in ethane, the
further addition of an atom of unit valency is required to produce C2H5OH. Again,
in three molecules of benzene, three hydrogen atoms may be substituted by a nitrogen
atom to form triphenylamine, (C6H5)3N, but if a carbon atom takes the place of
three hydrogen atoms an additional hydrogen atom is required to form triphenyl-
methane, (C6H5)3CH. The apparent isomorphism of salts like sodium nitrate,
NaNOa, and calcium carbonate, CaCOs, is likewise explained by the volume-valency
hypothesis.
The hypothesis that each valency in a given compound has the same volume
is not generally accepted. T. W. Richards, for instance, argues that the relationship
between benzene and tetrabromobenzene lands its supporters in an impossible
position since it requires the subsidiary assumption that all the remaining carbon
and hydrogen atoms in benzene should nearly double their volume when four atoms
of bromine are substituted for hydrogen. It is more in accord with H. Kopp's
COxMBINATION BY VOLUME 245
observations that the atomic volume of hydrogen in combination is much larger than
that of free hydrogen. In the nitro-dihalogeno-benzenes, observed by E. Repossi,
the introduction of chlorine causes a marked shortening of the fo-aids ; and in the
rhombic chloro- and bromo-ethanes, observed by B. Gossner, the symmetrical
substitution of chlorine lengthens the a-axis in relation to the 6-axis, and the un-
symmetrical substitution of the same element shorten the a-axis. In these and
many other examples there are real differences in the molecular volume and
crystalline form ; but the greatest variation occurs with the molecular volume,
probably because a change in volume produces only the cube root of its proportional
effect when applied in any one axial direction. The atomic volumes of chlorine
and bromine respectively at — 33'5° and 63° — which are like fractions of their boiling
points — are 21-8 and 25*5 respectively. Although the atomic volumes change
when these elements enter into combination, under similar conditions, the chlorine
would be expected to occupy less space than bromine, and this is what is actually
observed. It is urged that the method of equivalence parameters employed for
testing the hypothesis of valency- volume is a mathematical device which has the
effect of reducing widely deviating data to apparent harmony ; and of hiding
differences in crystal forms. The observed data are the axial ratios a:b :c.
According to W. Barlow and W. J. Pope, the molecular volumes of liquids of normal
paraffins calculated on the basis that one volume of carbon is equal to four volumes
of hydrogen (H=2*97 c.c. ; C=1I'88 c.c.) are in close agreement ; but T. W.
Richards showed that the agreement is quite as good when calculated on other
hypotheses, say, when the atomic volume of carbon is twice that of hydrogen.
T. V. Barker has shown that on the volume- valency hypothesis the atomic units
in potassium chloride are approximately the same, as they are likewise in potassium
iodide : consequently, since the atomic volumes of these two salts are respectively
37*49 and 53'06, the atomic volumes of the potassium atoms in these salts are
respectively 18* 74 and 26'53. There is every reason to believe that the atomic
volume of an element in the free state may be much greater than when it is combined,
so that there is no inherent objection to the hypothesis that the atomic volume of
an element varies a little in its different forms of combination. Again, T. V. Barker
has shown that the molecular volumes of potassium and ammonium iodides are
respectively 53*06 and 59'62. On the volume -valency hypothesis, " the iodine
atom in potassium iodide has a volume equal to 26*53, but in the ammonium com-
pound the iodine can only have one-eighth the volume of the molecule, since the
total valency-volume is eight — that is, volume of the iodine equals 7*45. It must
be assumed that when potassium is substituted by ammonium the iodine atom
experiences a shrinkage equal to about five-sevenths of its volume in the potassium
compound. Is it likely that this is really the case ? It must be remembered that
a comparison is not being made between the atomic volumes of an element in the
free and combined conditions, but rather its atomic volume in two compounds
which have an extraordinarily close chemical relationship." A. E. H. Tutton adds
that to explain the incontrovertibly proved iso-structure of ammonium and rubidium
sulphates, where there are 24 valency- volumes in the former cases and only 12 in
the latter, it is necessary to assume arbitrarily that the actual spheres of atomic
influence in the former are on a smaller scale — one-half indeed — than in the latter
in order to afford in the total the same volume. The congruency of the monoclinic
ammonium and rubidium nickel sulphates or selenates shows that the replacement
of two atoms of univalent rubidium by eight atoms of univalent hydrogen and two
atoms of ter- or quinque-valent nitrogen, produces no opening up of the structure.
This fact, says A. E. H. Tutton, is in entire antagonism with the hypothesis of valency-
volume of W. Barlow and W. J. Pope.
T. V. Barker concludes from his study of the valency- volume theory that so
far " the theory has received no general crystallographic support." A. Ogg and
F. L. Hopwood also concluded from the X-ray spectrometric examination of the
ammonium and alkali metal sulphates, that the replacement of eight potassium
246 INORGANIC AND THEORETICAL CHEMISTRY
atoms by forty atoms of the four NH4-radicles produces so little difierence from
what the replacement by eight rubidium atoms produces on the dimensions of the
elementary cell, as to furnish " conclusive evidence against the general truth of the
theory of crystal structure based on the closest packing of the constituent atoms or
their spheres of influence."
W. J. SoUas 1* assumed that, with the haloids of the alkali metals, the structural
unit is an aggregate of four molecules. He further assumes generally that the
structural units are proportional to the atomic volumes, not of the free elements,
but of the elements in the compounds concerned ; nor is the sum of the atomic
volumes equal to the molecular volume, for an allowance is made for interstitial
space. W. J. Sollas also assumed that the structural units are loosely packed;
he builds up structures with his units which give results in harmony with the
observed geometrical forms of the crystals and with some of their physical properties.
He has applied the hypothesis to the alkali halides, silver iodide, titanium dioxide,
and cassiterite. T. V. Barker adds that loose packing is equivalent to assuming
that some of the atoms have a greater volume than is initially supposed, and with
sufficiently loose packing, it does not seem difficult to obtain any structure whatever
irrespective of the atomic volumes.
References.
1 A. le Royer and J. B. A. Dumas, Ann. Pharm. Chim., 92. 408, 1821.
2 H. Schroder, Pogg. Ann., 56. 553, 1840 ; Wied. Ann., 4. 435, 1878 ; 11. 997, 1880 ; 14.
657, 1881 ; Ber., 10. 848, 1871, 1877 ; A. Horstmann, ib., 19. 1582, 1886 ; F. Krafft, ib., 15.
1711, 1882 ; 17. 1371, 1884 ; A. Bartoli, Ath Accad. Lincei, (3), 19. 577, 1884 ; G. Tschermak,
Liebig's Ann., 112. 129, 1859 ; 114. 25, 1859 ; W. Lessen, ib., 254. 42, 1889 ; C. M. Guldberg,
Zeit phys. Chem., 5. 374, 1890 ; T. E. Thorpe, Journ. Chem. Soc., 63. 775, 1893 ; G. le Bas, ib.,
91. 122, 1907 ; I. Traube, Zeit. anorg. Chem., 8. 12, 1895.
3 W. Herapath, Phil. Mag., (1), 64. 321, 1824 ; C. J. B. Karsten, Schweigger's Journ., 65. 394,
1832 ; F. Ammermiiller, Pogg. Ann., 49. 341, 1840 ; 50. 406, 1840 ; J. F. Persoz, Intro-
duction a Vetude de la chimie moliculaire, Paris, 1839; H. Kopp, Pogg. Ann., 47. 113, 1839;
52. 243, 262, 1844 ; Liebig's Ann., 41. 79, 1842 ; 96. ], 185, 303, 1855 ; 250. 1, 1889 ; Liebig's
Ann. iSuppl., 5. 303, 1867 ; The Kopp Memorial Lecture, T. E. Thorpe, Journ. Chem. Soc, 63.
775, 1893 ; P. F. G. Boullay, Ann. Chim. Phys., (2), 43. 266, 1830 ; L. Playfair and J. P. Joule,
Me7n. Chem. Soc, 2. 401, 1845.
* R. Gartenmeister, Liebig's Ann., 233. 249, 1886 ; J. Pinette, ib., 243. 32, 1888 ; P. Dobriner.
ib., 243. 11, 1888 ; E. Elsasser, ib., 218. 337, 1883 ; A. Zander, ib., 214. 138, 1882 ; 225. 74, 1884 ;
W. Lossen, ib., 254. 42, 1889 ; 214. 81. 1882 ; T. Liebisch, tb., 286. 140, 1895 ; T. Traube, ib.,
240. 43, 1887 ; R. SchifiE, ib., 220. 71, 278, 1883 ; Ber., 14. 2761, 1881 ; F. Krafft, ib., 15. 1711,
1882 ; 17. 1374, 1884 ; W. Stadel, ib., 15. 2559, 1889 ; H. Schroder, ib., 13. 1560, 1880 ;
P. Walden, ib., 29. 1699, 1896; R. Willstatter, ib., 41. 1480, 1908 ; J. C. Brown, Pror. Roy. Soc.,
26. 247, 1878 ; S. Feitler, Zeit. phys. Chem., 4. 66, 1889 ; F. Neubeck, ib., 1. 649, 1887 ;
W. Ramsay, J (mm. Chem. Soc., 35. 463, 1879 ; T. E. Thorpe, ib., 37. 141, 327, 1880 ; H. L. Buff,
Liebig's Ann. Suppl., 4. 129, 1865 ; A. Horstmann, Theoretische Chemie, Braunschweig, 1885 ;
G. le Bas, Journ. Chem. Soc., 91. 112, 1907 ; The Molecular Volumes of Liquid Chemical Compounds,
London, 1915 ; Phil. Mag., (6), 14. 324, 1907 ; (6), 16. 1908 ; (6), 27. 344, 741, 976, 1914 ; (6),
28. 439, 1914 ; Chem. News, 98. 85, 1908 ; 99. 206, 1909.
5 L. Playfair and J. P. Joule, Mem. Chem. Soc., 2. 401, 1845 ; R. Schiflf, Liebig's Ann.,
107. 64, 1858; T. E. Thorpe and J. I. Watts, Journ. Chem. Soc., 37. 102, 1880; F. W. Clarke,
Amer. Journ. Science, (3), 8. 428, 1874.
' L. Meyer, Liebig's Ann. Suppl., 7. 354, 1870 ; T. Thomson, Chemistry of Inorganic Bodies,
London, 1. 14, 1831 ; W. Borchers, Die Beziehungen zwischen aquivahnt Volumen und AUnrige-
MJic^f, Halle, 1904.
' F. Exner, Sitzber. Akad. Wien., 91. 850, 1885 ; Monatsh., 6. 249, 1885 ; P. A. Guye, Ann.
Chim. Phys., (6), 21. 208, 1890 ; I. Traube, Ann. Physik, (4), 5. 552, 1901.
8 T. W. Richards, Proc. Amer. Acad., 37. 1, 399, 1901 ; 38. 293, 1902 ; 39. 581, 1904 ; Journ.
Amer. Chem. Soc, 26. 399, 1904; 29. 117, 1907; 30. 8, 1908; 31. 158, 1909; 34. 971, 1912;
35. 381 1913; 36. 617, 2417, 1914; Journ. Chem. Soc, 99. 1201, 1911; T. W. Richards
and W. N. Stull, New Method of Determining Compressibility, Washington, 1903 ; T. W. Richards,
W. N. Stull, F. N. Brink, and F. Bonnet, The Compressibilities of the Elements and their Periodic
Relations, Washington, 1907 ; E. Griineisen, Ann. Physik, (4), 33. 1239, 1910 ; W. Miiller-Erzbach,
Ber., 14. 217, 2043, 1881 ; F. Ephraim and E. Michel, Helv. Chim. AcUi, 2. 266, 1919 ; F. Ephraim
and P. Wagner, i6.,50. 1088, 1917 ; H. H. Stephenson, C/ie^n. News,±{^. 178, 187, 1911; P.Kirwan,
Versuche und Beobachtungen iiber die specif sche Schwere und der Anziehungskraft verschiedener
Salzarten. Berlin, 1785 ; H. Davy, Collected Works, London, 5. 133, 1840 ; J. D. van der Waals
COMBINATION BY VOLUME 247
Die Continuitdt des gasfdrrnigen und fliissigen Zustdndes, Leipzig, 46-59, ] 886 ; Zeit. phys. Cfiem.,
38. 257, 1901 ; W. WitkofEsky, Krak. Anz., 305, 1905 ; H. K. Oiines, Proc. Acad. Amsterdam,
448, 1908 ; Crnnm. Univ. Leiden, 102. 1, 1907 ; G. S. Johnson, Journ. Chem. Soc., 31. 252, 1877 ;
I. Traube, Ueber den Raum der Atome, Stuttgart, 1899; W. Sutherland, Phil. Mag., (5), 27. 305,
1889 ; (6), 4. 632, 1902 ; A. P. Mathews, Journ. Phys. Chem., 17. 154, 181, 252, 320, 331, 337,
481, 520, 603, 1913 ; 18. 474, 1914.
9 H. Schroder, Pogg. Ann., 56. 553, 1840 ; Wied. Ann., 4. 435, 1878 ; 11. 997, 1880 ; 14.
656, 1881; Ber., 10. 1848, 1871, 1877 ; 13. 1560, 1886; Die, Molecular volume der chemischen
Verbindungen im festen und fliissigen Zustdnde, Mannheim, 1843 ; W. Ostwald, Lehrbuch der
allgemeinen Chemie, Leipzig, 1. 387, 1891.
" G. le Bas, Journ. Chem. Soc, 91. 112, 1907; PhU. Mag., (6), 14. 324, 1907; (6), 16. 60,
1908 ; (6), 27. 344, 741, 976, 1914 ; (6), 28. 439, 1914 ; Chem. News, 98. 85, 1908 ; 99. 206, 1909 ;
The Molecular Volumes of Liquid Chemical Compounds, London, 1916 ; Science Progress, 8. 663,
1914.
11 I. Traube, Ueber den Raum der At&me, Stuttgart, 1899 ; Ber., 25. 2524, 1892 ; 27. 3173,
1894 ; 28. 410, 2722, 2728, 3292, 1895 ; 29. 1023, 1896 ; 30. 265, 1897 ; 40. 130, 723, 1907 ;
Zeit. anorg. Chem., 3. 1, 1892 ; 8. 12, 1895; Liebig's Ann., 290. 44, 1895; Wied. Ann., 61.
380, 1897; 62. 490, 1897; Ann. Physik, (4), 5.552, 1901; (4), 22. 519, 1907; Zeit. -phys.
O^iem., 28. 115, 1899.
12 W. C. Roberts -Austen, Second Report to the Alloys Research Committee, London, 1893 ;
Third Report, London, 1895; Metallurgy, London, 1902; H. Tomlinson, Phil. Trans., 174. 1,
1883; Proc. Roy. Soc, 38. 42, 488, 1885; A. Wertheim, Ann. Chim. Phys., (3), 12. 385, 1844;
(3), 23. 52, 1849 ; W. Sutherland, Phil. Mag., (5), 32. 31, 215, 524, 1891 ; R. A Fessenden.
Journ. Franklin Inst, 142. 187, 1896 ; H. Crompton, Journ. Chem. Soc, 50. 315, 1896 ; Ber., 28.
148, 1895 ; 1. Traube, ib., 27. 2178, 1894 ; F. Osmond, Compt. Rend., 110. 346, 1890 ; Ann. Chim.
Phys., (5), 20. 66, 1880 ; E. Warburg and F. Tegetmeier, Wied. Ann., 41. 18, 1890 ; E. Warburg,
t6.,21. 622, 1884; 40. 1, 1890.
i» W. Barlow and W. J. Pope, Journ. Chem. Soc, 89. 1675, 1906 ; 91. 1150, 1907 ; 93. 1528,
1908 ; W. Barlow, Min. Mag., 17. 314, 1916 ; T. W. Richards, Journ. Amer. Chem. Soc, 35.
381, 1913 ; 36. 1686, 1914 ; W. Barlow and W. J. Pope, ib., 36. 1675, 1694, 1914 ; A. E. H. Tutton,
Proc Roy. Soc, 93. 68, 72, 1917 ; Phil. Trans., 217. 199, 1917 ; E. Repossi, Zeit. Kryst., 46.
202, 1909 ; B. Gossner, ib., 38. 154, 1904 ; 40. 84, 1905 ; T. V. Barker, Journ. Chem. Soc, 101.
2496, 1912 ; 107. 744, 1915 ; S. L. Penfield and W. T. H. Howe, Amer. Journ. Science, (3), 47
188, 1894; A. Ogg and F. L. Hopwood, Phil. Mag., (6), 32. 518, 1919.
1* W. J. Sollas, Proc Roy. Soc, 63. 270, 286, 296, 1898; 67. 493, 1900; 69. 294, 1902; 80.
267, 1908; T. V. Barker, Journ. Chem. Soc, 101. 2493, 1912
CHAPTER VI
THE CLASSIFICATION OF THE ELEMENTS
§ 1. The Classification of the Elements
Most of our systems of classification are artificial and without distinct lines of demarca-
tion. Being based upon limited knowledge, they have been formed upon apparent rather
than upon real similarities and differences ; and they are to our minds but artificial aids,
like crutches to cripples.— G. Gore (1878).
The classification of the elements has long been an attractive subject. It is only
by the aid of classification that the mind of man is able to cope with the multitudinous
facts presented by nature. In the words of F. Bowen (1866) ^ :
The first necessity which is imposed upon us by the constitution of the mind itself, is
to break up the infinite wealth of Nature into groups and classes of things with reference
to their resemblances and affinities, and thus to enlarge the grasp of our mental faculties,
even at the expense of sacrificing the minuteness of information which can be acquired
only by studying objects in detail.
The primary object of classification is to arrange the facts so that we can acquire
the greatest possible command over them with the least possible effort. This is
accomplished by arranging the facts in a systematic way. In all systems of classifi-
cation, the elements are assembled in a few groups or classes so that the members
of each group possess in common the greatest possible number of important
attributes ; and the attempt is made to collect together in one group the elements
which are alike in general properties, and to separate those which are unlike. No
one has succeeded in devising an unimpeachable system of classification for the
chemical elements, in which each element has only one peculiar place — when the
criterium of the classification is chemical behaviour. It is invariably found that
some elements are entitled with equal or almost equal consistency to a place in more
than one group. In the ideal system of classification, each class will be clearly
and sharply distinguished from every other class by some essential property or
properties which can be accurately defined, and readily recognized, and which are
common to the individual members of the class. In all the systems hitherto
proposed the different classes are more or less affiliated one to another, and J. P.
Cooke 2 has emphasized the fact that
Nature seems to abhor abrupt transitions, and shades off her bounding lines. Many
of the elements, while they manifestly belong to one series, have properties which ally
them to another.
The alchemists divided the metals into two classes, the perfect metals and the
semi-metals — the former included gold and silver ; the latter copper, iron, lead, tin,
and mercury. The former suffered no alteration when heated at the highest
available temperatures ; the base metals are changed under these conditions into
primitive earths. The former were accordingly called noble metals, the latter base
metals. At the beginning of the nineteenth century, J. J. Berzelius classified the
elements into two groups, the m£ials and the metalloids (meaning non-metals).
This simple division of the elements into two groups is confronted with many
difficulties because some simple substances — like antimony and arsenic — have the
general appearance of metals and yet behave chemically like the non-metals. These
248
THE CLASSIFICATION OF THE ELEMENTS
249
pseudo-metals exhibit characteristic properties of both classes. This simple
dichotonous division leads to vagueness, ambiguity, and contradictions as soon as
the attempt is made to formulate sharp clear-cut definitions of the metals and the
non-metals. The attempt to group the elements by a code of definitions seems to
be foredoomed to failure. There is a seductive simplicity about a definition which
may be attractive, but it is artificial and often misleading. As T. Campanella
(1590) expressed it : " Definition is the end and epilogue of science. It is not the
beginning of our knowing, but only of our teaching."
The attempt has been made to mark the metals by a term ending in um, and
the non-metals by a term ending in en, ine, or on. For example : Metals —
Aluminium, barium, sodium, magnesium, calcium, ferrum (iron), hydrargyrum
(mercury), etc. Non-metals — Boron, carbon, oxygen, silicon, chlorine, argon,
neon, krypton, etc. The idea persists in many but not all the modern names of the
elements. The time-honoured names, silver, gold, iron, copper, zinc, etc., have
alternative Latinized equivalents — argentum, aurum, ferrum, etc. — from which
their modern symbols are derived.
Philologists 3 tell us that the word metallon appeared in Greek literature about
the time of Herodotus (c. 450 B.C.), and it is supposed to have been borrowed from
some foreign language — possibly a Semitic word — since the Semites, represented by
the Phoenicians, had mines in the island of Thasos — not of the ^gean. The Semitic
meaning of the term was " to work iron like a smith," whereas the Greeks used it
not for a metal but for a mine or for any kind of mineral, including salt, found in
a mine. The resemblance of the Greek /AeVaXXov — a mine — to fieraXXdo) — meaning
** in quest of something "• — is thought to be accidental.
With aU its imperfections, J. J. Berzelius' subdivision of the elements into metals
and non-metals is so convenient that it is in common use when great precision is
not required. Very roughly, the properties of the metals can be contrasted with
those of the non-metals as indicated in the subjoined scheme — Table I.
Table I. — ^The Properties of the Metals and Non-metals contrasted.
Metals.
1. Form basic oxides.
2. Generally dissolve in mineral acids giving
off hydrogen.
3. Either form no compoimds with hydro-
gen, or form unstable compounds- —
usually non-volatile.
4. Solid at ordinary temperature (excepting
mercury).
5. Usually volatilize only at high tempe-
ratm-es.
6. When in bulk the metals reflect light
from polished or freshly cut surfaces.
7. Specific gravity is generally high.
8. Good conductors of heat and electricity.
Electrical resistance usually increases
with the rise of temperature.
9. More or less malleable and ductile.
10. Molecules usually monatomic in the
vaporous state.
Non-metals.
1. Form acidic oxides.
2. Do not usually dissolve easily in mineral
acids.
3. Form fltable compounds with hydrogen
— these are usually volatile.
4. Gases, liquids, or solids at ordinary
temperatures.
5. Excepting carbon, boron, and silicon, the
non-metals are either gaseous or volati-
lize at low temperatures.
6. Do not usually reflect light very well.
7. Specific gravity generally low.
8. Bad conductors of heat and electricity.
Electrical resistance usually decreases
with rise of temperature.
9. Malleability €Uid ductility are not well
defined.
10. Molecules usually polyatomic in the
vaporous or gaseous state.
If any particular property be selected as a criterion, it will be found that the
different metals can be arranged in a series which passes imperceptibly into the
non-metals without an abrupt change. P. P. von Weimarn,* indeed, has given
evidence which makes it not improbable that the different metals can exist in both
260 INOKGANIC AND THEORETICAL CHEMISTRY
forms. Hence, the metallic or non-metallic nature is not an unchangeable character-
istic of an element ; the metallic property may predominate at a low temperature,
the non-metaUic at a high one. This does not mean that metals and non-metals
are two distinct forms of matter like the three states of aggregation which all kinds
of matter undergo when the conditions — ^particularly temperature — are favourable.
To show how difficult it is to draw a hard-and-fast line of demarcation between
metals and non-metals, the non-metals arsenic, antimony, and tellurium would be
classed with the metals if we depended exclusively upon 6, 7, and 8 ; hence, some
introduce a third division — the metalloids — to include the hybrids — almost
analogous to the bastard metals of the alchemists — for elements which have properties
characteristic of both the metals and the non-metals. The term metalloid is some-
times used synonymously with non-metals. The metals lithium, sodium, potassium,
magnesium, and alimiinium have a low specific gravity ; and when potassium was
discovered in 1807, some argued that it could not be a metal because it was light
enough to float on water. The non-metals carbon, boron, and silicon are less
volatile than most metals. The non-metal hydrogen is a good conductor of heat ;
and the non-metal graphitic carbon is a good conductor of heat and electricity.
Hence the division of the elements into metals and non-metals is but a rough system
of classification, arbitrarily adopted because it is convenient.
In A. L. Lavoisier's classification, in his Traite elementaire de chimie (Paris, 1789),
the elements were arranged in four classes :
(1) Simple substances belonging to the three kingdoms which can be regarded as the
elements of bodies — light, caloric, oxygen, nitrogen, hydrogen. (2) Simple non-metallic
substances which are oxidizable and acidifiable — sulphur, phosphorus, carbon, muriatic
radicle, fluoric radicle, boracic radicle. (3) Simple metallic substances, oxidizable and
acidifiable — antimony, arsenic, bismuth, cobalt, copper, gold, iron, lead, manganese,
mercury, molybdenum, nickel, platinum, silver, tin, tungsten, zinc. (4) Simple substances,
salifiable and earthy — ^lirae, magnesia, baryta, alumina, silica.
Lavoisier frequently expressed his desire to keep within the limits of experience, but
that fact was not sufficient to prevent him from nursing the philosophical conception
of a world made from a very small number of elements. He seems to have regarded
light and heat as ponderable or material substances. Lavoisier's separation of
simple substances into metals and non-metals is interesting.
A. F. de Fourcroy, in his Systeme des connaissances chimiques (Paris, 1801),
divided the then known metals into five classes : The first class included brittle
metals which form acids by combining with oxygen — arsenic, tungsten, molybdenum,
and chromium ; the second class included brittle metals which do not form acidifiable
oxides — titanium, uranium, cobalt, nickel, manganese, bismuth, antimony, and
tellurium. The third class included metals which showed signs of ductility and are
oxidizable — zinc and mercury. The fourth class included ductile and easily
oxidizable — tin, lead, iron, and copper. The fifth class included the ductile metals
not oxidizable by fire — silver, gold, and platinum. The first three classes were
called demi-metaux ; the fourth class, me'taux imparfaits ; and the fifth class,
metaux parfaits.
The elements have been classed into acidic and basic, or, what amounts to the
same thing, into electro-negative and electro-positive elements — for example, J. J.
Berzelius,in his Ldrhok i Kemien (Upsala, 1818), where hydrogen with dual character-
istics separated the electropositive and the electronegative groups from one another.
The elements have also been classed according to their valency as A. Naquet did
in his Principes de chimie fondes sur les theories modernes (Paris, 1864). The
alchemists separated the metals into two groups — the base and the noble metals ;
and L. J. Thenard, in his Traite de chimie elemeritaire, theorique et pratique (Paris,
1818), imbued with A. L. Lavoisier's views on oxidation, relegated the non-
metals into one class, and classified the metals according to (i) their oxidizability
when roasted in air ; (ii) the stability of the oxides ; and (iii) their behaviour
when heated with steam. This system was modified by succeeding chemists
THE CLASSIFICATION OF THE ELEMENTS 251
and used by H. V. Regnault in his Cours elemcntaire de chimie (Paris,
1853). It was soon afterwards abandoned because it failed to accommodate itself
to new data and new knowledge.
J. B. A. Dumas, in his Traite de chimie appliquee aux arts (Paris, 1828), and later,
in a Memoire sur les equivalents des corps simples (1859),^ arranged the non-metals
in five groups or families. He founded his classification of the non-metallic elements
on the character of the compounds which they formed with hydrogen, on the volume
ratio of the two elements entering into combination, and on their mode of con-
densation.
I.— Hydrog^ne ; 11.^ — Fluor, chlore, brome, iode ; III.- — Oxygene, selenium, soufre,
tellure ; IV.- — Azote, phosphors, arsenic ; V.^ — Carbone, bore, silicium.
Hydrogen was supposed to be a metal. J. B. A. Dumas classed the metals, and in
general those bodies which do not unite with hydrogen, from the character of the
compounds they form with chlorine, and when possible, from the volume ratio of the
two elements entering into combination, and from their mode of condensation.
He further represented the atomic weights of the elements in a given
family as the sum of a series a-\-7d' -{-rid" -{- . . . Thus, fluorine =19 ; chlorihe
=19H-16-5=35-5 ; bromine=19+2 X 16-5+28=80 ; iodine=19+2x 16-5-1-2x28
+19=127. Similarly with the other families.
J. B. A. Dumas further stated that the classification of the metals ought to be
founded on the characteristic compounds which they form with chlorine ; and, as
a result of the further work of W. Odling,^ L. Gmelin, etc., systems were
devised in which family relationships were specially emphasized. Thus, W. Odling
(1857) arranged the elements in 13 groups :
I.— Fluorine, chlorine, bromine, iodine. II. — ^Oxygen, sulphur, selenium, tellurium.
III.— Nitrogen, phosphorus, arsenic, antimony, bismuth. IV.' — Boron, silicon, titanium,
tin. v.— Lithium, sodium, potassium. VT.— Calcium, strontiimi, barium. VII. — Majg-
nesium, zinc, cadmium. . VIII. — Beryllium, yttrimn, thorium. IX.- — Aluminium, zir-
conium, cerium, manium. X. — Chromium, manganese, iron, cobalt, nickel, copper.
XI.— Molybdenum, vanadium, tungsten, tantalum. XII. — Mercury, lead, silver. XIII.—
Palladium, platinum, gold.
In analytical chemistry, even to-day, the elements are conveniently classed
according to their behaviour towards certain reagents which are conventionally
taken as standards of reference. For instance :
Group I. — The addition of hydrochloric acid to a solution precipitates silver, mercuroua,
lead, and thalliimi chlorides ; tungstic oxide ; and possibly silicic acid and antimony
oxychloride.
' Group II. —The passage of hydrogen sulphide through the jfiltrate from the preceding
group, acidified with hydrochloric acid, precipitates mercury(ic), lead, bismuth, copper,
arsenic, antimony, tin, selenium, tellurium, gold, platinum, and molybdenimi either as
sulphides or as elements along with sulphur.
Group III. — The addition of ammonia, ammoni\mi chloride, and ammonivim sulphide to
the filtrate from the preceding group precipitates sulphides or hydroxides of iron, nickel,
cobalt, manganese, zinc, uranium, aluminium, chromium, titanium, beryllium, thallium,
zirconiiim, and the elements of the rare earths.
Group IV.— The addition of ammonium carbonate and alcohol to the filtrate from the
preceding group— after it has been evaporated to a small bulk— precipitates magnesium,
barium, strontium, and calcium carbonates.
Group v.— The salts of the alkah metals— lithium, sodium, potassium, rubidium, and
caesium — remain in solution.
The groups are subdivided again and again until finally salts of individual elements
remain.
In all these systems an element appears in more than one class ; or elements
with but few properties in common are grouped together ; or elements otherwise
related are separated into different classes. The properties of the elements used
as the basis of classification may also vary with the conditions under which the
properties are observed.
252 INORGANIC AND THEORETICAL CHEMISTRY
Were W. Odling's system to be revised in the light of existing knowledge, it
would furnish the most convenient system of classifying the elements — e.g. some
such system as that used in H. Moissan's Traite de chimie miner ale (Paris, 1904) ; and
also in L. Gmelin and K. Kraut's Handbuch der anorganischen Chemie (Heidelberg,
1907). It is usually supposed that the so-called periodic system (vide infra) is the
best that can be done, but that is doubtful. Only those facts which are known
can be classified ; and the perfection of any system of classification must necessarily
depend on the extent of our knowledge. The periodic system is usually adopted,
not because it furnishes the best possible arrangement according to existing know-
ledge, but because it holds alluring promises for the future. According to G. Gore :
The most perfectly philosophical classification of scientific truths can be made only
when their most essential and fundamental characters are discovered, and they are probably
those very characters which are the most difficult to find, and doubtless will be nearly the
last to be discovered.
References.
1 F. Bowen, A Treatise on Logic, Cambridge, U.S.A., 315, 1866 ; W. S. Jevons, The Principles
of Science, London, 2. 346, 1874 ; G. Gore, The Art of Scientific Discovery, London, 204, 1878.
2 J. P. Cooke, Amer. Journ. Science, (2), 17. 387, 1854.
3 P. Giles, Journ. Iron Steel Inst., 94. ii, 351, 1916 ; G. F. Zimmer, ib., 94. ii, 320, 1916.
* P. P. von Weimarn, Kolloid. Zeit., 13. 16, 1913; Journ. Russian Phys. Chem. Soc., 46.
1905, 1914 ; 47. 2177, 1915 ; 48. 1295, 1916; E. Jordes, Zeit. angew. Chem., 20. 2241, 1907.
5 J. R A. Dumas, Ann. Chim. Phys., (3), 55. 129, 1859.
« W. Odling, Phil, Mag., (3), 13. 423, 280, 1857 ; L. Gmelin, Handbuch der Chemie, Heidelberg,
1843.
§ 2. Triads, and the Law of Octaves
In all things there is order, harmony, and wisdom.' — ^H. Davy (1811).
Les propriit^s des corps sont les proprietes dea nombres. — A. E. B. de Chancourtois
(1862).
In all chemical changes one property at least remains unaltered, and the more
popidar systems of classification have been based on this property — the atomic
weights of the elements. The early efforts in this direction were seriously hampered
by the uncertainty in the numerical values of the atomic weights ; but after chemists
had cleared up the confusion associated with the atomic theory left by J. Dalton,
and obtained a consistent system of atomic weights, the results were more promising.
True enough, between 1816 and 1829, J. W. Dobereiner noticed some regularities
in the atomic weights of certain related elements, and, in a paper, Versuch zu einer
Gruppierung der elementdren Stoffe nach ihrer Analogie (1829), ^ he showed that most
of the chemically related elements either exhibited almost the same atomic weight —
e.g. iron, cobalt, and nickel — or else showed a constant difference when arranged
in sets of three. Thus, selecting one set from Dobereiner's list, and rounding off
the modern atomic weights,
Atomic Weight
Difference ....
Dobereiner's idea was taken up by a number of chemists — J. B. A. Dumas (1858),
J. B. P. Kremers (1852), J. H. Gladstone (1853), W. Odling (1857), and E. Lennsen
(1857) — and J. P. Cooke (1854) showed that the triads are probably only parts
of series, similar to the homologous series of organic chemistry, in which the differ-
ences between the molecular weights are a multiple of some whole number. He
struck a fatal blow at the doctrine by pointing out that the triads broke up natural
groups of the elements, so that instead of co-ordinating facts it tended in the opposite
direction. J. P. Cooke (1854) further arranged the elements in six groups or series
Jcium
Strontium. Barium.
40
87 137
47 50
THE CLASSIFICATION OF THE ELEMENTS 253
in which the members of each series followed a law of progression in which the
variation of the atomic weights could be expressed by a simple algebraic
formula.
In 1815, W. Prout 2 had suggested that the atomic weights of the elements were
probably exact multiples of the atomic weight of hydrogen, but with increasing
refinements in the methods of measurement, this suggestion did not fit with observa-
tions. In 1850, M. von Pettenkofer read a paper, Ueber die regelmassigen Ahstdnde
der Mquivalentzahlen der sogenannten einfachen Radicale (1850), before the Bavarian
Academy of Sciences, in which he tried to show that similar elements formed an
arithmetical series with the atomic weight of one of the elements as a whole number,
and another integer added or subtracted to obtain the series. Thus the atomic
weights of the three alkali metals then known were arranged
as a series 7+2w8, where w=0, 1, 2 . . . For example, Li=7 ;
Na=7+2x8=23; K=7-f2 X 8+2 X 8=39 ; and he obtained
analogous results with the alkaline earths, the chromium and
sulphur groups, and the halogens. Most probably in ignorance
of M. von Pettenkofer's series of 1850, modifications of
analogous series were advocated by both J. P. Cooke (1854)
and J. B. A. Dumas (1851), although the craze in quest of groups
and triads, more or less independent and distinct from one
another, seemed to divert attention from the establishment of
a continuous system including all the elements, and many felt
intuitively that the lists of Dobereiner's triads or Pettenkofet's
series were but fragments of a more general law — for instance,
A. Strecker in his Theorien und Experimente zur Bestimmung
der Atomgeivichte der Elemente (Braunschweig, 1859) :
It is scarcely reasonable to suppose that all the relations between
the atomic weights of chemically related elements are merely due to
chance. The discovery of the hidden meaning in these numerical
relations must be left to the future.
And M. Faraday (1852) :
We seem here to have the dawning of a new light, indicative of the
mutual convertibility of certain groups of elements, although under
conditions which as yet are hidden from our scrutiny.
' Some papers were published by A. E. B. de Chancourtois 3
between 1862 and 1863, in which it was proposed to classify
the elements by their atomic weights. A. E. B. de Chan-
courtois divided the circular base of a cyHnder into 16 parts,
representing the atomic weight of oxygen ; he then traced a
helix inclined at an angle of 45° ; and plotted the values of
the atomic weights as 'points caracteristiques vertical to the q 2 4- 6 8 lo I2 14 16
axes of the cylinder so that one-sixteenth part of a complete ^^^ i._a. E. B. de
rotation of the cylinder represented one unit. He called the chancourtois' Vis
helix so obtained the vis tellurique— the telluric screw. A , Tdlurique.
portion of the telluric screw unrolled from the surface of
the cylinder is illustrated in Fig. 1 as a plane surface. There is a tendency for
analogous elements to fall on the same vertical fine. The fundamental assump-
tion involved in this scheme is that the atomic weights of the elements can be
represented by whole numbers in accord with the formula w+16m, where m is an
integer. Hence he assumed that the differences between the atomic weights ought
to be constant. He tried to fill gaps in his Ust of elements by imaginmg new
varieties of the known elements— such as actually occurs in the case of carbon—
and these he called caracteres secondaires represented in Fig. 1 by a circle. A. E. B.
de Chancourtois' ideas were so much entangled with extraneous matters, and the
truth was so much obscured by useless and faulty speculations, that his work lay
^!.
r
^
^
^
^
^c
^.N
d^
^
S
F
^i
T'
^
^i-
^'
''i^
^
S
a
s
^i
^
^5^
§^
■ t:
Tn
^^
\
S^
^^-
^
^i
^
: -Vt
- ^
XZ'
^ ji
\
ZrJ
S^
S^
^
^
^^
-5s;
254
INOKGANIC AND THEORETICAL CHEMISTRY
buried for nearly thirty years, and it was only resurrected after Mendeleeff's system
had become famous. There are also vague hints in N. H. Marne's IJeher die Anzahl
der Elemente (Berhn, 1786), which, when read in the light of subsequent develop-
ments, have been taken as anticipations of the periodic law.
Again, between 1863 and 1866, J. A. R. Newlands ^ published a series of papers
in which he arranged the elements in the ascending order of their atomic weights,
and noticed that every succeeding eighth element was " a kind of repetition of the
first." Thus, copying Newlands' first table as it appeared in his communication
On the law of octaves (1865), where " the elements are arranged in the order of their
equivalents, with a few transpositions, it will be observed that elements belonging
to the same group usually appear on the same horizontal line."
Table II.^
-Newlands' Table
OF THE
Law
OF
Octaves
(1865).
No.
No.
No.
No.
No.
No.
No.
No.
H
1
F
8
CI
15
Co,
Ni22
Br
29
Pd
36
I
42
Pt, Ir
50
Li
2
Na
9
K
16
Cu
23
Rb
30
Ag
37
Cs
44
Tl
53
Gl
3
Mg
10
Ca
17
Zn
25
Sr
31
Cd
38
Ba,
V 45
Pb
54
B
4
Al
11
Cr
18
Y
24
Ce, La
33
U
40
Ta
46
Th
56
C
5
Si
12
Ti
19
In
26
Zr
32
Sn
39
W
47
Hg
52
N
6
P
13
Mn
20
A.S
27
Di, Mo
34
Sb
41
Nb
48
Bi
55
O
7
S
14
Fe
21
Se
28
Rh, Ru
35
Te
43
Au
49
Os
51
Note. — ^When two elements happen to have the same equivalent, both are designated
by the same number.
It will be observed, said Newlands, " that the number of analogous elements generally
differ by 7 or some multiple of seven ; in other words, members of the same group
of elements stand to each other in the same relation as the extremities of one or
more octaves in music. This peculiar relationship I propose to provisionally term
the law of octaves." Newlands noticed that elements belonging to the same group
usually appeared in the same column, and he declared that all the numerical relations
which had been observed among the atomic weights " including the well-known
triads, are merely arithmetical results flowing from the existence of the law of
octaves." Newlands' law of octaves did not attract much attention, probably
because faulty atomic weights seriously interfered with arrangement ; and because
the changes on triads and arithmetical series had been rung during the few preceding
years with tiresome persistence.
References.
1 J. W. Dobereiner, GUberfs Ann., 56. 332, 1816 ; 57. 436, 1816 ; J. B. A. Dumas, Gompt
Bend., 45. 709, 1857; 46. 951, 1858; 47. 1026, 1858: Ann. Chim. Phys., (3), 55. 129, 1859;
C. M. Despretz, Campt. Rend., 48. 362, 1859, 1817 ; Pogg. Ann., 15. 301, 1829 ; J. B. P. Kremers,
ib., 85. 56, 262, 1853 ; 99. 62, 1858 ; J. H. Gladstone, Phil. Mag., (4), 5. 313, 1853 ; W. Odling.
16., (4), 13. 423, 480, 1857 ; E. Lennsen, Liebig's Ann., 103. 121, 1857 ; M. von Pettenkofer, ib.,
105. 188, 1858 ; J. P. Cooke, Amer. Journ. Science, (2), 17. 387, 1854 ; M. Faraday, A Course of
Six Lectures on the Non-metallic Elements, London, 1852 ; Ostwald's Klassiker, 66, 1895.
2 W. Prout, Thomson's Ann* Phil., 6. 321, 1815 ; 7. Ill, 1816.
3 A. E. B. de Chancourtois, C(mipt. Rend., 54. 757, 840, 967, 1862 ; 55. 600, 1862 ; 56. 253, 467,
1217, 1863 ; Vis tellurique, classement naturel des corps simples ou radicaux obtenu au moyen
d'un systeme de classification helicoidal et numirique, Paris, 1863 ; L. de Boisbaudran and A. de
Lapparent, Campt. Rend., 112. 77, 1891 ; Chem. News, 63. 51, 1891 ; W. Crookes, ib., 63. 51,
1891 ; P. J. Hartog, Nature, 41. 186, 1899.
« J A. R. Newlands, Chem. News, 7. 70, 1863 ; 10. 59, 94, 95, 240, 1864 ; 12. 83, 94, 1865 ;
13. 113, 130, 1866 ; 25. 252, 1872 ; 26. 19, 1872 ; 27. 318, 1873; 32, 21, 1875 ; On the Discovery
of the Periodic Law, London, 1884.
THE CLASSIFICATION OF THE ELEMENTS 255
§ 3. The Periodic Law— D. I. Mendeleeff and L. Meyer
The periodic series is a brilliant and adequate means of producing an easily surveyed
system of facts which by gradually becoming complete will take the place of an assemblage
of the known facts.- — E, Mach,
D. I. Mendeleeff and L. Meyer, quite independently and, so far as we can tell,
quite in ignorance of Newlands' and Chancourtois' work, obtained a far clearer
vision of the law of octaves about 1869. D. I. Mendeleeff published his On the
correlation of the 'properties and atomic weights of the elements, in 1869, a year before
L. Meyer. Mendeleeff said : " When I arranged the elements according to the
magnitude of their atomic weights, beginning with the smallest, it became evident
that there exists a kind of periodicity in their properties." Otherwise expressed,
if the elements be arranged in the order of increasing atomic weights, the properties
vary from member to member in a definite way, but return more or less nearly to
the same value at fixed points in the series. D. I. Mendeleeff continued : " I
designate by the name ' periodic law ' the mutual relations between the properties of
the elements and their atomic weights. These relations are appUcable to all the
elements, and have the nature of a periodic function." Expressed more concisely,
Mendeleev's periodic law reads : The properties of the elements are a
periodic function of their atomic weights. A periodic function is one whose
value repeats itself at regular intervals. The interval is called a " period."
The ebb and flow of the tides, and the recurrence of the seasons are periodic
phenomena.
Mendeleeff' s table of the atomic weights was designed to tabulate the elements
in such a way as to exhibit the greatest number of relationships ; the early tables
were rather imperfect on account of imperfections in the atomic weight data, and
the paucity of our knowledge about the chemical characteristics of some of the
elements. The original tables were afterwards amended and modified owing to
improved data, and the discovery of new elements. The symbols of the elements
with their atomic weights have been arranged on a helix, on a spiral, and in numerous
other ways. Table III, not very different in style from one of Mendeleeff 's first
tables, is one of the simplest modes of arrangement, perhaps the best. The so-called
atomic numbers of the elements are indicated in brackets. These constants will be
discussed later. The fundamental principle of Mendeleeff's classification is that
the atomic weight of an element determines its position in the system.
When the elements are ranged in the order of their atomic weights, they form
(i) definite families or groups of elements with allied properties ; and (ii) series
in which the properties of allied elements recur in definite periods. The nine vertical
columns of the table are usually styled groups ; and the twelve horizontal lines,
series or periods. The properties of the elements and of their compounds are
consequently studied from the point of view of this system of classification. The
brief reviews of the family group given in this work will suffice to emphasize the
relationships of the members of any given group. The members of a group
have (1) general family properties like specific gravity, specific volume, laws of
combination which (2) gradually vary from the first to the last number, so that
the members of any particular group resemble one another more closely than do
any of the other elements. (3) Each family group differs from the others, but the
resemblances between the individual members of a family suggest that they have
been internally constructed on the same plan. The members of a series have
properties which (1) differ much from the first to the last member ; and (2) each .
series is more or less a repetition of that which precedes. D. I. Mendeleeff
emphasized the difference between the corresponding members of what he caUed
the odd series and the even series. The members of the odd series show com-
paratively greater analogies with the corresponding members of the odd series
than with the even series ; and likewise the members of the even series have more
256
INORGANIC AND THEORETICAL CHEMISTRY
1
1
• •.
.»
• •s
! i
as
1
1
1
s
1
1
:3i
e
0
s
1
>^.
S9
i
e
If
1
i^
^1
1
1
i!
1
mO
o
?
1 ! s
S
f
^rt
^
1 s r
1 ^
1 1
1 1
1
3
! ^ 1
i ^
1 1
1 1
1
' 1
^
s
o
s
CO
1
2 ^
11
1 II
la ;
10
i
t
1
2 i
1 1
2 s
1
s
CO
i 1
i
t
"*
Is
p §
' i
9
»
<
ill
i 1 ^
5
0
<
o
IS,
1
eo
« s
8 S i
E 1
1 i
M
t-l
o
i
1
s !
, « !
1 1
i!
: 1
n
bO
i
1
1 „_
s
«
8
T< f
S 5
? 1
i §^
o 1 H
^
8 i-
S bO
ss 1
p5 ^
i
3
M 0 1
« ^ 1
!s
s '
<
i^
▼^
y^
- S 1
^ 1
t 1
s
i
6
s
M
i
!
i '
■
1 : £
%
<
£
s
1
£
"" €
O
i
i
s :
1
1
1^
99
cc
•^ U5
» t-
00 OS
0 ^ i
i-H
1
' .
,
'. 1
. ]
~— ^—
— : — 1
.
1
(2
1
0
1
&
1
t3
t
o
4:9
tH
0
^
bU
^
^s^
1 i
F^
^
c
lJ 1
0
^
THE CLASSIFICATION OF THE ELEMENTS 257
analogies with corresponding members of the even series than with the odd series.
For example,
Group .
4th series
6th series
6th series
7th series
I II III IV
— Ti
K
Ca
Cu
Zn
Rh
Sr
V
VI
VII
V
Or
Mn
As
Se
Br
Nh
Mo
__
— Zr
Ag Cd In Sn Sb Te
where the 4th and 6th series resemble one another more closely than do the 4th and
5th, or the 5th and 6th. Each short period, it will be observed, contains eight
elements ; and each long period either contains nineteen elements of which three
are the so-called transition elements, or else it has provision made for nineteen
elements. Hjrphens are inserted in the spaces where the corresponding element
is unknown. The elements in the first short period are sometimes called group
elements or bridge elements, since they show a notable gradation of properties
from one to the other, and serve as links or bridges between the different groups.
The members of the next short period or series 3, are called typical elements
because they have the typical properties and characteristics of the group, and show
a rather wide divergence from neighbouring groups. After each typical element,
the different groups diverge into two SUb-groups.
The transitional elements. — It will be noticed that there is a distinct difference
between the members of the odd and the even series. The even series, say the
fourth and sixth, resemble one another more closely than the members of the odd
series, say the fifth and seventh. The lower oxides of the last members of the even
series resemble in many ways the first members of the odd series. Thus, the basic
oxides of chromium and manganese are in many ways similar to the oxides of copper
and zinc. Again, there are marked differences between the last members of the
odd series (halogens) and the first members of the next even series (alkali metals).
Those elements which cannot be placed in short periods fall in better with last
members of the even series, and the first members of the odd series. Thus, iron,
cobalt, and nickel fall between manganese and copper both with respect to chemical
properties and atomic weights :
Cr
Mn
Fe
Ni
Co
Cu
Zn
Atomic weight
52
54-9
55-8
59-0
58-7
63-6
65-4
Specific gravity-
6-9
7-4
7-8
8-7
8-8
8-9
6-9
Atomic volmne
7-5
7-4
71
6-8
6-7
71
9-5
SO also Ru — Rh — Pd -> Ag come just after the sixth series, and Os — Ir — Pt -> Au
after the tenth series. The inert gases are considered to form a kind of transition
between the last members of the odd series (halogens) and the first members of the
even series (alkali metals), and consequently also, they only occur in the horizontal
rows where transitional elements in the eight groups are absent.
The following arrangement. Fig. 2, modified from one by T. Bayley (1882),i
emphasizes the relationship and yet the individuality of the sub-groups, the character
of the transition elements, etc. Protyle represents an imaginarj^ primordial element
of elements, from which the ordinary elements are made ; and by extinct ekmenls
are understood imaginary elements of high atomic weight which may have once
been made from protyle, but which proved too unstable to endure under terrestrial
conditions and broke down into simpler elements of smaller atomic weight.
T. Bayley's table emphasizes the fact that while the atomic weights of the
elements progressively increase, their properties recur at definite intervals. No
well-known elements are omitted from the scheme, and with three exceptions the
order is that of the atomic weights, and the elements usually fall into virtually the
same groups as would have been obtained had they been arranged accordmg to
their chemical behaviour. Otherwise expressed, there is one element for each
place in the table, and each place in the table is intended for a defimte chemical
individual.
The valency of the elements shows a peculiar relation, for the maximum valency
VOL.
258
INORGANIC AND THEORETICAL CHEMISTRY
rises from 1 to 8 in passing along a given series from the first to the last group.
Thus,
Group
I
II
III
IV
V
VI
VII
VIII
Oxide
KjO
CaO
AlaOa
COa
P2O5
SO3
C1,0,
OS04
Valency .
1
2
3
4
6
6
7
8
The curve, Fig. 3, is obtained by plotting as abscisssB the atomic weights of the
elements, and for ordinates the higher oxides of the elements which correspond
Protyle
_.H :
He-Li-Be-B-C-N-O-F
I I I I I I I I
Ne Na Mcf Al Si P S CI
A K C
I
yir Rb %r Y Ir
! I I i I
Xe Cs Ba La Ce
1 1 1 1 1
i I I I I
Nt - Ra - Th
u Lx\ Ua Ue As Se Br
I i I I I I I
Ad Cd In %x\ Sb Te I
'^ ' I I I I I
r I
III III
Ta W - Os Ir Pt
i i 1 ill
I I I
Au Hd Tl
I r I
I I I I
Pb Bi
I I I I
' Extinct Elements. -••
Fig. 2. — ^T. Bay ley's Modification of the Periodic System.
to water and which can form hydrates with water, or unite together to form salts.
There are several elements whose highest salt-forming oxide, corresponding with
the family type, is not known, but, adds D. Carnegie 2 :
Chemistry is by no means a completely worked-out science, wanting nothing, and the
periodic law would be at fault did it fail to mirror forth such shortcomings and imper-
fections as still exist.
50 100 150 200
Fig. 3. — Periodic Curve of the Ideal Higher Oxides of the Elements.
250
The periodicity of the curve, Fig. 3, is perhaps the most prominent feature.
There is a blank for fluorine because it is not known to form any oxide ; there are
also blanks in places supposed to correspond with unknown elements. The elements
on the horizontal lines form families the members of which have many analogous
properties. Elements with decided metallic characters collect towards the troughs
of the wavy curve, and the non-metallic elements collect towards the crests. The
basicity of the oxides decreases in passing upwards from trough to crest. Thus,
lithium and sodium, and beryllium and magnesium oxides are strongly basic ;
boron and aluminium oxides are but feebly basic, and they also show acidic
THE CLASSIFICATION OF THE ELEMENTS
259
properties ; carbon and silicon dioxides are distinctly acidic, not basic ; while the
nitrogen, phosphorus, sulphur, and chlorine oxides are strongly acidic/
Again, the minimum valency rises from 1 to 4, and then falls to unity in passing
through the different groups. Thus,
Group
I
II
in
IV
V
VI
VII
Compound
KH
CaHa
(AIH3)
CH4
PH3
SH-
CIH
Valency
1
2
3
4
3
2
1
The maximum valencies of boron and aluminium may be quadri- not ter-valent
and if this suspicion proves well founded, these two elements will not fit the table!
Note the increasing acidity of the hydrides of carbon, phosphorus, sulphur, and
chlorine in passing from methane, CH4, to hydrogen chloride, HCl. The properties
of the hydrides of the elements in the first three groups are not so well known.
Similar remarks apply to the halides, and in no case is a simple halide known which
is higher in type than the maximum oxide indicated in the preceding scheme.
The quotient obtained by dividing the atomic weight of an element by its specific
gravity in the solid condition is called the atomic volume of the element. Con-
sequently, the atomic volume represents the number of cubic centimetres
occupied by an amount of the element equal to its atomic weight expressed
in grams. The magnitude of the atomic volume thus corresponds with the looseness
ou
1
—
1
~
'■"
■"
~
■'
"■
...
—
1
Cfi
"J
rR>
^
^ tt
K
1
V
1
1
1
Xe/
'^ ^n
Kr
\
/
I
7
1
l<>r
f
)Ka
: i!2
/ 1
1
.0
Rr
T
^V
1 1
c;
Na«.TI
l Zf
fe
p,,^'
a
e
1
Hi
^ or.
N^A 1
fJXx.'
" ^Zi
k^
v
Dk. _
- - m
'^ \^^'~ '
A«
. J c
;r X
4Sb
1
^r
J^
- -t
X ~,l
IP ~ iSc —
«"
Ga -
t-M
d
'
O.n ,
- U^
\
>i tNo/
ICnlMr
y^?
'Aq
XL
-h:/^"
Ee*
^1
Y^T
0
S r>
•
n JB
J_F^
Vo^V 1 1 1
1
-P"rrt
LL
L
_
L
_.
ii
ij
a
1 1
20
4^0
60
160
180
200
220 240
80 100 120 140
Atomic Weights
Fig. 4. — Relation between Atomic Volumes and Atomic Weights.
of texture or porosity, so to speak, of the solid element. In 1870, in a paper. Die
Natur der chemischen Eleniente als Function ihrer Atomgewichte, Lothar Meyer 3
showed that when the atomic volumes of the elements are plotted with the atomic
weights, a periodic curve showing a number of maximum and minimum points is
obtained, as illustrated in Fig. 4. Certain portions of the curve are incomplete
owing to the lack of data.
Most of the well-defined physical and chemical properties of the elements are
periodic ; for instance, specific gravity, atomic volume, melting point, hardness,
malleability, ductility, compressibility, coefficient of expansion, thermal conduc-
tivity, latent heat of fusion, refraction equivalents for light, colour, electrical con-
ductivity, magnetic power, etc. When the numerical values of these properties and
the atomic weights of the elements are tabulated on squared paper, a curve is
obtained which is broken up into periods as is the case with the atomic volumes
— Fig. 4. The specific heats of the elements are unique in furnishing a non-
periodic curve. According to Dulong and Pe tit's rule, if x denotes the specific
heat of an element with an atomic weight y, at ordinary temperatures, we have
aj7/=6"4. This is obviously a hyperbolic curve as indicated in Fig. 5, and not a
periodic curve like Fig. 4. J. Dewar's observations * of the specific heats of the
elements at 50° K., or —223°, give a periodic curve running almost parallel with
the atomic volume curve. Fig. 4.
There have been several attempts to represent the relation between the
260 INORGANIC AND THEORETICAL CHEMISTRY
atomic weight of an element and the order in which it stands in the list of elements
arranged in an ascending order of atomic weights, by algebraic formulai. E. J. Mills ^
used «<;=15(p—0"9375=^), where the value of p ranges from 1 to 16, and x from 1 to 50.
It is perhaps not surprising that the atomic weights can be represented by such
a formula since the term 0 9375 approximates so closely to a whole number that
any number can be expressed in decimals by this formula. J. H. Vincent repre-
sented the atomic weight of 58 elements by means of the formula w={n-{-2y^^,
where n represents numbers rising by unity for each successive element from
n=3 up to %=60. The results were good. J. B. Rydberg and F. M. Flavitzky used
trigonometrical functions, and G. J. Stoney a logarithmic function to represent
the magnitude as well as the periodicity of the atomic weights of the elements.
In 1870, H. Baumhauer ^ represented the relation between the elements
graphically by means of a spiral ; in 1876, L. Meyer used a helix after the manner
of A. E. B. de Chancourtois (1862) ; and various other forms of periodic curve
have been recommended by E. von Huth, W. Spring, J. E. Reynolds, W. Crookes,
G. J. Stoney, S. HaughtoD, E. Loew, and others. The periodic law of D. I.
Mendeleeff does not represent an uninterrupted function ; it does not correspond
with a continuous change of properties with a continuous variation of atomic weights.
MendeleefE therefore expressed himself against the attempt to represent the periodic
relations of the elements by geometrical curves. In the Faraday lecture, 1889,
D. I. Mendeleeff said :
The periods of the elements have a character very different from those which are so
simply represented by geometers. They correspond to points, to numbers, to sudden
changes of the masses, and not to a continuous evolution. In these sudden changes destitute
of intermediate steps or positions, in the absence of elements intermediate between, say,
silver and cadmiima, or aluminium and silicon, we must recognize a problem to which no
direct application of the analysis of the infinitely small can be made. Therefore, neither
the trigonometrical functions proposed by Rydberg and Flavitzky, nor the pendulum-
oscillations suggested by Crookes, nor the cubical curves of the Rev. Mr. Haughton, which
have been proposed for expressing the periodic law, from the nature of the case, can repre-
sent the periods of the chemical elements.
References.
1 T. Bayley, Phil. Mag., (5), 13. 26, 1882 ; T. Carnelley, Chem. News, 53. 157, 169, 183, 197, 1885.
2 D. Carnegie, Watfs Dictionary of Chemistry, 3. 813, 1893.
3 L. Meyer, Liebig's Ann. Suppl., 7. 354, 1870 ; OstwaWs Klassiker, 68. 1895.
* J. Dewar, Proc. Boy. Sac, 89. A, 158, 1914.
5 E. J. Mills, Phil. Mag., (5), 18. 393, 1884 ; (5), 21. 151, 1886; T. Carnelley, ib., (5), 29. 97,
1890 ; J. H. Vincent, ib., (6), 4. 103, 1902 ; G. J. Stoney, ib., (6), 4. 411, 504, 1902 ; Proc. May.
Soc., 44. 115, 1888 ; J. E. Reynolds, Journ. Chem. Soc., 81. 612, 1902 ; Chem. News, 54. 1, 1886 ;
A. Minet, Ccrnipt. Rend., 144. 8, 1907 ; N. Delauney, ib., 145. 25, 1907 ; 106. 1405, 1888 ; 109.
526, 1889 ; M. Zangerle, Ber., 4. 570, 1871 ; B. M. herach. Die Zahlenverhdltnissen des Planeten-
sy stems und die Atomgewichte, Leipzig, 64, 1879 ; B. N. Tchitcherin, Bull. Soc. Imp. Nat. Moscow,
1, 1890; J. B. Rydberg, Bihang.Vet. Akad. Svenska, 2, 1885; F. M. Flavitzky, ^ Function expressing
the Periodicity of the Chemical Elements, Kazan, 1887 ; Zeit. anorg. Chem., 11. 264, 1896 ; M.
Topler, Sitzber. Isis Dresden, 10, 1896.
® H. Baumhauer, Die Beziehungen zwischen dem Atomgewichte und der Natur der chemischen
Elemente, Braunschweig, 1870; L. Meyer, Die modernen Theorien der Chemie, Breslau, 1876;
L, R. Gibbs, Proc. Elliott Soc. Charleston, 77, 1875 ; E. von Huth, Das periodische Gesetz der
. Atomgevnchte und des naturliche System der Elemente, Frankfurt, 1884; W. Spring, Tableau
representant la loi periodique des elements chimiques, Liege, 1881 ; J. E. Reynolds, Chem. News,
54. 1, 1886 ; W. Crookes, i6.,54. 117, 1886 ; G. J. Stoney, ib., 57. 163, 1888 ; Phil. Mag., (6), 4.
411, 1902; S. Haughton, ib., (5), 58. 93, 102, 1888; E. Loew, Zeit. phys. Chem., 23. 1, 1897;
A. Bilecki, Zeit. anorg. Chem., 108. 113, 1919; 98. 86, 1916.
§ 4. The Gaps in Mendeleeff's Tables of the Elements
The periodic law has given to chemistry the prophetic power long regarded as the
peculiar dignity of its sister science astronomy. — H. C. Bolton.
Both Meyer and Mendeleeff considered it necessary to leave gaps in their tables
for undiscovered elements, and more particularly in order to keep certain related
THE CLASSIFICATION OF THE ELEMENTS
261
Mendeleeff boldly prophesied that the
1-00
075
0-50
elements in the same vertical column
missing elements would be
discovered later, and in
some cases even predicted
their properties in consider-
able detail. For instance,
when Mendeleeff announced
the law, there were two blank
spaces in group III, the mis-
sing elements were called
eka-aluminium and eka-
boron respectively ; and
another space below titanium
in group IV, represented a
missing element which in
this case was called eka-
silicon. The hypothetical
character of these elements
was considered to be an
inherent weakness of the
law, but the weakness was
turned to strength when
gallium, scandium, and germanium appeared duly clothed with those very properties
which fitted closely with Mendeleeff's audacious prognostications. This fit attracted
Table IV. — Comparison of Predicted and Observed Properties of Germanium.
0-25
n
~
~
—
—
~
—
—
—
—
—
-rLi
■■
~
~
~
~
~~1
'~
~
-
B«
»
U-
-48
i—
V
-
.
~
~!
r
^Ca
mI
>!•
K
.^
.r
O-
_
M"^^
^
\7
£
tSn^
ce
,
_
w_
_h
i^4^bA
J-
1 — 1
LJ
U
L^
'r
Alljl-
jej
1
'W'^
U—
lai]
L_
-T I
50
200
250
100 150
Atomic Weights
Fig. 5. — Curve showing the Relation between the Specific
Heats and Atomic Weights of the Elements (0°).
Eka-silicon, Es (predicted in 1871).
Atomic weight, 72.
Specific gravity, 5 '5.
Atomic volume, 13.
Element will be dirty grey, and on calcina-
tion will give a white powder of EsOo.
Element will decompose steam with difficulty.
Acids will have a slight action, alkalies no
pronounced action.
The action of sodium on ESO2 or on EsKgFg
will give the element.
The oxide EsOg will be refractory and have
a sp.gr. 4-7. The basic properties of the
oxide will be less marked than TiOo and
SnOg, but greater than SiO 2-
Eka-silicon will form a hydroxide soluble in
acids, and the solutions will readily de-
compose forming a metahydrate.
The chloride EsClj will be a liquid with a
boiling point under 100*^ and a sp.gr. of
1-9 at 0°.
The fluoride EsF^ will not be gaseous.
Eka-silicon will form a metallo- organic com-
pound Es(C2H6)4 boiling at 160°, and with
a sp.gr. 096.
Germanium, Ge (discovered in 1886).
Atomic weight, 72*3.
Specific gravity, 5-47.
Atomic volume, 13-2.
The element is greyish- white and on ignition
furnishes a white oxide GeOa-
The element does not decompose water.
The element is not attacked by hydrochloric
acid, but it is attacked by aqua regis.
Solutions of KOH have no action, but ife
is oxidized by fused KOH.
Germanium is made by the reduction of
GeOa with carbon, or of GeKjFs with
sodium.
The oxide GeOg is refractory and has a
sp.gr. 4-703. The basicity is very feeble.
Acids do not precipitate the hydrate from
dilute alkaline solutions, but from con-
centrated solutions, acids precipitate GeO,
or a metahydrate.
Germanium chloride, GeCl4, boils at 86°,
and has a sp.gr. at 18°, 1*887.
The fluoride GeF4.3H20 is a white solid mass.
Germanium forms Ge(C2Hj)4, which boils at
160°, and has a specific gravity slightly
less than wat«r.
262 INORGANIC AND THEOKETICAL CHEMISTKY
considerable attention, and served to strengthen the faith of chemists in the funda-
mental truth of the periodic law. In illustration, the predicted properties of eka-
silicon and subsequently discovered properties of germanium are quoted side by side
in Table IV.
The confirmations of MendeleefE's predictions of the properties of eka-aluminium
(gallium), and of eka-boron (scandium) were equally striking. This dramatic
achievement focused attention on the generalization ; but it is only fair to say
that the predictions and their subsequent verification are not such positive proofs
of the truth of the periodic law as some suppose. It is certainly wrong to say,
as C. Winkler did, " it would be impossible, to imagine a more striking 'proof of the
doctrine of periodicity of the elements than that afforded by this embodiment of
the hitherto hypothetical eka-silicon," because gaps appeared in some of the older
systems of classification, and the properties of the missing members could have been
predicted, and atomic weights estimated by analogy with the other members of the
family, quite independently of the periodic law, and in some cases with better
results.
§ 5. The Application of the Periodic Law
A natural law only acquires scientific importance when it yields practical results, that
is, when it leads to logical conclusions which elucidate phenomena hitherto unexplained,
when it directs attention to occurrences till then imknown, and especially when it calls forth
predictions which may be verified by experiment. — D. I. Mendeleeff.
D. I. MendeleeS pointed out that the periodic law could be employed in : 1. The
classification of the elements ; 2. The estimation of the atomic weights of elements
not fully investigated; 3. The prediction of the properties of hitherto unknown
elements ; and 4. The correction of atomic weights.
1. The classification of the elements.— T. H. Huxley (1864) has said : " By
the classification of any series of objects, is meant the actual or ideal arrangement
together of those which are like, and the separation of those which are unlike ; the
purpose of this arrangement being to facilitate the operations of the mind in clearly
conceiving and retaining in the memory the characters of the objects in question."
If a new element possesses one of the marked characteristics of a given class, it
follows that it will probably possess the remaining characteristics. If a new element
of the alkaline-earth family be discovered, a great many other properties could be
inferred with a high degreee of probability of their being right. In fine, enthusiasts
say that the periodic system is superior to all the older methods of classifying the
elements, for the law makes it possible to build up a system of the greatest possible
completeness free from much arbitrariness, and it furnishes strong circumstantial
evidence of the correctness of the reasoning employed by Cannizzaro in deducing
values for the atomic weights of the elements.
2. The estimation of the atomic weights of the elements. — On account of
practical difficulties, it is not always possible to fix the atomic weight of some
elements by vapour density determinations (Avogadro's rule), and by specific heat
determinations (Dulong and Petit's rule), and the atomic weights of these elements
were frequently assigned on somewhat uncertain grounds. According to C. L.
Winkler, indium has the equivalent weight 37 "8. The correct atomic weight must
be some multiple of this, and for no special reason, the atomic weight was once
taken to be 37'8x2=75*6. In that case, indium would fall between arsenic and
selenium where it would be quite mis-matched. Mendeleef! proposed to make
indium tervalent, like aluminium, so that the atomic weight became 37"8x3==113'4,
and the element fell in the table between cadmium and tin where it fits very well.
The subsequent determination of the specific heat of indium, 0*0577, corroborated
the change made by Mendeleeff in the atomic weight -from 75*6 to 113-4. Beryllium,
THE CLASSIFICATION OF THE ELEMENTS 263
uranium, and a number of the rare earths at one time did not fit very well into the
table, but MendeleefE's alteration of the supposed atomic weights to make these
elements fit the table were subsequently justified by vapour density determinations
of the volatile chlorides, or by specific heat determinations.
3. The prediction of the properties of hitherto undiscovered elements.
— When an empty space occurs in Mendeleeff's table, it is assumed that an element
will one day be discovered which will fill that place ; and conversely, if a new element
were found to correspond with a place in the table already filled, it would be sus-
pected that the supposed element is not really elemental. In attempting to imitate
Mendeleefi, and predict the properties of missing elements in the table, attention
is paid to the composition and properties of the more important compounds —
hydroxides, oxides, haloid salts, etc. — so as to bring out (1) the family characters
of the group to which it belongs ; (2) the character of the series to which it belongs ;
(3) its position in the series and group so that a comparison can be made with the
properties of other known elements similarly situated in neighbouring groups or
series ; and (4) the relations of the particular group and series in which it occurs
with other groups and series. In order to avoid introducing new names when speak-
ing of unknown elements, represented by gaps in the table, Mendeleeff designated
them by prefixing a Sanscrit numeral — eha (one), dwi (two), tri (three), etc. — to
the names of the preceding analogous elements of the odd or even numbered series
of the same group. Thus, the unknown elements of group I will be called eka-
C8esium,and dwi-csesium. Were strontium unknown, it would be called eka-calcium.
In addition to the prediction of germanium, gallium, and scandium already dis-
cussed, Mendeleefi foretold the possible discovery of eka- and dwi-ctesium ; of eka-
niobium — -En=146 ; of eka-tantalum — Et=235 ; of dwi-tellurium— Dt=212 ;
and of the analogies of manganese : eka-manganese— Em=100 ; and tri-manganese
-Tm=190.
The case of the so-called inert gases is of more recent date. The discovery
of argon and helium could not have been predicted from Mendeleefi's periodic law,
but after these elements had been discovered, and accommodated in the periodic
table between the strongly acid halogen family and the strongly basic alkaU metals,
the probable existence of other similar inert gases was indicated. When an ex-
haustive search was made, krypton, neon, and xenon were discovered with pro-
perties and atomic weights which could have been predicted from the arrangement
which was made for argon and helium in Mendeleeff's table.
4. The correction of the values of atomic weights.— If the atomic weight
of an element does not fit with the regular course of, say, the atomic volume curve.
Fig. 4, the atomic weight is probably in error. Thus, the atomic weights of
platinum, iridium, and osmium at that time were probably too high, and subsequent
determinations verified this inference. For example, the atomic weights of these
elements were :
In 1870
In 1919
§ 8. Some Defects in the Periodic Law
The scientific value of thoroughly sound hypotheses is enhanced daily both by known
facts that they are continually assimilating, and new facts that they are continuaUy
revealing. — J. Ward (1899).
There are some misfits in the Mendeleeff's table as we have it to-day, owing
to the fact that at least three pairs of elements would be mis-matched if they were
• simply classed according to their atomic weights : argon (39*88) and potassium
(39-10) ; cobalt (58-97) and nickel (58-68) ; and tellurium (127-5) and iodine (12692).
Platinum.
Iridium.
Osmium.
196-7
196-7
198-6
195-2
193-1
190-9
264 INORGANIC AND THEORETICAL CHEMISTRY
G. Kriiss and F. W. Schmidt (1889)i attributed the difficulty with Cobalt and nickel to
the presence of a hitherto undiscovered element in nickel which they named gnommm.
This explanation, however, had to be discarded. It did not survive the ordeal
remorselessly applied to conjectures of this kind. No gnomium has yet been
found. Again, the case of iodine and tellurium has been studied with relentless
vigour stimulated largely by D. I. Mendeleeff's prediction : " The atomic weight
of tellurium 7nust be between 123 and 126, and cannot be 128." Iodine most
certainly belongs to the same group as the other halogens, and tellurium undoubtedly
belongs to the selenium group ; these elements are accordingly placed among
their own family relations in spite of the fact that if their atomic weights were
alone considered tellurium woidd be ranked with the halogens, and iodine with
selenium. B. Brauner (1889) suggested that ordinary tellurium is a complex
containing a- and ^-tellurium ; and it was inferred that true tellurium — say a-Te —
has an atomic weight 125, and that the other form of this element has a higher
atomic weight, and will find a place in the periodic system in the valency below
tellurium. D. I. Mendeleeff called this undiscovered element dwi-tellurium, Dt,
and he sketched some of its physical and chemical properties ; but tellurium, said
G. WyroubofE, has been tortured in every conceivable manner : it has been melted,
sublimed, oxidized, hydrogenized, phenylated, dissolved, crystalHzed, fractioned,
and precipitated ; yet nothing but failure has followed all attempts to get an
atomic weight lower than iodine or to fraction the element into two others. Nothing
has developed which would warrant a belief in Mendeleeff's " must." Hence, in
spite of the fact that " the laws of nature admit of no exception," faith in the law
has led chemists to allocate these discordant elements according to their chemical
properties and not according to their atomic weights. To put the matter bluntly,
the procedure runs : It is necessary either to reject the periodic law or to reject
the number 127'5 for tellurium ; the periodic law cannot be rejected because it is
the very embodiment of truth, nay, truth itself ; ergo, in spite of all evidence to the
contrary, the number 127"5 must be wrong. Bode's law of astronomy successfully
predicted the asteroids and allocated their proper place in the solar system ; but
the subsequent discovery of Neptune did not agree with Bode's law. The law was
accordingly abandoned and it is now regarded as a curiosity. Mendeleeff's law
may have to go the same way. B. Brauner's assumption that tellurium is a mixture
of true telluriimi with a higher homologue, may be a good working hypothesis for
stimulating experiments on this element, but to maintain the thesis against all
evidence to the contrary ' ' may afford an easier way out of the difficulty than by
working steadily at the cause of the discrepancy, but it affords at best a
feeble and undignified cover for one's retreat." This method must be dubbed
unscientific, but the circumstantial evidence justifies the procedure in the ex-
pectation that a consistent system will ultimately grow from the truth and error
engrafted into the " law." It is not very probable that the principle underlying
the periodic law will be abandoned because it is founded on a vast assemblage
of facts of different kinds ; and because it seems to be plastic enough to fulfil
subsequent requirements. The central problem in inorganic chemistry, said
W Ramsay (1904), is to answer the question : Why this incomplete concordance ?
Allocation of hydrogen. — The location of hydrogen in the table, as already
indicated, is difficult, because hydrogen seems to be without companions. It is
univalent, and thus falls either with the alkali metals (D. I. Mendeleeff, 1869 ;
G. Martin, 1901) or with the halogens (0. Mason, 1896 ; W. Crookes, 1898 ; W.
Ramsay, 1901). Although D. I. Mendeleeff 2 rather inclined to the beUef that
hydrogen occupies an " isolated independent position," he said that in virtue of
" its salt-like oxide H2O, and the salts H, it must stand in the first group ; " that
" the nearest analogue to hydrogen is sodium which also stands in an odd series of
the first group ; " and that " the more remote analogues of hydrogen are copper,
silver, and gold." The attempts to displace hydrogen from its position at the head
of the alkaU metal group, and to place it with the halogens have not been very
THE CLASSIFICATION OF THE ELEMENTS 265
successful, but in either case many of the arguments appear rather strained and far-
fetched ; they run pro et con. somewhat as follows :
(1) Unlike hydrogen, the monad alkali metals appear to bo monatomic, but hydrogen
too is probably monatomic at a high enough temperature.
(2) If placed at the head of the halogen table, hydrogen is in close contiguity with the
other gaseous elements, but the extreme mobility and lightness of the hydrogen molecules
may be a powerful factor in determining its gaseity ; after all gaseity is a mere accident
of temperature.
(3) Hydrogen is electropositive like the alkali metals, but it is not now considered to
be a metal ; hydrogen does not exhibit the metallic properties characteristic of the family
of alkali metals, and towards lithium it behaves like nitrogen, oxygen, and the halogens
in forming a hydride. This argimient is of little weight when no objection is raised to the
allocation of nitrogen and bismuth ; or of carbon and lead in the one family group.
(4) The difference between two consecutive elements \isually ranges between 15 and 20,
and this agrees better with superposing hydrogen above fluorine than above lithiimi (7) ;
as G. Martin (1901) has pointed out this argument simply depends on the arbitrary selection
of subtraction as a criterion ; if division be selected, quite a different conclusion is obtained.
Thus, progressing upwards from potassium, the ratio K : Na = l-7 ; Na : Li = 3-3 ; and,
following the same rule, Li : H = 6-9, which is near to the observed value.
(5) If hydrogen be placed above lithium, six gaps for undiscovered elements are crowded
in between hydrogen and helium, or helium must come in an unnatural intermediate
position, say, above carbon or nitrogen. In view of the gaps in the old periodic tables
which were subsequently filled, there is, however, no particular objection to the assumption
that these undiscovered elements have a real existence even if they have not yet been
discovered.
(6) The mutual replacement of hydrogen and the metals which has led to the acids
being regarded as salts of hydrogen, establishes a clear analogy between hydrogen
and the alkali metals ; against this it must be remembered that there is an equally striking
analogy between hydrogen and the halogens, for these elements can mutually replace one
another in many organic compounds with no more effect on the general properties of the
resulting compounds than is produced by the siibstitution of one halogen with another.
This argument loses much weight if it be remembered that the behaviour of a compound
is determined by its constitution rather than by the chemical nature of the atoms them-
selves, and that " the most diverse radicles may displace other radicles in a compound
and perform a similar function to that of the displaced radicles without materially affecting
the fundamental characteristics of the body into which they have entered."
(7) If the behaviour of the halogens towards oxygen be selected as a criterion, the
diminishing stability of the oxygen compounds with diminishing atomic weight culminates
in fluorine. No stable oxides are known, hydrogen oxide is a very stable compound,
totally unlike the halogen oxides. The halogens in their known oxides have a maximtim
valency of seven, while the maximum valency of hydrogen is one.
(8) Then again, there is a great contrast between the stable hydrogen compounds with
the halogens, and the instability of the hydrides of the alkali metals. In a rough sort of
way the former property suggests dissimilarity ; the latter, similarity. Hence also
B. Braimer (1901) asks : How can such a positive element as hydrogen stand at the head
of such negative elements as the halogens ? The elements at the head of a sub-group are
always more negative and less positive than the lower members of that sub-group.
Accordingly, it will be evident that the position of hydrogen has not been definitely
settled, and that hydrogen appears to be a rogue element quite out of place in the general
scheme. Some suppose that hydrogen is a member of an extinct or yet undiscovered
series of independent elements, but whether hydrogen is the alpha or the omega
is indeterminate because it would be eHgible for a place either in group I or group
VII according to the properties selected for comparison. The supposed first
member of the series is called " proto-fluorine " ; so also the elements " proto-
beryllium " and " proto-boron " have been invented, the former with an atomic
weight 1-33, and the latter, 2. All this, however, is mere speculation.
Allocation of the rare earths. — This also presents some difficulties. Most
of the rare earths can be distributed about the table according to their atomic
weights, or they can be relegated to a class by themselves. B. Brauner (1902),
who has made a special study of the rare earths, considers that they should all be
grouped together with cerium between barium and tantalum so that " Ce, 140"25
in the table stands for : Ce, 140'25 ; Pr, 140-6 ; Nd, 144-3 ; Sa, 150-4 ; Eu, 152 ;
Gd, 157-3 ; Tb, 1592 ; Dy, 1625 ; Er, 167-7 ; Tm, 1685 ; Yb, 172-0 . . . This
has been called the asteroid theory of the raie earths. The properties of the
266
INORGANIC AND THEORETICAL CHEMISTRY
rare earths, however, are not well enough known to give us much confidence in
the various proposals which have been made for dealing with them ; and con-
sequently, Mendeleeff considered that the inotallation of these elements should
be deferred ; a similar remark appUes to the radioactive elements. Here F. Soddy
and A. Fleck (1913) 3 assume :
All members occupying the same place in the periodic system are chemically identical
with one another, and are not separable from one another by chemical process, although
their atomic weights may vary over several units.
The rare earths do not fall all in the same group in this sense because several of the
members fit well enough into the table, thus, ytterbium — Yb, 172 — fits into group
III, series 10, etc. The so-called isotopic elements will be discussed later.
K the properties of the elements are dependent on their atomic weights
the existence of two elements with different properties and approximately
the same atomic weights should be impossible. Hence the difficulty with
elements hke cobalt and nickel ; ruthenium and rhodium, etc. The peculiarities
of these elements would never have been suspected from the periodic law. It might
also be added that some experiments with the radioactive elements have led to the
inference that " different elements not necessarily of identical atomic weight, do
occupy the same place in the table, and that when this occurs, these elements possess
an identical chemical nature." The evidence as to the identity of chemical pro-
perties is not very strong when it is remembered how very few chemical tests have
been made owing to the small amount of available material. Not very long ago
praseodymium and neodymium were considered to have identical chemical
properties.
Twin elements. — R. Lorenz^ has shown that certain elements have atomic
weights which approach each other in pairs, and which differ from each other by
at most 1-4 units ; and he appUes the term twin elements to pairs of elements
whose atomic weights approach one another very closely— within li — of one another.
For example :
Diff.
Diff.
Boron-carbon .
. 1-009
Nickel-cobalt .
. 0-660
Sodimn -magnesium .
. 1-322
Seleniima-bromine .
. 0-893
Aluminium-silicon .
1-320
Palladium-silver
. 1-238
Phosphorus-sulphur .
. 1033
Tin-antimony
. 1-190
Potassium-calcium .
0-864
Iodine-tellurium
. 0-736
Vanadium-chromium
0-940
Tantalum-timgsten .
. 1-20C
Manganese-iron
0-910
Lead-bismuth
. 1-099
The elements usually show many similarities in their chemical behaviour, and their
separation presents some difficulties. Most of the twin-elements usually follow
one another in immediate succession, so that the atomic weight of a member of one
pair differs from that of the corresponding member of the next pair by approximately
4 or a multiple of 4, e.g. Na and Al, 4*022 ; Mg and Si, 4*02 ; Al and P, 3-95 ;
Si and S, 3 "663 ; etc. Lorenz shows that elements which do not form twin pairs
may follow this rule if they be regarded as representing twin pairs with other
unknown elements — the exceptions are H, Be, N, Zn, Ga, Rb, Y, Zr, Nb, In, Cs, Ba,
Ir, Au, and some rare earths elements.
Some elements are allocated places in the table according to their atomic
weights in opposition to their chemical properties. For instance, copper,
silver, and gold fall into one group with the alkali metals. The tervalency of gold
appears to be unconformable with the valency of its companions although in its
present position the series PtCl4, AuCls, HgCl2, and TlCl is suggestive. Berylhum
is pecuharly placed from this point of view. Thallium is very like lead, but its
sulphate and some other salts are quite different from lead salts. At least three
pairs of elements have been placed according to their properties irrespective of their
atomic weights, as indicated by the misfits mentioned in the preceding section.
Again, the so-called type-elements, Li, Be, B, C, N, 0, F, which stand at the heads
THE CLASSIFICATION OF THE ELEMENTS 267
of the family groups — ^the vertical columns of MendeleefE's table — usually have
properties quite at variance with the other members of the family. In 1870,
MendeleefE attributed this to their low atomic weight, for he said :
The elements of the first two series have the least atomic weights, and in consequence
of this very circumstance, although they bear the general properties of the group, they
still show many peculiar and independent properties.
The difficulty still remains, for these elements have not yet been altogether reconciled
to the groups to which they should be closely analogous. The test of any given
classification of the elements arises when the arguments why a given element should
be included rather in one class than in another are reviewed. For instance, in spite
of the unique properties of fluorine or of lithium, could the former be included in
any group other than the halogens, or lithium in any group other than the alkali
metals ? The answer is in the negative.
Some elements which appear to be chemically similai are separated
in the table. For example, copper and mercury ; silver and thallium ; barium
and lead ; etc. The position of these elements in the table gives no hint of these
characteristics. Still, it might be argued that these elements exhibit many essential
difEerences. Thus, the physical properties of the cupric and mercuric chlorides
and sulphates show great contrasts. The stabihty of cuprous and mercurous
chlorides is also very difierent. Lead and barium dioxides appear to have a different
constitution. The unstable thalUum sesquioxide, TI2O3, corresponds with the other
— more stable — sesquioxides in the group, but there are many important points
of resemblance between thallium and the alkaU metals, and between silver and
lead. The extension of the periodic law to include compounds as well as elements
is not always satisfactory. Many examples will appear when the different family
groups are reviewed ; and J. Locke (1898) ^ inquires : Why should the relations
between magnesium and zinc be emphasized and the closer relation between
magnesium and manganese be ignored as if the explanation were not conditioned
by an equally important law of nature ? The worship of the periodic law as a fetish
may stimulate the pursuit of remote analogies in one direction, and close the door
to the search for closer analogies in other directions.
Multivalent elements. — According to Mendeleeff, when an element Hke copper
forms two series of compounds in one of which it has the same valency as its neigh-
bour in a horizontal row, the compounds of the neighbouring elements are similar.
This is confirmed by the close resemblance between the bivalent compounds of
copper and zinc ; but, on the other hand, the close proximity of scandium to
titanium does not seem to confer on tervalent titanium compounds any of the
characteristic properties of scandium. Hence, Mendeleeff's classification does not
make clear the fact that heterologous elements — i.e. elements belonging to different
groups in the periodic table — may give analogous compounds when in the same
form of combination below their highest valency — e.g. silver and thallous chlorides.
The compounds of ferric iron resemble those of tervalent aluminium and chromium,
while those of ferrous iron are like those of bivalent zinc and magnesium. Again,
the cuprous and cupric compounds have Uttle more in common than has hydrogen
sulphide and sulphuric acid— each pair of compounds has the same element— copper
in the one case, sulphur in the other. Still further, the compounds of bivalent
vanadium resemble those of magnesium ; tervalent vanadium, those of alummium ;
quadrivalent vanadium, those of silicon; and quinquivalent vanadium, those of
phosphorus. Hence, J. Locke (1898) and G. A. Barbieri (1914) recommended that
the compounds of an element with different valencies be regarded as belonging to
so many different elements. Ferric and ferrous iron are just as distinct primary
forms of matter as electricity and heat are forms of energy, and the one can be con-
verted into the other, or into metalHc iron. When a ferric salt is reduced to a
ferrous salt, or into metallic iron, the form of matter analogous to tervalent iron
has ceased to exist.
268 INORGANIC AND THEORETICAL CHEMISTRY
The higher oxides. — G. Wyrouboif represents the actual state of our knowledge
of the higher oxides of the elements by a chart constructed like the ideal curve,
Fig. 3. The selection of the characteristic oxides by Mendeleeff (Fig. 3) is quite
arbitrary, there appears to be no guiding principle. Sometimes it is the lower
oxide which is selected — e.g. BaO in place of Ba02 ; sometimes a higher — e.g.
Mn207 in place of Mn02, MnO.. etc. Sometimes it is the more stable oxide — e.g,
BaO and not Ba02 '■> ^^d sometimes it is the less stable one — e.g. CU2O, not CuO.
The curve of the actual oxides will doubtless be modified by future researches,
but it is far less regular and has more the character of a zig-zag line. No
longer can the higher oxides ranged along the same horizontal line be said to
have any relation with their chemical analogies ; the best established of which
may disappear ; and harmonious order is replaced by la plus ahsolue anarchie.
Against these views it has been urged that Mendeleeff purposed selecting the
highest salt-forming oxide in his table, and that he did not regard such oxides
as K2O2, Ba02 as salt-forming oxides. ^ If he distinctly specified the
salt-forming oxides, it is urged that the generalization cannot be impugned by the
consideration of another class of oxides altogether. Mendeleeff claims that the
true peroxides, Ba02, Cr207, Ti02, H2O2, cannot form corresponding salts, whereas
Pb02 and Bi205 are distinctly salt-forming oxides in that the one corresponds
with the plumbates, and the other with the bismuthates. However, the existence
of the persulphates, pertungstates, and permolybdates does not harmonize with
D. I. Mendeleeff's views. The explanation of D. I. Mendeleeff, moreover, does not
account for copper— cuprous oxide, CU2O, should be the maximum salt-forming
oxide of this element, whereas cupric oxide, CuO, is the commoner salt-forming
oxide.
G. Wyrouboff, in a paper, Sur la classification periodique des elements (1896),7
has the idea that the periodic system is a very interesting and highly ingenious
table of the analogies 'and the dissimilarities of the simple bodies — a mere catalogue
raisonne of the elements ; and further, allowing Mendeleeff's dictum — the laws
of nature admit of no exception — the periodic law must be accepted or rejected
as a whole ; there are numerous cases where the periodic law conflicts with facts ;
ergo, the law ought to be rejected. G. Wyrouboff adds that the periodic classification
is a sterile combination of numbers ; it is an ingenious combination of observations
arbitrarily selected ; and one of those vague formulae which satisfies subjective
conceptions, without corresponding with any objective reality. His pro-
posal to reject the periodic law is somewhat precipitate, for we do not feel quite
satisfied that the supposed misfits are not due to defective knowledge. Some can
see a distinction between failure and incomplete success. Although a single con-
tradictory fact is fatal to a hypothesis, the hypothesis is not to be rejected on the
first prima facie conflict with reality. This does not justify the relegation of in-
consequent facts into obscurity, but rather indicates the need for keeping them in a
conspicuous position. The law may have to be ultimately abandoned ; it is only
excessive zeal which could say qu'aucun argmnent chimique ne pouvait pretaloir
contre la hi periodique. What M. Faraday ® said a century ago (1819) apphes
to-day :
Much as the present stage of knowledge owes to the tendency of the human mind to
methodize, and therefore to facihtate its labours, still it may complain that in some direc-
tions it has been opposed and held down to error by it. All method is artificial and all
arrangement arbitrary. The distinction we make between classes, both of thoughts and
things, are distinctions of our own ; and though we mean to found them on nature, we are
never certain we have actually done so. That which appears to us a very marked distinctive
character may be really of very subordinate importance, and where we can perceive
nothing but analogies and resemblances, may be concealed nature's greatest distinctions.
THE CLASSIFICATION OF THE ELEMENTS 269
References.
* G. Wyrouboff, Les actualites chimiques, 1. 18, 1896 ; B. Brauner, Journ. Chem. <Sfoc., 55.
382, 1889 ; L. Staudenmeier, Zeit. anorg. Ghent., 10. 189, 1895 ; D. I. Mendeleefif, Journ. Ghent.
8oc., 55. 634, 1889; G. Kruss and F. W. Schmidt, Ber., 22. 11, 2026, 1889.
2 D. I. Mendeleeff, Jcmm. Buss. Ghent. Soc., 1. 1869 ; G. Martin, Ghent. News, 84. 9, 73,
154, 1901 ; B. Brauner, ib., 84. 233, 1901 ; 0. Masson, t6.,73. 233, 1896 ; W. Crookes, Proc. Roy.
Soc., 63. 408, 1898.
3 A. Fleck, Journ. Ghent. Soc, 103, 381, 1052, 1913; F. Soddy, Nature, 91, 669, 1913.
4 R. Lorenz, Zeit. anorg. Ghent., 12. 329, 1896 ; Ghent. News, 74. 211, 234, 1896.
5 J. Locke, Anter. Ghent. Journ., 20, 581, 1898.
« D. I. Mendeleeff, Ber., 15. 242, 1882.
' G. Wyrouboff, Les actualites chimiques, 1. 18, 1896.
8 B. Jones, The Life and Letters of Faraday, London, 1. 304, 1870.
CHAPTEE VII
HYDEOQEN
§ 1. The Occurrence of Hydrogen in particular and of the Elements
in general
It can be said that hydrogen is more widely distributed in the universe than any other
element, since its presence has been recognized spectroscopically in all the heavenly bodies.
— E. Baub (1907).
The element hydrogen occurs free in nature in comparatively small quantities.
Free hydrogen has been detected in the gaseous exhalations from volcanoes and
fumaroles. Thus E,. Bunsen,i in 1853, found 25*14 per cent, of hydrogen, along
with carbon dioxide and hydrogen sulphide, in the gases from the Reykjalidh
fumaroles in Iceland ; H. Moissan (1902) reported 8" 12 per cent, of free hydrogen
in the gases from Monte Pelee (Martinique), where the disastrous eruption of 1902
occurred ; A. L. Day and E. S. Shepherd found 10*2 per cent, of hydrogen in the
gases from the crater at Kilauea, Hawaii, in 1912 ; and F. Fouque, 29*43 per cent,
in the gases from Santorini (^gean Sea).
Table I.— Composition of Some Natural Gases.
EeykjaUdh,
Monte Pelee,
KUauea,
Santorini,
Stassfurt
Iceland.
Martinique.
Hawaii.
^gean Sea.
Mines.
Hydrogen
2514
8-12
10-2
29-43
93-050
Oxygen ....
i —
13-67
. —
0-32
0-185
Nitrogen ....
0-72
54-94
11-8
32-97
5-804
Carbon dioxide
50-60
15-38
73-9
36-42
0-180
Carbon monoxide
. — .
1-60
4-0
. — .
. — .
Siilphur dioxide
. — .
. — ,
. — .
. — .
—
Hydrogen sulphide .
24-12
—
—
^-
^-~
Ethane
—
5-46
■ —
0-86
0-778
Hydrogen has also been detected in natural gas. The presence of hydrogen
in these gases has been denied, although many published analyses include free
hydrogen 2 — for instance, S. P. Sadtler reported 4:'79 to 22*50 per cent, of hydrogen
in the gases from the petroleum springs of Pennsylvania ; C. C. Howard found an
average of 1*76 per cent, of hydrogen in natural gas from Indiana and Ohio ; E. H. S.
Bailey found none in the gas from Kansas, and F. C. Phillips none in the gas from
Vancouver ; free hydrogen has been found among the gaseous inclusions in the
anhydrite or salt deposits of Stassfurt, Leopoldshall, and Wiehczka ; H. Precht's
analysis is given in Table I. Confirmatory observations have been made by
H. Rose, H. Precht, E. Reichardt, E. Erdmann, etc.^ Occluded gases containing free
hydrogen have been obtained from granites and other rocks .^ Thus, W. A. Tilden
found :
H2
61-68
88-42
12-49
61-93
36-15
270
Granite (Skye)
Gabbro (Cornwall)
Pyroxene, Gneiss (Ceylon)
Gneiss (Seringapatam) .
Basalt (Antrim) .
Na
COa
CO
CH4
513
23-60
6-45
302
1-90
5-50
216
2-03
1-16
77-72
8-06
0-56
0-56
31-62
5-36
0-51
1-61
32-08
20-08
10-00
HYDROGEN 271
A. Gautier obtained 13461 c.c. of hydrogen from 100 grams of granitic rock. A
small part of the gases existing in rocks is entrapped in minute cavities and pores.
Some suppose that the larger part is occluded in the mineral much as palladium
can occlude hydrogen, while others assume that the gases are not free in the minerals
but are liberated through chemical reactions when the mineral is heated in vacuo
— e.g. methane comes from organic matter present ; carbon oxides from carbonates ;
nitrogen from nitrides ; and hydrogen from the decomposition of steam by iron or
its oxides or salts. Free hydrogen has also been found in the occluded gases of
meteorites— in quantities varying from 0'2 to nearly 50 c.c. per gram of soUd.
T. Graham obtained 16'53 c.c. of gas from 5*78 c.c. of the Lenarto meteoric iron, and
the gas contained 85" 68 per cent, of hydrogen by volume ; 9'86 of nitrogen ; 4'46
of carbon dioxide ; and no carbon dioxide. The meteorite contained 90*88 per
cent, of iron ; 8-45 of nickel ; 0-66 of cobalt ; and 0*002 of copper. A. W. Wright's
analyses of the gases occluded in iron and stoney meteorites gave
CO2
CO
CH^
Ha
N2
Iron type
8-6-14-4
12-5-67-7
18-2-76-8
1-5-51
Stone type
. 35-4-8 10
1-7-4-4
00-3-6
13-1-57-9
1-7-3-5
h=Q
h=20
^=80
h=im
ft=800
10X1013
8X1013
43X1012
182X10"
3X1010
78X101'
43xl0i«
520X1011
35X10'
0
21X101'
7x1016
25X1011
3X10«
0
W. Ramsay and M. W. Travers found a gram of a sample of meteoric iron from
Toluca contained 2*8 c.c. of occluded hydrogen ; from Charcas (Mexico), 0*28 c.c. ;
and from Rancho de la Pila (Mexico), 0*57 c.c. W. E. Hidden, W. M. Flight,
J. Dewar and G. Ansdall, M. W. Travers, J. W. Mallet, etc., have also found hydro-
gen in meteorites. Hydrogen has also been reported by J. Parry, and A. Pictet
and L. Ramsey er as occluded in cast iron, steel, coke, and coal.
Although hydrogen is being constantly liberated from the earth's solid crust,
the proportion actually found in the atmosphere at sea level is very small ; at
higher altitudes, the proportion is probably greater, but the atmosphere is there
so attenuated that the actual amount is very small. J. H. Jeans estimates that
the number of molecules of hydrogen per c.c. at a height h kilometres, is
Hydrogen
Nitrogen
Oxygen
F. C. Phillips 5 could detect no hydrogen at a height of 7000 feet above sea level,
but owing to exhalations from natural gas and the decay of organic matter, he believes
it will be found at still higher levels ; H. and E. Erdmann found it at a height of
800 metres. J. L. Gay Lussac could find none in air collected at an altitude of
6,636 metres during his balloon ascent in 1804. The estimates which have been made
of the amounts of free hydrogen in the atmosphere are somewhat discrepant. J. B. A.
Dumas and J. B. J. D. Boussingault's analyses of air in 1841, indicated the presence
of hydjogen, but they thought the gas must be present in air in the form of methane,
CH4. In 1898, A. Gautier calculated that 100 litres of air contain from 11 to 18
c.c. of hydrogen ; and, in 1900, G. D. Liveing and J. Dewar found hydrogen and
helium to be always present in the more volatile portions obtained by the fractional
distillation of liquid air. Lord Rayleigh's estimate, from spectroscopic observations,
is about one-sixth of that of A. Gautier ; and A. Leduc's observations on the density
of air also indicate that A. Gautier's estimate is too high. G. Claude found air to
contain less than O'OOOOl part of free hydrogen.
Spectroscopic observations 6 of the sun's chromosphere show what appear to be
stupendous red flames of incandescent hydrogen with calcium and a few other
elements in some cases towering over 300,000 miles (M. Fenyi, 1892) into space,
and 100,000 miles in width (C. A. Young, 1872)— thousands of times larger than
the earth on which we live. These prominences, as they are called, have been observed
to shoot nearly half a million miles within 10 minutes of time. Spectroscopic
observations also show that hydrogen is present in nebulae and certain stars.
Combined hydrogen is common. Water contains one-ninth of its weight of
272
INORGANIC AND THEORETICAL CHEMISTRY
hydrogen. We really know nothing about the hydrogen as it is combined with
oxygen in water. The fact is that when water is decomposed under certain con-
ditions, this proportion of hydrogen is obtained. It is the foQon jparler to say that
the compound contains the element, or that the element occurs in or is present in
the compound, when the element can be obtained from the compound by suitable
methods of decomposition. The occurrence of an element, refers not only to the
conditions under which the free element may be found, but also to those natural
compounds which contain the element united with other elements. Hydrogen
occurs combined not only with oxygen as water, but also with sulphur as sulphuretted
hydrogen ; with chlorine as hydrochloric acid ; with nitrogen as ammonia ; and
more rarely combined with phosphorus, iodine, bromine, and carbon. It is one
of the chief constituents of animal and vegetable tissue. Hydrogen also is present
in nearly all organic compounds, in all acids, and in many gases — the hydrocarbons
(petroleums), hydrogen sulphide, etc.
A. Gautier (1901) attributes the formation of free hydrogen in nature, (1) to the
decomposition of water at a red-heat by ferrous salts, etc. ; (2) to the decomposition
by heat of hydrocarbons formed from natural carbides ; and (3) to the decomposi-
tion of such compounds as iron nitride by steam — 2NFe3-f 6H20=2NH3+6FeO
-I-3H2, a reaction observed when a crystalline iron nitride from Etna lava is treated
with water.
Quantitative distribution of the elements. — By comparing a large number
of analyses of rocks, etc., F. W. Clarke (1916) "^ has tried to estimate the percentage
composition, by weight, of the earth's crust (J mile deep) together with the ocean
and the atmosphere. His result is :
Per cent.
Per cent.
Per cent.
Oxygen .
. 5002
Magnesium .
. 2-08
Barium
0-08
Silicon .
. 25-08
Hydrogen .
0-95
Manganese
0-08
Aluminium
. 7-30
Titanium
0-43
Strontium
0-02
Iron
. 4-18
Chlorine
0-20
Nitrogen
0-03
Calcium .
. 3-22
Carbon
0-18
Fluorine
010
Sodium .
. 2-36
Phosphorus .
Oil
Bromine
0-008
Potassium
. 2-28
Sulphur
0-11
All other elements
0-41
He also emphasizes the fact that in the solid crust, the lighter elements
predominate over the heavier, so that the abundant elements all have an atomic
weight less than 58. The heavier metals occur only in trivial amounts. The
mean density of the earth — 5*4-5*6 — however, is about double that of the average
of the surface rocks. This has led to the assumption that the heavier elements
are concentrated in the interior — a supposition, says F. W. Clarke, which is possibly
true, but unprovable.
If we try to get an estimate of the relative number of atoms of the different
kinds of elements distributed in the half-mile crust, the ocean and the atmosphere,
Clarke's numbers must be divided by the corresponding atomic weights of the
elements. We thus obtain for the percentage number of atoms in the half-mile
crust :
Oxygen .
Hycfiogen
Silicon .
Aluminium
63-81
Sodium
16-30
Magnesium
15-87
Calcium
4-68
Iron
1-72
1-61
1-40
1-29
Potassium
Carbon .
Titanium
Chlorine
102
0-27
0-16
0-11
This gives a better idea of the relative distribution of the elements from the
chemical point of view than the actual weights in the preceding list.
The occurrence of the elements and the periodic law. — ^D. I. Mendeleeff ^ i^as
drawn attention to the fact that the elements which occur most abundantly on the eariKs
surface have small atomic weights ; the converse does not necessarily apply, for some
elements with small atomic weights — e.g. lithium (7), beryllium (9), and boron (11)
— are by no means abundant. There are some exceptions to D. I. MendeleefE's
rule ; strontium, for instance, appears to be less abundant than barium. The
HYDROGEN 273
elements in the groups of MendeleefE's table often occur more abundantly in passing
from the first to the second member, and afterwards decrease with increasing
atomic weight— ^.^r. the alkali family ; the tetrad group ; etc. J. H. Gladstone
argued that the average vapour density of the elements which are plentiful is less
than that of the elements which are common ; and with those which are common, it
is less than with the rare elements ; and with the rare elements less than with the very
rare elements. Consequently, says J. H. Gladstone, as the earth cooled from the
vaporous state, those elements having the least vapour density must have tended
to remain near the surface, while those with a high vapour density accumulated more
towards the centre, and therefore occur most rarely on the surface crust.
In a paper On the periodic law, and the occurrence of the elements in nature
(1884), T. Carnelley has made a special study of the occurrence of the elements
from the point of view of the periodic law. He shows that with the exception of
carbon, nitrogen, oxygen, sodium, magnesium, aluminium, and silicon, the elements
belonging to the odd series of Mendeleeff's table are, as a rule, easily reduced to the
free state, while those elements belonging to the even series are usually reduced with
difficulty. The exceptional elements correspond with the exceptional character of
the atomic volume curve — Fig. 4, Cap. VI — where the curve instead of contmuing to
fall as it reaches carbon, begins to rise until it comes to sodium. The ready re-
ducibihty of the elements of the odd series corresponds with their common occurrence
in the free state. Excepting carbon, nitrogen, oxygen, and group VIII, the
eletnents belonging to the even series do not occur in the free state, v)hereas elements
belonging to the odd series generally, and sometimes frequently, occur free. The following
elements of the odd series, for instance, are frequently found in a free state : copper,
silver, gold, mercury, arsenic, antimony, bismuth, sulphur, selenium, tellurium,
lead, and tin, while gallium, indium, and thalhum are so sparsely distributed
that not enough is known to justify a definite statement about them. The four
halogens — fluorine, chlorine, bromine, and iodine — zinc, and phosphorus are the
only notable exceptions because sodium, magnesium, aluminium, and sihcon have
already been accounted for, and the exceptions, carbon, nitrogen, and oxygen, in
the even series also belong to the peculiar part of the atomic curve. The elements
of group VIII all occur native, and this tendency is the more marked the greater
the atomic weight. The four halogens are the most electro- negative of the elements,
and they occur in nature united with the most electro-positive elements, and except-
ing a few secondary products — oxy chlorides and sulphochlorides — are never found
in combination with oxygen or sulphur.
T. Carnelley further shows that excepting the halogens and the members of group
VIII, the elements belonging to the odd series of Mendeleeff's Table rarely occur in nature
as oxides, but usually occur as sulphides {or double sulphides), selenides, tellurides, or
arsenides — i.e. in combination with negative elements belonging to an odd series.
In illustration, the following elements of the odd series commonly occur as sul-
phides, selenides, tellurides, or arsenides — copper, silver, zinc, cadmium, mercury,
gallium, indium, thallium, lead, antimony, sulphur, selenium, and tellurium. On
the other hand, arsenic, bismuth, and tin are generally found in this form ; while gold,
sodium, magnesium, aluminium, sihcon, and phosphorus rarely if ever occur so com-
bined. The tendency of the mernbers of the odd series in any particular group to occur
in nature as sulphides increases, and the tendency to occur as oxides or double oxides
diminishes as the atomic weight increases. For example, in group II, magnesium nearly
always occurs as double oxide (carbonate, siHcate, etc.) ; zinc occurs commonly as
sulphide, sometimes as oxide ; cadmium, found as sulphide, never as oxide ; and
mercury as sulphide, or metal, never as oxide. In group IV, sihcon occurs always
as oxide or double oxide (silicates) ; tin, almost always as oxide, sometimes as
sulphide ; and lead generally as sulphide, rarely as oxide. Similarly with the
other groups.
On the other hand, elements belonging to the even series usually occur as oxides
or double oxides — silicates, carbonates, sulphates, alumiinates, etc.— i.e. in combination
VOL. T. T
274 INORGANIC AND THEORETICAL CHEMISTRY
with an element belonging to an even series. In illustration, the following elements
of the even series commonly occur as oxides or double oxides — lithium, potassium,
rubidiimi, caesium, beryllium, calcium, strontiimi, barium, boron, scandium, yttrium,
lanthanum, ytterbium, carbon, titanium, zirconium, cerium, thallium, vanadium,
niobium, didymium, tantalum, chromium, terbium, tungsten, and manganese.
Molybdenum and manganese (rarely) occur as sulphides. Nitrogen (in the nitrates)
and molybdenum frequently occur as oxides or double oxides.
Excepting iron, nickel, and cobalt the elements of group VIII rarely occur in
a combined state. Iron usually occurs as oxide, frequently as sulphide ; cobalt
usually as sulphide or arsenide, and sometimes as oxide ; while nickel almost always
resembles the elements of the even series, cobalt the elements of the odd series.
Hence the three triads of elements in group VIII show a gradual passage from
the even to the odd series.
T. Carnelley summarizes the facts : Elements standing on the falling portions
of the curve of atomic volumes are reducible with difficulty, and never occur in a free
state in nature or in combination as sulphides, but always in combination with oxygen,
forming oxides or double oxides— -e.g. silicates, sulphates, carbonates, etc. Elements
on the rising portions of the curve are easily reducible, and almost always occur more
or le^s in the free state in nature, and also in combination with sulphur, but rarely
with oxygen. It is possible that the exceptions are apparent, not real, and that
with increasing knowledge the anomalies will disappear, for the coincidences are
too many, and the exceptions too few, to lead to any conclusion other than that
the relative abundance of the elements is somehow connected with their position
in the periodic classification.
W. D. Harkins^ has noticed that when the elements are arranged in the order
of their atomic numbers, the even numbered elements are in every case more abundant
in meteorites than the adjacent odd numbered elements ; similar remarks apply
to the earth's lithosphere. In all cases also the elements of low atomic number
and low atomic weight occur much more abundantly than elements with a high
atomic weight.
References.
1 R. Bunsen, Pogg. Ann., 38. 215, 1853 ; H. St. C. Deville and F. Leblanc, C(ympt. Rend.,
47. 317, 1858 ; H. St. C. Deville, F. Leblanc, and F. Fouque, ib., 55. 75, 1862 ; 56. 1185, 1863 ;
H. Moissan, ib., 135. 1085, 1902 ; F. Fouque, Santorini et ses eruptions, Paris, 1879 ; A, L. Day
and E. S. Shepherd, Compt. Rend., 157. 958, 1027, 1913 ; A. Gautier, ib., 150. 1564, 1910.
2 C. Engler, Ber., 21. 1816, 1888 ; S. P. Sadtler, Rep. Second Geol. Sur. Pennyslvania, 1. 146,
1876 ; C. C. Howard, Ann. Rep. U.S. Geol. Sur., 11. i, 592, 1891 ; F. C. Phillips, Journ. Amer.
Chem. Soc., 20. 696, 1898; Amer. Chem. Journ., 16. 406, 1894; E. H. S. Bailey, Univ. Quart.
Kansas, 4. 1, 1895 ; A. L. Day and E. S. Shepherd, Bull. U.S. Geol. Sur., 24. 673, 1913 ; A. Brun,
Quelgues recherches sur le volcanisme, Geneve, 1911 ; W. Libby, Amer. Journ. Science, (3), 47.
372, 1894; J. Janssen, Compt. Rend., 64. 1303, 1867 ; 97. 601, 1883.
3 H. Rose, Pogg. Ann., 48. 353, 1839 ; R. Bunsen, ib., 83. 197, 1851 ; J. B. A. Dumas, Ann.
Chim. Phys., (3), 43. 316, 1855 ; H. Precht, Ber., 12. 557, 1879 ; E. Reichardt, Arch. Pharm., (2),
103, 347, 1860 ; E. Erdmann, Ber., 43. 777, 1910 ; J. Acosta, Compt. Rend., 36. 779, 1853 ;
W. Lindgren and F. L. Ransome, Prof. Paper U.S. Geol. Sur., 54. 252, 1906.
* A. Gautier, Compt. Rend., 131. 647, 1276, 1900; 132. 58, 189, 740, 932, 1901; 142.
1382, 1465, 1906; 148. 1708, 1909; 149. 84, 1909; Ann. Mines, (10), 9. 316, 1906; Bull. Soc.
Chim., (4), 5. 977, 1909; K. Hiiltner, Zeit. anorg. Chem., 43. 8, 1905; W. A. Tilden,
Chem. News, 75, 16^, 1897; Proc. Roy. Soc., 59. 218, 1896; W. Ramsay and M. W.
Travers, ib., 60. 442, 1897 ; T. Graham, Proc. Roy. Soc., 15. 502, 1867 ; J. W. Mallet, ib., 20. 365,
1872 ; F. Wohler, Pogg. Ann., 146. 297, 1872 ; W. E. Hidden, Amer. Journ. Science, (3), 31. 461,
1886 : A. W. Wright, ib., (3), 9. 294, 459, 1875 ; (3), 11. 253, 1876 ; (3), 12. 165, 1876 ; M. W.
Travers, Proc. Roy. Soc, 64. 130, 1898; 60. 156, 1896; J. Parry, Amer. Chem., 4. 254, 1874;
A. Pictet and L. Raraseyer, Ber., 44. 2486, 1911 ; T. Graham, Compt. Rend., 64. 1067, 1867;
J. W. Mallet, Amer. Journ. Science, (3), 2. 10, 1871 ; J. Dewarand G. Ansdell, Proc. Roy. Inst.,
11. 541, 1886; W. M. Flight, Phil. Trans., 173. 893, 1882.
6 J. L. Gay Lussac, Ann. Chim. Phys., (1), 72. 265, 1809 ; J. B. A. Dumas and J. B. J. D.
Boussingault, Compt Rend., 21. 1005, 1841 ; Ann. Chim. Phys., (3), 3. 257, 1841 ; A. Gautier,
Ccmipt. Rend., 127. 693, 1898 ; 130. 1353, 1667, 1677, 1900 ; 131. 13, 86, 535, 647, 1276, 1900 ;
135. 1025, 1902 ; 136. 21, 598, 1903 ; Ann. Chim. Phys., (1), 22. 5, 1901 ; Lord Rayleigh, Phil.
HYDROGEN 275
Mag., (6), 3. 416, 1902 ; A. Leduc, Compt. Rend., 135. 860, 1332, 1902 ; 136. 21, 1903 • G Qaude
t&., 148. 1454, 1909 ; G. D. Liveing and J. Dewar, Proc. Roy. Soc., 67. 468, 1900 : F. C PhiUips'
Journ.Amer Chem. Soc, IT 801, 1895; H. and E. Erdmann, Ergeb. Konig. Preuss. Aeronaut]
Obs 6. 221, 1911; J. H. Jeans, The Dynamical Theory of Oases, Cambridge, 356, 1916;
A Gautier, Compt Rend., 132. 58, 189, 1901 ; R. T. Chamberlin, Tfie Oases in Rocks, Washington,
1908.
« J. N. Lockyer, Compt. Rend., 86. 318, 1878 ; A. Corau, ib., 86. 315, 530, 1878'; C. A. Young
Amer. Journ. Science, (3), 20. 353, 1880 ; H. W. Vogel and A. Paalzoflf, Ber., 13. 274, 1880.
' F. W. Clarke, The Data of Geochemistry, Washington, 34, 1916.
8 D. T. Mendeleeff, Zeit. Chem., 5. 405, 1869; J. H. Gladstone, Phil. Mag., (5) 4 379 1877 •
T. Carnelley, ib., (5), 18. 194, 1884 ; J. Waddell, Chem. News, 113. 289, 1916! ' ' ' '
^ W. D. Harkins, Journ. Amer. Chem. Soc, 39. 856, 1917.
§ 2. The Preparation and Purification o! Hydrogen
The progress of aeronautics, and the use of hydrogen for the so-called hardening
or hydrogenation of oils, and the synthesis of ammonia, has incited industrial
chemists to improve the existing processes and to develop new methods for preparing
this gas. The use of hydrogen for the inflation of balloons was proposed by J. A. C.
Charles in 1783, soon after H. Cavendish's work on inflammable air ; and the first
balloon sent up from British soil on November 25, 1793, was inflated with hydrogen.
The lifting power of hydrogen is about 1*2 kilograms per cubic metre, or about
68J lbs. per 1000 c. ft.
If the molecular weight M (hydrogen=2) of a gas be expressed in ozs., a gram-molecule
of the gas will occupy nearly 22-3 c. ft., at n.p.t. ; an ounce of the gas wiU thus occupy
22'3JM c. ft., and a pound of the gas will occupy nearly 357/M. c. ft., or a cubic foot of the
gas will weigh ilf/357, or 0'0028M lb. Again, the lifting power of a given volume, v, of any
gas is equal to the difference between the weight of the gas and the weight of an equal
volimie of air at the same temperature and pressure. The lifting power of a gas of mole-
cular weight M at sea level is therefore 0-0028?; (28-98— M) lbs., so that the lifting power of
500,000 c. ft. of hydrogen at n.p.t. is nearly 38,000 lbs. If a gas had a lifting power greater
than hydrogen its molecular weight would be less than 2, and in the extreme case, a gas
with a vanishingly small atomic weight would have a lifting power about 7^ times that of
hydrogen .
The gas for dirigible balloons must be free from such impurities as are liable to
attack the fabric of the balloon ; and for military purposes as small a plant and as
small a weight of material as is practicable must be employed. Cost is not then of
prime importance. In addition hydrogen is also used as a combustible in special
cases — e.g. the oxy hydrogen blowpipe ; for the manufacture of quartz glass ; the syn-
thesis of gems — e.g. rubies and sapphires ; in the melting of platinum ; and in the
autogenous welding of steel, iron, copper, and various alloys. It is also used mixed
with nitrogen as an inert atmosphere in the manufacture of tungsten filaments for
lamps, and in fusing tungsten powder into rods.i
Some processes for the preparation of hydrogen have been previously discussed.
The more important methods which have been suggested for preparing hydrogen
are based upon the electrolysis of water, the decomposition of water by chemical
reactions, the action of metals on dilute acids, the decomposition of metal hydrides,
the action of metals and alloys on alkaline lye, the decomposition of hydrocarbons,
and the formation of hydrogen as a product of secondary importance— i.e. as a by-
product—in processes pursued for other substances called the primary or main
products of the reaction.
(1) The electrolysis of aqueous solutions. —Highly ipmifiedw&tei is a poor conductor
of electricity, but if the water be made slightly acid or alkahne it can be readily
electrolyzed. Hydrogen gas is liberated at the cathode, oxygen at the anode. In
addition to the work of decomposition, a certain proportion of current is expended
in warming up the liquid, so that the total energy expended is the sum of that
276 INORGANIC AND THEORETICAL CHEMISTRY
absorbed as heat and that which does chemical work. The heating effect is pro-
portional to the resistance, R ohms, and to the square of the current, C amps. The
total electrical energy of the current — C amps, and E volts — is CE watts ; the
chemical work done by the current is CE—C^R, and this is approximately 69 Cals.
per gram-molecule of water. One coulomb of current decomposes 0*0933 mgrm
of water ; and gives 0"0829 mgrm. or 0058 c.c. of oxygen ; 0'01034: mgrm. or
0*1150 c.c. of hydrogen ; that is, 0*0933 mgrm or 0*1725 c.c. of mixed gases. Other-
wise expressed, one ampere hour of current decomposes 0*3351 grm. of water
liberating 0*298 grm. or 207*2 c.c. of oxygen ; 0*0373 grm. or 414*4 c.c. of hydrogen ;
or 0*3353 grm. or 621*6 c.c. of mixed gases. With an intensity of 1*5 volts, virtually
no current passes through the electrolyte, because at least that potential is required
to start the electrolysis and provide the energy for the decomposition. With a rather
higher voltage, a current passes through the electrolyte, but no visible evolution of
gas occurs with electrodes of the usual size until the solution is saturated with
hydrogen and oxygen gases. A solution of one part of sulphuric acid with ten of
water, in an electrolytic cell fitted with platinum electrodes, is readily broken down
by a current with an intensity of not less than 1*7 volts ; sodium hydroxide
solutions have a similar minimum voltage of 1*69 volts, and potassium hydroxide,
1*67 volts. The electrolysis proceeds with a copious evolution of gas when a current
with an intensity of 2 or 3 volts is used.
The strong tendency towards oxidation at the anode and towards reduction
at the cathode renders the electrodes particularly liable to attack during the elec-
trolysis. Graphite or carbon anodes suffer some oxidation ; platinum and gold
anodes are not attacked ; and lead anodes are superficially oxidized, but the film
of lead dioxide seems to protect the metal from further attack. Gold cathodes are
not perceptibly attacked ; platinum cathodes slowly blacken, probably owing to
the absorption of hydrogen and its subsequent decomposition whereby a film of
platinum-black is formed; and lead cathodes slowly blacken and disintegrate super-
ficially so that powdered lead collects on the floor of the cell. The action is probably
similar to that with platinum. Iron makes a good cathode with alkaline solutions.
The electrodes, of course, must also resist the chemical action of the acid or alkaline
electrolyte. If the cathode can unite with an alkali metal, or occlude or
dissolve hydrogen gas, it is liable -to disintegrate rapidly, presumably because
of the successive formation of an alloy and its decomposition by the water of the
electrolyte.
The proportion of gases liberated during the electrolysis of water is always less than
the theoretical amount, and the volume of oxygen less than half that of hydrogen.
Temporar}^ losses at the beginning of the electrolysis may occur through the absorp-
tion of hydrogen at the cathode, or of oxygen by the superficial oxidation of the
anode. These losses, however, cease when the cathode is saturated, and when a
continuous protective coat of oxide is formed on the anode. There is a certain loss
caused by the recombination of hydrogen and oxygen dissolved in the electrolyte
to form water. The dissolved hydrogen unites with oxygen at the anode ; and the
dissolved oxygen with the hydrogen at the cathode. The diminution in the volumes
of the two gases is in the same proportion as they are evolved. If the electrolysis
proceeds while the system is under pressure, more amperes will be transmitted for
the same applied voltage, or the same number of amperes can be passed with a smaller
applied voltage. In this case, the output of gas per ampere hour decreases, and the
heating effect is diminished, because, under pressure, (i) more hydrogen and oxygen
dissolve in the electrolyte and this increases the amount of recombination ; and (ii)
the conductivity of the solution is increased while the resistance and consequently
also the heating effect is decreased.
In the electrolysis of dilute sulphuric acid, oxygen may be lost owing to the for-
mation of persulphuric acid or persulphates. The amount so lost depends on the
temperature, current density, and the concentration of the acid. Thus, 0. Schon-
herr2 found that at ordinary temperatures the percentage amounts of oxygen lost in
HYDKOGEN 277
the formation of persulphuric acid with acids of different concentrations — specific
gravities — and different current densities—amperes per sq. metre — ^are :
Sp. gr. H2SO4 . . 1-15 1-20 1-25 1-30 135 1-40
15 amps. . — — — 118 39 230
50 amps. . — 4*4 29*3 47*2 60-5 67-7
100 amps. 7-0 20-9 43-5 51-6 71 -3 75-7
are
In commercial work the concentration of the acid and the current density
adjusted to fall well below the minimum here indicated. The persulphuric acid in
solution may pass by diffusion or circulation to the cathode where it is reduced
by the hydrogen back to sulphuric acid. Under these circumstances the net result
is a loss of hydrogen and oxygen in the proportions in which they are liberated —
the persulphuric acid acts as an intermediate compound. The higher the tempera-
ture the less the loss of oxygen by the formation of persulphuric acid, and running
the cell warm is one of the best remedies. There may be losses of oxygen from the
anodic formation of hydrogen 'peroxide. The conditions which particularly favour
the formation of this compound — over 60 per cent, sulphuric acid — do not usually
obtain in the electrolytic cells used for the generation of hydrogen and oxygen.
There is also a tendency to form a little ozone, which decreases the total volume of
oxygen given off — but the presence of ozone is often an advan-
tage since the oxygen gas is then a more efficient oxidizing Oj^ygen^ Hydros^-
agent. The conditions which favour the formation of ozone
are high current density ; highly concentrated sulphuric acid —
say up to 50 per cent. ; and low temperatures. The amount
of ozone formed is small when working with a low current
density, very dilute acid, and warm solutions.
When but small quantities of electrolytic hydrogen are
required in the laboratory the anode is formed by immersing
the platinum wire in a mass of liquid zinc amalgam (Fig. 1)
which absorbs the oxygen to form zinc oxide and finally zinc
sulphate ; the cathode is a platinum plate. A current from an
accumulator giving between 4 and 6 volts, suffices to work the ^^- ^;^^gen**^on
cell charged with dilute acid (1 : 10) — the gas is comparatively ^ ^^lyi scale of
pure. If alkali lye and nickel electrodes be used, the hydrogen electrolysis
will be contaminated with hydrocarbons derived from traces
of carbides, etc., in the caustic alkali ; but, according to H. B. Baker (1902),3
an aqueous solution of purified crystalline baryta gives a gas of the highest degree
of purity. H. B. Baker used a platinum plate as cathode, and a platinum wire
dipping 'in zinc amalgam for the anode. The zinc amalgam absorbs the oxygen.
According to E. W. Morley, if the alkaline solution used in the electrolysis
contains carbonates, hydrocarbons may be formed.
On the industrial scale, a great many systems for the electrolysis of water have
been designed and used ; many of these are described in V. Engelhardt's Die Elec-
trolyse des Wassers (Halle a. S., 1902) and M. N. Schoop's Die industnelle Electrolyse^
des Wassers (Leipzig, 1901). In most of the arrangements, dilute sulphuric acid
is electrolysed between lead electrodes, or sodium hydroxide solutions between iron
or nickel electrodes. The main practical difficulties are : (i) The slight attack of
the electrodes by the electrolyte, and (ii) the necessity of preventing the admixture
of the two gaseous products of the electrolysis without increasing the internal re-
sistance of the cell too much bv means of diaphragms, say asbestos. There are four
leading types of cell : * (1) The filter-press type ; (2) the tank type : (3) tHe Dell
type ; and (4) the metal partition type. * . .
0. Schmidt'sfilter-presstype of cell is based on the fact that if a conducting material
be placed between the anode and cathode of an electrolytic cell, the system behaves
as if it were reallv two cells, because while hydrogen is liberated at the cathode and
oxygen at the anode, oxygen is also liberated from the side of the diaphragm tacing
278
INORGANIC AND THEORETICAL CHEMISTRY
Rubber Rings
/TTT'
Oxygen
OuHel-.
Diaphragms^
Liberated.
"Diaphragms.
Fig. 2. — Diagrammatic Arrangement of O. Schmidt's
Filter Cells.
the original cathode and hydrogen from the side facing the original anode — pro-
vided the voltage drop between the original anode or cathode and the conducting
partition is less than the minimum voltage required for electrolysis. The cell is
made of recessed iron plates and they are clamped together with an intermediate
partition with a rubber border and porous asbestos cloth in the centre. Hence,
a series of each pair of plates forms one cell and two half cells. The iron plates are
thus insulated from one another
so that if the cells be filled
with electrolyte, and the end
plates be made respectively
positive and negative poles,
the current passes from pole to
pole through the electrolyte
and the iron plates. A series
of these plates are clamped
together as indicated in Fig. 2.
The hydrogen and oxygen
liberated at alternate sides of
these plates pass through suit-
able passages into fcoUecting
pipes. The asbestos partitions
prevent the mixing of the gas.
The hydrogen is about 99 per
cent, purity, the oxygen 75J
per cent. ; about 5*9 c. ft. of hydrogen and 3 c. ft. of oxygen at mean temperature
and pressure are collected per kilowatt hour ; with a series of 40 plates, about 2*5
volts are absorbed in each cell and the current density about 2 amps, per sq. deci-
metre. The apparatus is compact, but the rubber joints of the diaphragm require
close attention to prevent leakage.
The International Oxj^gen Co.'s tank tjrpe of cell has a mild steel tank fitted with
an iron cylinder perforated with holes and hung from the cast-iron lid of the tank by
means of a conducting rod. An asbestos curtain surrounds the
inner cylinder. The lid of the tank is insulated from the outer walls
and from the central rod. The cell is filled with a 10 per cent, solu-
tion of sodium hydroxide. The lid is also fitted with suitable outlet
pipes so that during the electrolysis the hydrogen liberated from the
walls of the tank and the oxygen liberated on the walls of the inner
cylinder are collected in separate pipes. At 20°, and 2992 in. pressure,
3'051 c. ft. of oxygen and 5*950 c. ft. of hydrogen were collected per
kilowatt hour, and the purity of the hydrogen was 99*70 per cent.,
and of the oxygen, 98"34 per cent. In M. U. Schoop's bell type of
cell, a perforated tube — made of iron if the electrolyte be alkaline,
and of lead if the electrolyte be acid — is surrounded by a glass or
porcelain tube as illustrated in Fig. 3. The electrodes are arranged
in pairs as shown in the diagram. Two pairs are placed in each cell.
One pair acts as anode, the other pair as cathode. The glass tubes
collect the gases, and there is little risk of mixing the gases. In
^^U S~h ^^^ Schuckert's system, the electrodes are arranged so that the gas from
Electrodes. each electrode collects in a separate cell. In P. Caruti's diaphragm
type of cell, pierced metal diaphragms are employed. The electro-
lysis of brine solutions for caustic alkali and chlorine furnishes hydrogen as a
by-product. The cost of the electrolytic process per 2000 cubic feet of hydrogen
(and 1000 cubic feet of oxygen) is about 19^. on the assumption that the electrical
energy costs \d. per unit.
(2) The decomposition of water. — A large number of oxidizable substances can
be employed for this purpose. In his Essai de mecanique chimique fondee sur la
I Gas
T. OuHeh.
9
ii
i
HYDROGEN 279
ihermochimie (Paris, 2. 521, 1879), M. Berthelot showed that in very many cases,
the decomposition of water with the liberation of hydrogen can be effected by
substances, which, in uniting with the oxygen of the water, give off more heat than
occurs when hydrogen unites with the same proportion of oxygen. The results of
the action of the alkali metals and the metals of the alkaline earths, previously
indicated, are represented by the equations :
iNa H-Ol-H Na-O-H , H ^. ;;"7H-0:-H H . 0-H
iNa+H;^0|-H = Na-O-H+H' ^^^ j'^^H-Oi-H = H+^*<0-H
In the former case, one atom of hydrogen in each molecule of water, H — 0 — H, is re-
placed by an atom of sodium whereby a solution of sodium hydroxide, Na — 0 — H, is
formed ; in the latter case, one atom of calcium replaces one atom of hydrogen in
each of the atoms of water, and thus forms one molecule of calcium hydroxide, Ca(OH) 2.
The reactions with the alkali metals are violent, but they are controlled by amalga-
mating the metals with mercury. An alloy of sodium with lead — called commercially
hydrone — generates hydrogen when in contact with water.
According to H. Fleck and H. Basset,^ amalgamated magnesium decomposes
cold water, while the metal alone has no appreciable action on cold water. H. St.
C. Deville could detect no appreciable action between aluminium and boiling water.
J. B. Bailie and C. Fery showed that if the aluminium be amalgamated, it is rapidly
oxidized by boiling water : 2A1+6H20->2A1(0H)3+3H2. In the commercial
application of this process, it is not necessary to amalgamate the aluminium
directly, since the metal reduces solutions of mercuric salts to the metallic state :
2Al+3HgCl2->2AlCl3+3Hg, and any excess of aluminium present is automatically
amalgamated.
H. Wislicenus and L. Kaufmann prepare the amalgamated aluminium turnings, freed
from oil by treatment with aqueous soda luitil a copious evolution of hydrogen has set in,
by rinsing them once with water, and acting on the metal for one or two minutes with a
half per cent, solution of mercuric chloride ; the process is then repeated, ajid the metal
finally well washed with water, alcohol, and ether, and kept under light petroleum ready
for use. The last traces of water may, after a little while, cause a reaction vigorous enough
to raise the petroleum to its boiling-point. This amalgamated aluminium decomposes
water with violent evolution of hydrogen.
The aluminium should be free from copper. The alloy duralumin contains
94 per cent, of aluminium and 4 per cent, of copper, and it is but slightly attacked by
boiling water even in the presence of a mercury salt. In M. Baupre's process an
intimate mixture of aluminium powder, mercuric chloride, and potassium cyanide,
which is quite stable in dry air, gives a close approximation to the theoretical yield
of hydrogen when gradually added to water at about 70°. S. Uyeno's alloy of
aluminium (78 to 98) with a small proportion of zinc (I'S to 15) and tin (0'5 to 7'0
per cent.) cast into plates and then amalgamated with mercury, decomposes hot
water. If a little potassium permanganate be added to boiling water, in which
aluminium powder is suspended, there is a continuous evolution of hydrogen. If
too muchpermanganate be present, the reaction is retarded. Chlorates, perchlorates,
and nitrates do not act in place of the permanganate as catalytic agents. H.
Fosterling and H. Philipps make a mixture of metallic sodium and aluminium
silicide, Al2Si4, into briquettes, and preserve them in air- and water-tight boxes.
In contact with water, a mixture of sodium silicate and aluminium hydroxide is
formed, Al2Si4+8Na+18H20=2Al(OH)3+4Na2Si03+15H2. The method has been
called the sical ^process.
O. Prelinger 6 found that manganese decomposes cold water slowly, and hot
water rapidly. According to J. A. Wanklyn and L. Carius, reduced iron does not
decompose water at 50°, but it does decompose water at 100°, and E. Ramann
obtained 12 c.c. of hydrogen by boiling water for an hour with 10 grms. of iron,
reduced in hydrogen. It is not (5lear how much of this hydrogen was occluded in
the metal. N. J. B. G. Guibourt found the reaction between iron and water is
280 INOKGANIC AND THEORETICAL CHEMISTRY
accelerated by salts of mercury, copper, and the less easily oxidizable metals. M. Lorin
also found the reaction to be appreciable at 40° in the presence of ammonium salts.
M. Meusnier and A. L. Lavoisier prepared hydrogen in 1784 by passing steam over
red-hot iron ; and the hydrogen used for inflating the first balloon sent up from
French soil, in 1794, was prepared by this method. The metals chromium, nickel,
and cobalt act similarly. The reaction with steam and iron is usually represented
by the equation: 3Fe+4H20=Fe304+4H2. It is not clear, however, what
particular mixture of iron oxides is actually formed, for the observed results would
accord with the formation of ferrous oxide, FeO, on the back reaction, so that for
equilibrium: 3FeO+H20^Fe304+H2. M. Gillard (1850)7 used the iron-steam
process industrially ; and he reduced the spent oxide with carbon monoxide and
hydrogen derived from the action of steam on heated carbon. The process was soon
abandoned as commercially unsatisfactory.
Porous briquettes — made from spent pyrites, obtained as a by-product in the
manufacture of sulphuric acid, and clay — have been used as a source of the iron.
The oxide of iron can be reduced to the metal and used over and over again by heating
the oxidized briquettes with a reducing agent — e.g. water gas, coal gas, or producer gas.
C. Jacoby used a mixture of finely divided iron with twice its weight of a hydroxide
of calcium, barium, or strontium calcined to dull redness. In the modification
called A. Messerschmitt's iron contact method, the red-hot iron in a shaft retort, re-
sembling a gas producer, is sprayed with a jet of steam, and the iron oxide is reduced
by a stream of water gas, producer gas, or coal gas. There are two types of plant,
the so-called single retort and the multiple retort processes. The process is efficient
and cheap when the right kind of iron is employed. Spongy iron acts more effi-
ciently, presumably because it presents a larger surface for a given amount of iron.
J. Jacob heated iron by part of the hydrogen produced in the reaction. A great
many patents have been obtained for modifications of this process. The gas employed
for the reduction must be free from certain impurities which gradually reduce the
activity of the iron, by forming a layer of impurities (siliceous dust, sulphides, etc.)
on the surface of the metal. In practice, the iron lasts from eight to thirty days,
but it may be more or less revivified by periodically heating it in a stream of air.
A. Messerschmitt ^ used natural ores of manganese, or manganese and iron
with the idea of lowering the reaction temperatures ; 0. Dieifenbach and W. Molden-
hauer used alloys of iron with manganese, chromium, tungsten, titanium, aluminium,
and related elements, as well as briquetted mixtures of the oxides ; the Badische
Anilin und Soda Fabrik used iron oxide fused with refractory oxides — zirconia,
magnesia, or silicates ; C. F. Jaubert used briquettes of iron oxide with fireclay,
pumice, or magnesia, together with a small amount of manganese, chromium, copper,
or lead oxide. Using lime purifiers to remove hydrogen sulphide and carbon dioxide
a gas of 99'75 per cent, purity can be obtained by this process. The efficiency of
the process rapidly deteriorates, and the maintenance of the temperature of the
reaction consumes much of the hydrogen.
F. Bergius ^ argues that liquids are more reactive than gases, and that the
liquid state can be maintained from the boiling point to the critical temperature if
a sufficient pressure be used, and he has patented a process for the preparation of
hydrogen by the action of liquid water on iron. It was found that the gas was of
99*95 per cent, purity because liquid water does not attack the dangerous impurities
• — sulphur and carbides — of iron so readily as steam. The speed of the reaction is
accelerated in the presence of sodium chloride, ferrous chloride, and copper — e.g. at
300°, 230 c.c. of hydrogen per hour were obtained with pure water and iron ; 1390
c.c. per hour if ferrous chloride be present ; and 1930 c.c. per hour if in addition some
copper be present. The yield of hydrogen per hour in the last case was nearly
doubled by raising the temperature of the reaction from 300° to 340°. The hydrogen
can be charged directly into cylinders without further compression. A pressure
apparatus of 10 gallons capacity produced 3000 c. ft. of hydrogen per day.
The metalloids boron, silicon, and carbon can deoxidize steam. H. Moissan 1° gays
HYDROGEN 281
amorphous boron begins to act at a red heat and the reaction then proceeds with
incandescence ; E. Vigouroux found amorphous silicon begins to react at a red heat,
and crystalline silicon at a rather higher temperature. Steam is also decomposed
in its passage over red-hot coke ; hydrogen and carbon monoxide, CO, are the main
products of the reaction : C+H20=C04-H2. The mixture of gaseous products
by this reaction is called water gas. There is also a side reaction : C-f-2H20
=C02-f 2H2, and according to 0. Dieffenbach and W. Moldenhauer,ii the latter
reaction occurs if the coke be saturated with 10 per cent, potassium carbonate, and
when the coke is also mixed with five times its weight of lime, the carbon dioxide is
absorbed by the lime if the temperature of the reaction be between 550° and 750°.
F. Bergius also claims that the carbon dioxide reaction alone occurs if the water be
maintained in the liquid state by pressure at 340°, and if one per cent, of thallium
chloride be mixed with the coke. The carbon dioxide is then removed by lime.
The Badische Anilin und Soda Fabrik worked with finely-divided iron as catalyst
between 400° and 500°, and from 4 to 40 atm. pressure ; J. L. Buchanan and E. B.
Maxted used a catalyst of finely-divided iron and copper for the same purpose.
According to J. J. Coquillion water vapour and carbon monoxide are decomposed
in the presence of red-hot platinum, forming carbon dioxide and hydrogen ;
L. Macquenne says that the transformation is complete in 25-30 hrs. in the presence
of platinum sponge. The reaction is of the balanced type : C0+H20^C02+H2.
0. Hahn found the equilibrium condition to be Ph2^co2"^=-Ph„o^coj where the
equilibrium constant K has the value 0*05 at 400° ; 0*1 at 500° ; "0-3 at 600° ; O'G
at 700° ; and 0*9 at 800°. Hence, low temperatures favour the formation of
hydrogen, but, at the same time, lead to slow reactions. Consequently, the successful
production of hydrogen from water gas requires the use of catalytic agents to
accelerate the reaction at as low a temperature as possible. L. Mond and C. Langer
patented the use of nickel at 350°-400°, or of cobalt at 400°-450° as catalysts ;
C, Ellis and B. E. Eldred used iron, nickel, or manganese ; W. Naher and K. Miiller,
rhodium or palladium asbestos ; L. Vignon, iron or platinum ; and H. S. Elsworthy,
iron or nickel.
Another problem is to separate the hydrogen from the carbon monoxide, when
the first reaction predominates. There have been several proposals ; among others:
(1) B. C. Sykes and S. Blamires, J. Pullmann and H. S. Elsworthy, and A. Longsdon
recommend removing the hydrogen by driving the gas under pressure through porous
tubes. The hydrogen and methane diffuse through the walls of the tubes faster than
the carbon monoxide. A. Jouve and G. Gautier (1906) 12 reduced the percentage
amount of carbon monoxide from 45 to 8 per cent, by passing water gas through a
porous partition through which the hydrogen diffuses far more rapidly than the carbon
monoxide. (2) C. von Linde, A. Frank, and N. Caro (1906) subject the water gas to a
preliminary cooling with liquid air whereby the carbon monoxide and dioxide are
liquefied, the hydrogen remains as a gas. When the liquid bv-product is allowed
to gasify, it can be burnt as a fuel gas. (3) A. Frank (1906) passed the dried mixture
over calcium carbide, CaC2, at a temperature exceeding 300°. The carbon monoxide
and dioxide form calcium oxide and carbonate, and free carbon— any nitrogen
present forms calcium cyanamide. In the former case, CaC2+CO->CaO-h3C ;
and in the latter case, CaC2+N2=CaCN2+C. (4) When a mixture of steam and
carbon monoxide is passed over lime at about 500°, the carbon monoxide is oxidized
and an equal volume of hydrogen is evolved : CaO+H20+CO->CaCOo4-H2.
According to M. G. Levi and A. Piva the reaction occurs in two stages:
CaO+H20-f2CO--(H.COO)2Ca— calcium formate ; and (HCOO)2Ca=CaC03-fCO
+H2. Hence, if water gas— say, a mixture of equal volumes of hydrogen and
carbon monoxide— be passed over lime under these conditions, hydrogen will be
substituted in place of the carbon monoxide. (5) E. K. Rideal and H. S. Taylor 13
have shown that even in the presence of large quantities of hydrogen, carbon
monoxide can be oxidized to the dioxide in the presence of suitable catalysts— a
mixture of iron and chromium oxides to which small quantities of ceria and thoria
282 INORGANIC AND THEORETICAL CHEMISTRY
have been added. This catalyst completely oxidizes carbon monoxide between
200° and 300°, but does not attack hydrogen. The problem of removing a relatively
large proportion of carbon dioxide by lime is a serious one. The basal patent is by
C. M. T. du Motay (1880), and the Greisheim Elektron Co. added 5 per cent, of iron
powder to act as a catalyst. This action was studied by W. E. Engels.
In the so-called auto-comhustion processes of G. F. Jaubert,i* a combustible
substance and an oxidizing agent are kindled in the presence of water. The
combustion once started continues and the water is decomposed. In one case, a
mixture of lime and f errosilicon at a high temperature is exposed to steam, and a little
gunpowder is used to start the reaction : 3FeSi6+4:0H2O=Fe3O4-l-18SiO2+40H2.
In another, a mixture of iron filings, potassium perchlorate, and hydrated lime, or
a mixture of ferrosilicon, litharge, and soda lime, is used. Here the reducing agent
decomposes the water in the hydrated body.
J. E. G. Lahousse (1905) ^^ recommended the decomposition of steam by passing
it over red-hot barium sulphide, BaS, which is thereby oxidized to barium sulphate,
BaSO^, thus, in symbols, BaS+4:H20=BaS04+4:H2. The sulphate can be reduced
back to the sulphide by heating it with coal or producer gas : BaS04+2C
=2C02H-BaS. In the analogous process by G. Teissier and P. Chaillaux, barium
sulphate is heated with manganous oxide : BaS04-l-4MnO->BaS-|-4Mn02 ; and the
product raised to a white heat : BaS+4Mn02->BaS+4MnO+202. When the
reactions complete, steam under pressure is passed over the mixture and hydrogen
is set free : BaS+MnO-h4H20~>BaSq4+MnO+4H2. The solid products of the
reactions are then ready to be employed in a fresh cycle.
(3) The action of metals on dilute acids. — The usual laboratory method of pre-
paration is to act upon zinc with dilute sulphuric acid (1 : 8) or hydrochloric acid
(1:2) as already described. Aluminium, magmesium, or iron may be substituted
for the zinc. The reaction is symbolized : M"+H2S04=MS04+H2, or M"-f 2HC1
=MCl2+H2, where M" represents a gram-atom of the bivalent metals. The
process of chemical change results in the substitution of two atoms of hydrogen
in the acid by an equivalent bivalent atom. According to E. W. Morley, the
purest redistilled zinc contains a little absorbed carbon monoxide which ultimately
finds it way into the hydrogen gas prepared from the zinc. The impurities in the
commercial metals — carbon, sulphur, silicon, phosphorus, antimony, and arsenic —
form the corresponding gaseous hydrides. Sulphuric acid may contain : (i) sulphur
dioxide which is evolved as such or partly reduced to hydrogen sulphide ; (ii)
nitrogen compounds which form nitrogen and nitrous oxides ; (iii) arsenic and
selenium, which form the corresponding hydrides. Hydrochloric acid made from
sulphuric acid may also contain these same impurities. If the zinc and acid are
pure, the resulting hydrogen has a high degree of purity, but it is evolved so slowly
as to make the process of little practical use.i^ N. A. E. Millon noticed that the
speed of evolution is much accelerated if a little platinic chloride be added to the
hydrochloric acid. The platinic chloride is decomposed, and platinum metal is
deposited on the zinc so as to form a kind of voltaic couple. A similar result was
obtained by J. C. d' Almeida with platinum wire or finely divided platinum in contact
with zinc, and C. Gourdon found that in these cases the hydrogen is given off quickly
at the surface of the platinum — not the zinc. A similar effect is obtained with a
small quantity of a salt of copper, silver, gold, tin, antimony, bismuth, nickel,
cobalt, or one of the less oxidizable metals. The reaction between zinc and sulphuric
acid was reported by F. Selmi to be accelerated by the presence of sulphates of
manganese, magnesium, or iron, but to be retarded by sulphates of potassium,
sodium, or aluminium.
Iron is little used in the preparation of hydrogen for laboratory purposes since
the gas is contaminated with hydrocarbon gases derived from the carbides in the
iron. These impurities give the hydrogen an unpleasant odour. In stationary
military camps, hydrogen is sometimes made in this manner, but the main objection
is the enormous quantity of metal and acid needed for filling an airship or balloon
HYDKOGEN 283
of but moderate size — say, 250,000 or 500,000 cubic feet capacity. In the former
case, about 18 tons of iron and 31 tons of acid are needed for an inflation. The
captive balloon at the 1878 Paris Exhibition had a capacity of about 883,000 c. ft.
and it is said to have required 190 tons of sulphuric acid and 80 tons of iron for an
inflation. This method of making hydrogen for inflating balloons is reported to have
been used in 1861 in the American Civil War. Unless specially treated, the impurities
in the hydrogen produced by this process are liable to rot the fabric of balloons and
airships. L. V. Pratis and P. Marengo patented a process for the purification of
the gas by first scrubbing it with water ; and then passing it through a solution of
a lead salt. Several other methods of cleaning the gas have been patented.
F. -Konther patented a process for recovering the acid in the preparation of
hydrogen by the action of hydrochloric acid on iron ; he heated the ferrous chloride
to a high temperature with steam : 3FeCl2+4H20->Fe304+6HCl+H2.
G. V. Barton patented the recovery of the zinc sulphate produced as a by-product
when zinc is used in place of iron, by treating the solution with sodium carbonate
or hydrocarbonate, and igniting the precipitated zinc carbonate to form zinc oxide
— the zinc white used as a pigment by paint-makers. J. WannschafE and J. Savels-
berg 17 proposed to make hydrogen and zinc oxide by the action of zinc chloride
solutions on waste zinc.
(4) The decomposition of metal hydrides. — ^The hydrides of the alkali and alkaline
earth metals readily decompose water at ordinary temperatures liberating hydrogen
and forming the corresponding metal hydroxide : CaH2+2H20->Ca(OH)2+2H2.
The commercial powder hydrolith — devised by G. F. Jaubert is — contains 90 per
cent, of calcium hydride, and it gives of! hydrogen by mere contact with water. A
kilogram of the solid gives about a cubic metre of hydrogen. It has been used by
the French army for filling observation balloons. The process is rather expensive
though ver}^ convenient. In the commercial process for making calciimi hydride,
CaH2, a little nitride, Ca3N2, is formed at the same time, and this reacts with water
producing ammonia : Ca3N2+6H20=3Ca(OH)2+2NH3. The ammonia, being
very soluble in water, is readily removed by scrubbing the gas with water. The
lightness — low specific gravity — of lithium hydride, Li2H, or Li4H2, has led to its
being suggested as a means of restoring hydrogen to an airship should there be
a serious loss of this gas from any cause. 27'76 lbs. of lithium hydride and 66*6 lbs.
of water give 1000 c. ft. of hydrogen at about 5° and 760 mm. This amount of
hydrogen would have a buoyancy of 74'06 lbs., and if the products of the reaction
were dropped overboard, every 94*36 lbs. of materials would increase the buoyancy
of the airship 168-42 lbs. The cost of lithium hydride is, however, too great for the
general realization of this proposal.
(5) The action of^netals and alloys on alkaline lye—'Uydiogen gas can be obtained
by warming tin, aluminium, or zinc with a dilute solution of sodium hydroxide (50
grms. of the hvdroxide per 500 c.c. of water). If the metals are free from carbon, the
resulting gas is fairly pure. J. P. Cooke and T. W. Richards (1888) i» used this
process in some atomic weight determinations and purified the gas by passing it
over solid potassium hydroxide, calcium chloride, concentrated sulphuric acid, and
phosphorus pentoxide. When aluminium is used, the method is sometimes called
the hydrik or aluminal process. We are told that the Russians prepared hydrogen
for their war balloons in Manchuria during the Russian-Japanese war, 1904-5, in
this way. The reactions with aluminium and zinc are respectively represented :
(NaOH /ONa NaOH „ ^ONa^„
2Ar'+ NaOH=2A1^0Naf 3H2 ; and Zn"-f{j^^Qjj=Zn<Qj^g^+H2
(NaOH \ONa
Here it will be observed that one hydrogen atom in each molecule of sodium
hydroxide is replaced by the respective elements— three by the tervalent alu-
minium, and two by the bivalent zinc; sodium aluminate, Al(0Na)3, is
a by-product in the former process; and sodium zincate, Zn(0Na)2, in
284 INORGANIC AND THEORETICAL CHEMISTRY
the latter case. With tin, sodium stannate, Sn(0Na)4, is formed. We
see, therefore, that under the stated conditions, an atom of sodium or potassium
can displace only one of the two hydrogen atoms in the water molecule — H2O ;
and that tin, zinc, or aluminium can displace the other hydrogen atom.
According toH. Williams (1881),20 zinc dust heated with water alone gives hydrogen
gas. Water alone, however, does not appreciably decompose boiling water ; but,
as shown by W. Wilson, decomposition does occur if copper turnings are also present.
L. Meyer and T. Leykauf observed the evolution of hydrogen during the action of
zinc and water on crystals of copper sulphate ; the action in these cases is similar
to that of the zinc-copper couple. H. Schwarz obtained hydrogen by heating a
mixture of zinc dust and calcium hydroxide, and the process was patented by
W. Majert and G. Richter for the generation of hydrogen in the field or the in-
flation of observation balloons. Hydrogen may be prepared by heating slaked lime
with coal— C. M. Tessie du Motay and C. R. Marechal (1868)— C+2Ca(OH)2
=CaO+CaC03+2H2.
Under the name of the silicol process the preparation of hydrogen by the action
of sodium hydroxide on the element silicon has been recommended by Siemens and
Schuckert (1911). and on the cheaper ferrosilicon — called silicol — or other silicon
alloys by G. F. Jaubert (1908).2i
An intimate mixture of powdered sodium hydroxide and ferrosilicon when
moistened with water gives off hydrogen rapidly, and becomes incandescent. When
ferrosilicon is added to a cold solution of sodium hydroxide, the metal is not rapidly
attacked, but owing to the heat of the reaction, the temperature of the liquid gradu-
ally rises as ferrosilicon is added, and then reaches a stage — about 80°— where the
ferrosiUcon is instantly attacked. Consequently, if ferrosilicon is added faster than
it is attacked by the caustic lye, when the temperature has risen to the critical stage
the accumulated ferrosilicon may be instantly attacked causing an explosive pressure
of gas in the gas generator. The reaction is not a simple oxidation of the water by
the silicon : Si+2H20 — Si02+2H2, followed by the formation of sodium silicate :
Si02+2NaOH=Na2Si03+H2O, so that the composite equation is symbolized :
Si+2NaOH+H20=Na2Si03+2H2, since, instead of 2*86 parts of sodium hydroxide
being required for one part of sihcon, the ratio is less, 1 : 1'7. This shows that a
richer silicate is simultaneously formed, possibly by the action of the simple sodium
sihcate on the silicon : say, Si+Na2Si03+2H20=Na2Si205+2H2, where the com-
posite reaction is symbolized : 2Si4-2NaOH4-3H20— Na2Si205+4H2. This requires
a ratio of silicon to sodium hydroxide of 1 : 1"43. The ratio 1 : 1*72 is the one usually
recommended. With a 10 per cent, sodium hydroxide and 92 per cent, silicol solution
P. L. Teed found at 15° and 30 in. barometric pressure :
Weight ratio, Si : NaOH . . 1 : 0-745 1 : 1-065 1 : 1-480 1 : 3-200
Hydrogen per lb. of silicol . 13-62 14-30 1536 16'80 c. ft.
With 30 per cent, sodium hydroxide solution with the ratio by weight Si : NaOH
=1:0-852, 1:2-13 and 1:3*19, the respective yields of gas were 19*35, 23*90,
and 23*58 c. ft. ; while with 40 per cent, sodium hydroxide, and a weight ratio of
Si : NaOH of 1 : 1*58 and 1 : 3*19, the respective yields of hydrogen were 24*10
and 24*50 c. ft. Hence, a 40 per cent, solution of sodium hydroxide with a ratio
by weight of Si : NaOH of 1 : 1*6 would appear to be best, but if the solution be
too concentrated, it becomes so viscous or even solid during the action that the
plant is difficult to clean ; if the solution be too dilute, tlie frothing is troublesome.
A little mineral grease is sometimes used with the soda lye to prevent frothing.
The usual solution employed contains about 25 per cent, of sodium hydroxide, and
the ratio of Si : NaOH by weight nearly 1 : 1*72. With the 25 per cent, caustic lye
and a wide generator there is no difficulty with frothing when no grease is used.
Plants are in use generating 1500 to 6000 c. ft. per hour by the siUcol process.
It furnishes 2500 c. ft. of gas per hour. The sodium hydroxide is dissolved in water
in a tank, and run into a generator fitted with a stirrer run by a benzene motor.
HYDROGEN 285
The silicol in the container is fed into the generator by an automatic feed also
operated by the benzene motor. The hydrogen then passes into the condenser or
scrubber, where it is cooled and freed from steam. The heat generated during the
action reaction raises the temperature in the generator above 100°, sometimes to
120°. Thermometers inserted in the generator enable the temperature to be observed
and if necessary controlled. An excessive pressure due to the rapid generation of
hydrogen is prevented by suitable seals. Precautions are of course taken to prevent
the. generation of an explosive mixture of hydrogen and air — usually before the caustic
lye is run into the generator — by allowing hydrogen from the gas-holder to blow
back hydrogen equivalent to four times the volume of the plant. When the genera-
tion of gas from a given charge has ceased, the sodium silicate solution and sludge
are run from the generator by a trapped discharge pipe ; the generator is washed
out with water and is ready for another charge. Low-grade ferrosilicon is not
economical because it does not give equivalent yields of hydrogen. At least 80
per cent, ferrosilicon is needed to give satisfactory yields :
Snicol . . . . 50 70 85 90 per cent. Si
Yield . . . .12^ 27J 91 97 per cent, of theory
An 84 per cent, ferrosilicon has iron 69 per cent., aluminium 5"3 per cent., and carbon
0'2 per cent. The sodium carbonate impurity in the caustic lye and the iron of the
ferrosilicon are found unaltered in the sludge. The sludge contains 37 per cent,
of silica ; 6 per cent, sodium carbonate ; 20 per cent, of soda calculated as Na20 ;
about 10 per cent, of matter insoluble in water and the remainder is water. The
ferrosilicon gives a 90 per cent, yield of hydrogen, the loss being mainly due to the
protective action of impurities, leakages, and the solubiHty of hydrogen in the water
used in the scrubber. The gas has from 99 to 99"9 per cent, purity and is virtually
free from arsine and phosphine, for less than 0"01 per cent, of these gases are present ;
the acetylene amounts to, about 0*005 per cent. The main impurity is air. This
process gives a large hydrogen production from an inexpensive plant ; but the cost
of the reagents is high. The process is therefore advantageous where large quantities
of hydrogen are periodically required, but not where there is a continuous demand.
Slaked lime substituted in place of sodium hydroxide gives very poor yields of
hydrogen, but a mixture of the two gives better results than with sodium hydroxide
alone. The sodium silicate formed in the first stage of the reaction between the
silicon and sodium hydroxide is probably converted into the hydroxide again :
Na2Si03+Ca(OH)2=2NaOH+CaSi03. A mixture, devised by G. F. Jaubert
(1910), containing 25 parts of 90-95 per cent, ferrosilicon or manganosihcon, 60 of
sodium hydroxide, and 20 slaked lime, is commercially known as hydrogenite. It
is sold compressed in cakes contained in metal cartridges to protect the mixture from
moisture, and when ignited — by forcing a hot ball or wire into the briquette — forms
sodium and calcium silicates with the evolution of hydrogen. A kilogram of the
powder gives from 270 to 370 litres of hydrogen.
(6) The decomposition of hydrocarbons. — When hydrocarbon gases are heated to
about 1000°, they dissociate forming lamp-black, hydrocarbons, and hydrogen.
In 1911, R. Pictet 22 patented a process for heating a mixture of water and hydro-
carbon vapour — e.g. petroleum — whereby a mixture of hydrogen with some carbon
monoxide is formed. The proportion of the latter depends on the proportion of
water present in the original mixture. The Carbonium Co. at Friedrichshafen, the
headquarters oiE the Zeppelin airships, has obtained hydrogen by decomposing
acetylene, C2H2, prepared from water and calcium carbide, by electric sparks or by
explosion ; preferably, according to J. Machtoff, under a pressure of 4 to 6 atmo-
spheres, C2H2=2C+H2. The carbon wins a high price as a high-grade lamp-black.
In F. G. L. Rincker and L. Wolter's process the gas produced by the distillation
of crude tar, oil, or other hydrocarbon is passed through hot coke or charcoal.
The 6 to 8 per cent, of carbon monoxide and hydrogen sulphide in the washed gas
is partly removed by hot soda lime. The Badische Anilin und Soda Fabrik suggest
286 INORGANIC AND THEORETICAL CHEMISTRY
reducing the temperature of decomposition by the use of a refractory oxide like
magnesia impregnated with nickel oxide for reducing the temperature of decom-
position ; 0. DiefEenbach and W. Moldenhauer, the use of wire gauze of nickel,
cobalt, platinum, etc., with short contact and sudden cooling of the products so
as to avoid the reduction of carbon dioxide to the monoxide : and the Berlin
Anhaltische Maschinenbau, and C. Ellis, the use of coke as a contact catalyst.
7. By-product hydrogen. — Hydrogen is a by-product in the manufacture of
chlorine by the electrolysis of aqueous solutions of alkali chlorides ; 2NaCl4-2H20
=2NaOH+H2+Cl2. Chlorine is given off at one electrode and hydrogen at the
other. The sodium chloride is broken down by the electric current into sodium
and chlorine ; and the sodium, in contact with water, reacts as indicated in the
equation : 2Na+2H20=2NaOH-l-H2. Hydrogen is also a by-product in the manu-
facture of oxalates from formates, which in turn are made by treating caustic potash
with carbon monoxide (producer gas) under pressure : KOH+CO=H.COOK ; and
2H.COOK=K2C204+H2. L. Vignon (1911) showed that if carbon monoxide be
passed over powdered calcium hydroxide at 300°, there is no action ; at 400°, a
mixture of hydrogen and hydrocarbons — chiefly methane, CH4, and ethylene,
C2H4 — is evolved.23 The proportion of hydrogen increases as the temperature rises
until at 600° only hydrogen is formed : CO+Ca(OH)2=CaC03+H2. The forma-
tion of the hydrocarbons is probably preceded by the formation of calcium formate
and oxalate as intermediary products of the reaction : 4CO+2Ca(OH)2+CaO
=3CaC03+CH4 ; and 4CO+4Ca(OH)2+CaO=4CaC03+2H2+C2H4. A mixture
of sodium formate and calcium oxide yields nearly pure hydrogen. Steam, carbon
monoxide, and calcium hydroxide form calcium formate at 250°-300° ; at 500°-600°,
there is almost a quantitative yield of hydrogen ; the presence of sodium hydroxide
lowers the temperature of the reaction. In 1861, M. Berthelot noticed that when
either the formates or the oxalates of the alkalies are heated with the alkali hydroxide;
hydrogen is formed ; and R. Pictet prepared hydrogen by heating sodium formate,
H.COONa, with soda lime. The reaction is symbolized H.COONa+NaOH
->Na2C03+H2 ; similarly, with sodium oxalate, (C00Na)2, or potassium oxalate :
Na2C204+2NaOH=2Na2C03+H2. The reactions are not particularly convenient
when compared with other methods of preparation, but E. H. Amagat used the
process in his work on the effect of pressure on the volume of gases.
There are several other reactions in which hydrogen is formed. These reactions
are of greater theoretical than practical interest. For example, A. Bruno 24 found
that when iron filings are shaken with water saturated with carbon dioxide for
36-48 hrs., the carbon dioxide is replaced by hydrogen : Fe+C02+H20=FeC034-H2.
According to M. Lorin, most of the ammonium salts and the salts of the amine bases
develop hydrogen when treated with zinc at ordinary temperatures, or better at
40° ; iron acts more slowly. Ammonium nitrate is not appHcable since it gives
nitrous oxide at 50° ; similar remarks probably apply to the nitrates of the amine
bases. W. P. Winter regards the alleged evolution of hydrogen by the action
of water on sodamide as probably due to the presence of some free sodium.
A. Descamps found that potassium cobaltocyanide, K4CoCy6, decomposes water at
ordinary temperatures giving off hydrogen and forming the cobalticyanide, K3CoCyg.
M. Berthelot found that acidified solutions of chromous salts gradually decompose
with the evolution of hydrogen ; and R. Peters noted that the reaction is accelerated
by the presence of platinum black. Hydrogen is given off by uranyl acetate in
sunlight. A. Sieverts found that in aqueous solutions platinum, silver, and copper
cause hypophosphorous and phosphorous acids to split off hydrogen, respectively
forming phosphorous and phosphoric acids : H3P02+H20=H3P03+H2 ; and
H3P03+H20=H3P04+H2. Hydrogen is often evolved when organic compounds
— e.g. oleic acid — are fused with alkalies. Hot soda-lime converts alcohol into
acetic acid with the evolution of hydrogen : C2H50H+NaOH=CH3COONa-|-2H2 ;
formaldehyde, H.COH, gives off hydrogen in the presence of cuprous oxide and
alkali : H.COH -|-NaOH==H.COONa+H2 ; and also with sodium peroxide :
HYDROGEN 287
2HCOH+Na202=2HCOONa+H2. M. Kernbaum found water can be decomposed
by sunlight in accord with the equation : 2H20=H2024-H2, and by exposure to
ultra-violet light ; and A. Dobierne observed a similar reaction with radium rays.
W. Loeb detected hydrogen among the products obtained when the electric arc is
formed under water.
According to G. PoUacci, W. Palladin, and J. Stoklasa and W. Zdobnicky, the
gases obtained as exhalations from the green parts of plants contain some hydrogen,25
and F. Selmi found that hydrogen is liberated by fungi and moulds — arsenic, anti-
mony, and sulphur are converted by the same organisms into the corresponding
hydrides. S. Kostytscheff found the gas formed in the respiratory gases of some seed
plants. H. Tappeiner found that the gas is formed in the fermentation of cellulose,
and in the decay of certain organic bodies. C. Oppenheimer and A. Krogh detected
hydrogen among the intestinal gases of many animals.
The purification of hydrogen gas.— Most of the methods of preparing hydrogen
furnish a gas contaminated with impurities which can generally be removed by
treating the gas with suitable reagents. The nature of the impurities depends upon
the character of the materials employed in the preparation of the gas. Air is one
of the commonest of impurities, and it comes from the air dissolved in the liquids
used in the preparation of the gas ; from the air originally present in the apparatus ;
and by leakage through rubber and other joints. This impurity can be considerably
reduced by using liquids previously boiled ; and mercury sealed glass joints for the
apparatus. Oxygen may also be derived from the potassium permanganate solution
sometimes used for washing the gas. According to H. Debray,^^ traces of oxygen
can be removed by passing the hydrogen gas over red-hot copper, over spongy
platinum, melted sodium, or through a solution of chromous chloride, but the
nitrogen from the air is not removed by this treatment. According to E. W.
Morley,27 from O'Ol to 0"001 per cent, of nitrogen in the gas can be recognized
sp ectros copically .
The hydrocarbon gases in hydrogen are derived from carbonaceous impurities
in the metals and acid. E. W. Morley states that the electrolysis of dilute alkali
hydroxides, or dilute hydrochloric acid, may furnish a gas contaminated with traces
of carbon compounds. When the alkali contains carbonates, or the acid organic
matter, M. Berthelot says that hydrogen prepared from the metals always contains
carbon compounds. The electrolysis of water acidified with purified sulphuric acid
is recommended by E. W. Morley 28 as the safest process to use for hydrogen of a
high degree of purity, and M. Berthelot recommends washing the electrolytic
hydrogen with potassium permanganate solution and scrubbing it with fused
potassium hydroxide.
The metal used in preparing the gas may contain occluded carbon oxides, or
sulphur, selenium, phosphorus, silicon, antimony, and arsenic, which contaminate
the hydrogen with gaseous compounds. Sulphuric acid may also be reduced by
hydrogen to sulphur dioxide or even to hydrogen sulphide.29 The acids may
also contain nitrogen, selenium, and arsenic compounds. Hydrocarbons can be
removed by passing the gas through alcohol, or, according to J. W. Dobereiner
and J. Stenhouse,30 by passing the gas through a tube packed with recently
ignited wood charcoal, or paraffin wax. M. Donovan recommends passing the gas
through an oxidizing liquid — say fuming nitric acid followed by a solution of ferrous
sulphate to absorb the nitrous fumes ; E. Varenne and E. Hebre used a sulphuric
acid solution of potassium dichromate followed by potassium hydroxide ; and
E. Schobig used potassium permanganate in a similar way. The sulphur, selenimii,
and sihcon compounds are absorbed by passing the gas through a solution of potas-
sium hydroxide ; and the hydrides of arsenic, antimony, and phosphorus, accord-
ing to A. Lionet,3i are decomposed by passing the gas over red-hot copper turnmgs,
but they are usually removed by passing the gas through solutions of the salts of
metals— e.^. mercuric chloride, silver sulphate, lead nitrate— as indicated by J. B. A.
Dumas in 1843 ; or else by passing the gas through a saturated solution of potassmm
Calcium
Zinc
Zinc
Calcium
Sulphuric
chloride.
bromide.
chloride.
bromide.
acid (1-838).
Moisture
. 0-0021
0-0011
0-0008
0-0002
0-0000025
288 INORGANIC AND THEORETICAL CHEMISTRY
permanganate, and then through a 5-10 per cent, solution of silver nitrate. Accord-
ing to H. Reckleben and G. Lockemann (1908), bromine, bleaching powder, and
potassium hypochlorite are good absorbents for arsine ; the same gas is said to be
removed by bubbUng hydrogen through petroleum spirit cooled by liquid air at
—110°, and H. K. Onnes recommends freezing out the various impurities from
hydrogen by cooHng it with liquid air. Carbon monoxide can be removed by passing
the gas over heated soda lime : 2NaOH+CO->Na2C03+H2.
Hydrogen has been purified in special cases by passing it into a glass tube
containing metallic palladium previously evacuated, and heated red-hot. The
metal is allowed to cool in contact with the gas. The hydrogen is assumed to be
alone absorbed by the metal ; and to be given off again when the tube containing
the metal is heated. The tube containing the palladium is of course attached to
the apparatus into which the hydrogen is to be introduced.
Drying gases. — In 1766, H. Cavendish dried gases by passing them through a
tube containing a hygroscopic salt. Anhydrous calcium chloride and concentrated
sulphuric acid are most commonly used as drying agents, and in special cases phos-
phorus pentoxide is employed. Several attempts have been made by R. Fresenius
(1865),32 E. W. Morley (1885), and G. P. Baxter and R. D. Warren (1911), to estimate
the amount of moisture left in a gas after the desiccating agents have done their
work. From these investigations it appears that the weight of residual water
(grams) left in a litre of gas dried by different desiccating agents is as follows :
Phosphorus
pentoxide.
0-000000025 grm.
The phosphorus pentoxide should be freed from the lower oxides of phosphorus
by distillation over platinized asbestos in a current of oxygen. Alumina dehydrated
at a low temperature is said to be a little superior as a drying agent to sulphuric
acid, and has the additional advantage that it can be revivified by reheating
after it has done its work.
Storage. — The gas is compressed in steel cylinders called hombs — under a pressure
of about 100-150 atmospheres. There has been a number of fatal accidents from
the explosion of the cylinders. In most cases it is certain that an explosive mixture
of hydrogen and oxygen has been introduced into the cylinder, or an oxygen
cylinder still containing oxygen has been charged with hydrogen. The cylinders
of hydrogen or the taps of hydrogen cylinders are often coloured red to lessen the
risk of accidentally using a cylinder of hydrogen for one of oxygen. We are told
that in the Boer War, the British transported hydrogen compressed in cylinders
at about 200 atm. pressure, and that it required 50 horses to transport sufficient
for a balloon 400 cubic metres capacity.
The cost of hydrogen per cubic metre prepared by the different processes, is,
according to E. D. Ardery (1916) 33 : Steam on iron, Vbd. ; distillation of crude
oils and tar, I'lbd. ; ferrosilicon, lOd. ; water gas, 2-5^. ; iron and sulphuric acid,
12'bd. ; hydrogenite, 16d. ; hydrolith, 44c^. to 50c^. ; silicon and caustic soda, b3d.
References.
1 H. L. Barnitz, Met. Chem. Eng., 14. 391, 1910 ; E. D. Ardery, ib., 14. 260, 333, 1916 ; A. W.
Crossley, Pharm. J (mm., 92. 604, 637, 676, 1914 ; H. S. Redgrove, Chem. Trades Journ., 60.
359, 1917.
2 0. Schonherr, Zeii. Elektrochem., 1. 417, 468, 1895 ; 2. 162, 245, 1895 ; J. W. Richards,
Journ. Franklin Inst., 160. 377, 1905 ; J. W. Richards and W. S. Landis, Trans. Amer. Elec-
irochem. Soc., 3. 105, 1903 ; 4, 111, 1903.
3 H. B. Baker, Journ. Chem. Sac., 81. 400, 1902 ; E. W. Morley, Amer. Chem. Journ., 12.
460, 1890.
* P. Garuti and C. R. Pompili, Brit. Pat. No., 23663, 1896 ; 12960, 1900 ; 2820, 1902 ; 27249»
1903 • P. Garuti, ib., 16588, 1892 ; G. B. Baldo, ib., 18406, 1895 ; W. B,. Knowles Oxygen Co.,
ib 1812, 1913; A. Delmard, German Pat. D.R.P., 58282, 1890; 0. Schmidt, ib., 111131,
1899 - Schuckert & Co., ib., 231545, 1910 ; M. U. Schoop, ib., 141049, 1900 ; E. Westphal, ib..
HYDKOGEN 289
133615, 1900 ; M. Hazard-Flamand, U.S. Pat. No., 646281, 1900 ; 1003466, 1911 ; J. H. Fischer,
E. G. Liining, and A. W. Collins, ib., 1004249, 1911 ; R. Moritz, ib., 981102, 1911 ; J. B. Burdett,
ib., 1086804,1914; C. Ellis, i6., 1087937, 1092903, 1914 ; I. H. I^vin, ib., 1094728,1914;
E. Leroy and R. Moritz, French Pat. No., 397319, 1908; L'Oxyhydrique fran9ai8, i6., 459957, 1912 ;
E. Benker, ib., 461981, 1913; Maschinenfabrik Siirth, ib., 462394, 1913; V. Engelhardt, Die
Elektrolyse des Wassers, Halle a. S., 1902 ; Easton, Pa., 1904 ; F. Brahmer, Chemie der Oase,
Frankfurt a. M., 1911 ; P. L. Teed, The Chemistry and Manufacture of Hydrogen, London, 1919 ;
R. Hammerschmidt and J. Hess, Chem. Ztg., 22. 123, 1898 ; G. Winssinger, ib., 22. 609, 1898 ;
0. Schmidt, Zeit. Elektrochem., 7. 295, 1901 ; A. Coehn and W. Caspari, ib., 6. 37, 1900 ;
E. W. Magruder, Amer. Chem. Journ., 19. 810, 1897.
5 H. St. C. Deville, Ann. Chim. Phys., 5. (3), 46. 415, 1856 ; M. C. Schuyten, Chem. Ztg., 20.
129, 1896; S. Uyeno, Brit. Pat. No., 11838, 1912; M. Baupr6, Crnipt. Rend., 147, 310, 1908;
French Pat. No., 392725, 1908; Chemische Fabrik Greisheim Elektron, Brit. Pat. No., 3188,
1908 ; H. Forsterling and H. Philipps, U.S. Pat. No., 977442, 1910 ; H. Wislicenus and L. Kauf-
mann, Ber., 28. 1323, 1983, 1895 ; J. B. Bailie and C. Fery, Ann. Chim. Phys., (6), 17. 248, 1889 ;
H. Fleck and H. Basset, Journ. Amer. Chem. Soc, 17. 789, 1895.
8 J. A. Wanklyn and L. Carius, Liebig's Ann., 120. 69, 1861 ; N. J. B. G. Guibourt, Ann.
Chim. Phys., (2), 11. 43, 1819 ; M. Lorin, Compt. Rend., 60. 745, 1865 ; M. Meusnier and A. L.
Lavoisier, Mem. Acad., 269, 1784 ; E. Ramann, Ber., 14. 1453, 1881 ; O. Prelinger, Monatsh., 14.
353 1893
^ M. Gillard, Journ. Pharm., (3), 17. 105, 1850; H. Giffard, Monit. Scient., (3), 3. 156,
1873.
8 A. Messerschmitt, French Pat. No., 461480, 1913 ; Badische Anilin und Soda Fabrik, ib.,
440780, 1912 ; C. F. Jaubert, ib., 418312, 1909 ; 0. Dieffenbach und W. Moldenhauer, German
Pat. D.R.P., 233347, 1910.
9 H. Lane, ^n^Pa^.ATo., 10. 356, 1903; 17591,1909; 11878,1910; H. Lane and G. Monteux.
French Pat. iVo.. 386991, 1908; A. Messerschmitt, Brit. Pat. No., 12117, 12242, 12243, 1912 ; 17691,
17692, 18028, 18942, 1913 ; J.Jacob, ib., 593, 1861 ; Badische Anilin und Soda Fabrik, ib., 27735,
1912 ; 2096, 1913 ; J. Oettli, ib., 16759. 1885 ; J. Betou, ib., 7518, 1887 ; V. B. Lewes, ib.,,
20752, 1890 ; H. G. Hills and H. Lane, ib., 10356, 1903 ; W. Naher and M. Noding, German Pat.
D.R.P., 279726, 1913 ; N. Caro, ib., 249269, 1910 ; H. Strache, ib., 253705, 1910 ; C. Jacoby,
Patentblatt, 10. 273, 1889 ; L. Vignon, French Pat. No., 373271, 1907 ; C. Dellwik and E. Fleischer,
ib., 395132, 1908 ; J. Pintsch, ib., 466739, 1913 ; H. E. Elsworthy, U.S. Pat. No., 778182,
1904 ; V. B. Lewes, Journ. Soc. Chem. Ind., 10. 824, 1891 ; 12. 437, 1893 ; W. Lettermann,
Journ. Gas-Beleuch, 39. 187, 1896; H. L. Barnitz, Met. Chem. Eng., 14. 391, 1916; 16.
611, 1917.
10 H. Moissan, Compt. Rend., 114. 617, 1892 ; E. Vigouroux, Bull. Soc. Chim., (3), 13. 616,
1895.
11 0. Dieffenbach and W. Moldenhauer, Brit. Pat. No., 8734, 1910 ; F. Bergius, German Pat.
D.R.P., 259030, 254593, 1911 ; 262831, 1912; Journ. Soc. Chem. lnd.,'S,2. 462, 1913 ; Brit. Pat.
No., 19002, 19003, 1912 -, Badische AniUn und Soda Fabrik, ib., 26770, 27117, 27055, 1912 ; 8864,
27963, 1913 ; 16490, 1914 ; J. Pullman and H. S. Elsworthy, ib., 22340, 1891 ; Chemische Fabrik
Greisheim Elektron, ib., 2523, 1909 ; C. M. Tessie du Motay, U.S. Pat. No., 229338, 229339,
229340, 1880 ; C. Ellis and B. E. Eldred, ib., 854157, 1907 ; H. E. Elsworthy, French Pat. No.,
355324, 1905 ; W. Naher and K. Miiller, German Pat. D.R.P., 237283, 1910 ; J. L. Buchanan
and E. B. Maxted, ib., 6476, 6477. 1914 ; J. J. Coquillion, Campt. Rend., 88. 1204, 1879 ;
L. Maquenne, Bull. Soc. Chim., (2), 39. 308, 1883 ; G. F. Jaubert, French Pat. No., 418312, 1909 ;
0. Hahn, Zeit. phys. Chem., 42. 705, 1903 ; 44. 513, 1903.
1'-^ B. C. Sykes and S. Blamires, Brit. Pat. No., 3332, 1891 ; Journ. Soc. Chem. Ind., 10. 353,
1891 ; A. Longsdon, tfe., 11. 671, 1892 ; A. Jouve and G. Gautier, French Pat. No., 372045, 1906 ;
M. G. Levi and A. Piva, Ann. Chim. AppL, 4. 1, 1914 ; Chemische Fabrik Greisheim Elektron,
Brit. Pat. No., 2523, 1909 ; 13049, 1912 ; 0. Diefifenbach and W. Moldenhauer, ib., 8734, 1910 ;
A. Frank, ib., 26928, 1906; E. Ellenberger, U.S. Pat. No., 989955, 1912; C. von Linde, tb.,
1020102, 1020103, 1027862, 1027863, 1912; W. T. Hoofnagle, ib., 1056026, 1913; French Pat.
No., 417983, 1911 ; G. Claude, ib., 375991, 1906 ; H. S. Elworthy, ib., 355324, 1905 ; J. PuUmann
and H. S. Elsworthy, Brit. Pat. No., 22340, 1891 ; L. Vignon, ib., 20685, 1907; W. Naher and
K. Miiller, ib., 20484, 1911 ; L. Mond andC. Langer, ib., 12608, 1888; C. Ellis and B. E. Eldred,
U.S. Pat. No., 854157, 1907.
13 Greisheim Elektron Co., Brit. Pat. No., 2523, 1909 ; W. H. Engels, Ueber die Wasserstoff-
gewinnung aus Kohlenoxyd und Kalkhydrate, und die Beschleunigung der Wassergassreaktton
durch Eisen, Karlsruhe, 1911 ; E. K. Rideal and H. S. Taylor, Analyst, 44. 89, 1919.
1* C. F. Jaubert, Brit. Pat. No., 9623, 1911 ; 5005, 1912.
15 J. E. G. Lahousse, French Pat. No., 361866, 1905; G. Teissier and P. ChaiUaux, \b.,
447688, 1912. „-^ ,:, c , •
16 N. A. E. Millon, Compt. Rend., 21. 37, 1850 ; L. C. A. Barreswill, t6.,21. 292, 18o0 ; F. belmi,
Ber., 13. 206, 1880; J. C. d' Almeida, Compt. Rend., 68. 442, 633, 1869; C. Gourdon, %b., 76,
1250 1873
1' G. Wannschaff and J. Savelsberg, Brit. Pat. No., 5511, 1911 ; H. Williams, ib., 8895, 1886 ;
H. Hawkins, ib., 15379, 1891 ; 25084, 1897; L. V. Pratis and P. Marengo, tb., 15509, 189/ ;
VOL. I. ^
290 INOKGANIC AND THEORETICAL CHEMISTRY
J. Fielding, ib., 17616, 1898; G. V. Barton, ib., 28534, 1910; F. Konther, Oerman Pat
D.R.P., 42456, 1888.
18 G. F. Jaubert, French Pat. No., 327878, 1902 ; Brit. Pat. No., 25215, 1907 ; A. Fowirsiols,
Rev. Gen. Science Pur. Appl, 26. 339, 1915.
" J. P. Cooke and T. W. Richards, Amer. Chem. Journ., 10. 81, 191, 1888 ; J. Thomsen,
Zeit. anorg. Chem., 11. 14, 1896 ; Lord Rayleigh, Chem. News, 59. 147, 1889 ; H. Modebeck,
Chem. Ztg., 29. 64, 1905 ; F. F. Runge, Pogg. Ann., 16. 130, 1829.
20 K WiUiams, Chem. News, 51. 146, 1882 ; 52. 205, 1883 ; H. Schwarz, Ber., 19. 1140, 1886 ;
W. Majert and G. Richter, German Pat. D.R.P., 39898, 1887; 42488, 1888; Brit. Pat. No.,
4881, 1887 ; W. Wilson, Gilberts Ann., 14. 238, 1803 ; L. Meyer, Ber., 9. 512, 1876 ; T. Leykauf,
Journ. prakt. Chem., (1), 19. 124, 1840 ; J. H. Gladstone, Chem. News, 34. 43, 1876 ; C. M. Tessie
du Motay and C. R. Mar^chal, Bull. Soc. Chim., (2), 9. 334, 1868.
" G. F. Jaubert, French Pat. No., 406930, 1910 ; Brit. Pat. No., 7494, 1913 ; ib., 17589, 1911 ;
G. F. Jaubert, ib., 153, 1911 ; H. S. Redgrove, Chem. Trades Journ., 60. 359, 1917; P. L. Teed,
The Chemistry and Manufactiire of Hydrogen, London, 1919 ; E. A. Weaver, Journ. Ind. Eng.
Chem., 12. 232, 1920.
22 R. Pictet, Brit. Pat No., 24256, 1910 ; 13397, 14703, 1911 ; Maschinenbau A.G., ib., 2054,
1914 ; R. Lessing, ib., 15071, 1909 ; Badische ^nilin und Soda Fabrik, ib., 12978, 1913 ; F. G. L.
Rincker and L. Wolter, Wasserstojf nach dem Rincker-Wolter-Verfahren, Berlin, 1909; French
Pat. No., 391867, 391868, 1908 ; E. Geisenberger, ib., 361462, 1905 ; C. Ellis, U.S. Pat. No.,
1092903, 1914; C. Bosch, German Pat. D.R.P., 268291, 1911 ; J. Machtoff, ib., 194301, 1906;
0. Dieffenbaeh and W. Moldenhauer, ib., 229406, 1909.
23 L. Vignon, Bull. Soc. Chim., (4), 9. 18, 1911 ; V. Merz and W. Weith, Ber., 13. 718, 1880 ;
G. Levi and A. Piva, Ann. Chim. Appl, 5. 271, 1916 ; M. Berthelot, Ann. Chim. Phys., (3), 61.
463, 1861 ; R Pictet, ib., (5), 13. 216, 1878.
2* A. Bruno, Bull. Soc. Chim., (4), 1. 661, 1907 ; M. Lorin, Compt. Rend., 60. 745, 1865 ; A.
Descamps, Ann. Chim. Phys., (4), 67. 330, 1868; M. Berthelot, ib., (4), 127. 24, 1898; W. Loeb,
Ber., 34. 917, 1901 ; M. Kembaum, Bull. Acad. Cracow, 583, 1911 ; A. Dobierne, Compt. Rend., 148.
703, 1909 ; Radium, 6. 45, 1909 ; W. P. Winter, Journ. Amer. Chem. Soc., 26. 1484, 1904 ;
A. Sieverts, Zeit. anorg, Chem., 64. 59, 1909; 76. 1, 1912; Zeit. phys. Chem., 91. 199, 1916;
R. Peters, ib., 26. 193, 1898 ; Pharm. Centrh., 59. 695, 1898.
26 G. PoUacci, Atti Bot. Univer. Pavia, 7. 97, 1902 ; H. Tappeiner, Ber., 15. 101, 1882 ; 16.
1734, 1740, 1883 ; G. van der Velde, Zeit. physiol. Chem., 8. 367, 1863 ; J. Stoklasa and
W.Zdobnicky, C^em. Ztg., 34. 945, 1910; W. Palladin, Ber.deut. bot. Ges.,29. 472, 1911; S. Kostyt-
scheff, ib., 24. 436, 1906; 25. 178, 1907; C. Oppenheimer, Biochem. Zeit., 16. 45, 1909;
A. Krogh, ib., 7. 24, 1908.
2« J. P. Cooke and T. W. Richards, Amer. Chem. Journ., 10. 81, 191, 1888 ; J. W. Dobereiner,
Schweigger's Journ., 42. 62, 1824 ; M. Siewert, Zeit. Ges. Naturwiss., 23. 1, 1864 ; H. Debray,
Dingier' 8 Journ., 166. 344, 1862.
27 E. W. Morley, Amer. Chem. Journ., 12. 460, 1890; Amer. Journ. Science, (3), 41.
220, 1891 ; G. Bischof, Kastner's Archiv., 1. 179, 1824 ; A. P. Dubrunfaut, Cmipt. Rend.,
69. 1245, 1869.
28 E. W. Morley, Amer. Chem. Journ., 10. 21, 1888 ; 12. 460, 1890 ; Amer. Journ. Science, (3),
41. 220, 276, 1891 ; Zeit. phys. Chem., 20. 242, 1891 ; M. Berthelot, Bull. Soc. Chim., (3), 5.
576, 1891.
29 M. J. Fordos and A. Gelis, Journ. Pharm. Chim., 27. 730, 1841 ; H. Kolbe, Liebig's Ann.,
119. 174, 1861.
*° J. Stenhouse, Liebig's Ann., 106. 125, 1868 ; E. Varenne and E. Hebre, Bull. Soc. Chim.,
(2), 28. 523, 1877 ; J. Habermann, Chem. Ztg., 13. 314, 1894 ; M. Donovan, Liebig's Ann., 21.
375, 1837 ; E. Schobig, Journ. prakt. Chem., (2), 14. 289, 1876 ; M. Berthelot, Bull. Soc. Chim.,
(3), 5. 576, 1891 ; C. Aschmann, Chem. Ztg., 21. 1049, 1898 ; C. Violette, Compt. Rend., 77. 940,
1873 ; J. W. Dobereiner, Schweigger's Journ., 3. 377, 1811.
31 J. B. A. Dumas, Ann. Chim. Phys., (3), 8. 189, 1843 ; H. Reckleben and G. Lockemann, Zeit.
angew. Chem., 21. 433, 1908; C. Renard, Compt. Rend., 136. 1317, 1903; A. Lionet, ib., 89.
440, 1879 ; H. K. Onnes, Proc. Akad. Amsterdam, 11. 883, 1909 ; E. H. Keiser, Chem. Journ.,
10. 249, 1888; J. Lowe, Dingler's Journ., 211. 193, 1874; F. G. Hahn, Ann. Pharm., 129. 57,
1864; J. J. Berzelius and P. L. Dulong, Ann. Chim. Phys., (2), 15. 386, 1820; C. Renard, Coinpt.
Rend., 136. 1317, 1903.
32 R. Fresenius, Zeit. anal. Chem., 4. 177, 1865 ; C. Voit, ib., 15. 432, 1876 ; H. C. Dibbits,
ib., 15. 121, 1876 ; P. A. Favre, Ann. Chim. Phys., (3), 12. 223, 1844 ; H. V. Regnault, ib., (3),
15. 129, 1845; J. D. van -der Plaats, Rec. Trav. Chim. Pays-Bos, 6. 45, 1899; E. W. Morley,
Amer. Journ. Science, (3), 30. 140, 1885 ; (3), 34. 199, 1887 ; Journ. Phys. Chem., 3. 241, 1905 ;
Jtmrn. Amer. Chem. Soc., 26. 1171, 1904; M. V. Dover and J. W. Marden, ib.. 39. 1317,
1917; G. P. Baxter and R. D. Warren, ib., 33. 340, 1911 ; A. T. McPherson, ib., 39. 1317,
1917 ; J. W. Marden and V. Elliott, Journ. Ind. Eng. Chem., 7. 320, 1910 ; H. Cavendish, Phil.
Tran*., 56. 201,1766.
" E. D. Ardery, Met. Chem. Eng., 14. 260, 333, 1916.
HYDROGEN 291
§ 3. Chemical Affinity
There are agents in nature able to make the particles of bodies stick together by very
strong attractions. And it is the business of experimental philosophy to find them out —
Isaac Newton.
All things act according to their nature. The atoms of the different elements join
because they possess a tendency to combine with definite other atoms, and one affinity is
overpowered by another stronger affinity so that atoms far from being pushed or passively
pressed into combination, are themselves actively pushing. P. Carus (1913).
The cause of chemical action has mystified man from the earliest ages, and there
• is no prospect of an immediate solution. The crucibles, pelicans, and alembics
of the working alchemists of the Middle Ages must have demonstrated every day
in a thousand different forms that matter seems to be endowed with properties
or to possess a kind of vis occulta — in virtue of which two or more dissimilar sub-
stances, when brought into contact, give rise to other forms of matter possessing
properties quite distinct from the original substances. The process of change is
called a chemical reaction. At present, chemical action can only be referred back to
the presence of a selective force, indwelling in the different kinds of matter, which leads
certain substances, under certain conditions, to undergo chemical change. This selective
force is called chemical affinity.
Nearly five centuries before Christ, Empedocles attributed the various changes
which occur in the form of matter to the operation of two motive forces which he
personified by assuming love to be a dynamic attractive force which induces the
union of substances, while an analogous repulsive force, hate, effects their separation.
Union was regarded as a marriage of the elements, decomposition a divorce ; love
unites, hate scatters :
All through hate are split to shapes diverse ;
Each through love draws near and yearns for each. . . .
Aristotle (c. 320 B.C.) rightly maintained that Empedocles' two forces — harmony
and discord ; love and hate ; attraction and repulsion — are two different aspects
of one motive force, because the formation of a new combination must involve the
disruption of a previous one, and the decomposition of one system must involve
the production of another system.
Not very long after the time of Empedocles, Hippocrates postulated that when
two substances unite to form a compound, they must possess one common principle
or bond of kinship, for, said he : " like unites only with like." In later centuries,
Hippocrates' maxim seems to have given rise to the idea that substances with
kindred qualities react chemically ; the greater the resemblance between two specific
forms of matter, the more likely are they to enter into combination, and the more
stable the resulting product. It was argued that the metals are akin to mercury,
and therefore " mercury devours the metals," or " the metals lick up the mercurj^"
cequalitas enim amicitice parens est ; and the word affinity was coined near the end
of the thirteenth century — in a work De rebus metallicis attributed rightly or wrongly
to Albertus Magnus ^ — in order to connote the idea that the ajfflnitas (relationship)
between combining substances is the cause of their union. Sulphur burns the metals,
said Albertus Magnus, because of the affinity it has for these substances — for the
metals were themselves supposed to contain the common principle sulphur. J. J.
Becher 2 expressed the same idea in 1669 when he stated that one substance
attrahet another in virtue of its qffinitas, and a reactio ensues ; the stronger the
attraction, the more vigorous the reaction. J. Mayow 3 used the term affinity in
the same sense in 1674. He said :
Nitro-aerial spirit and sulphur are engaged in perpetual hostilities with one another
. . . and each has a great affinity and relationship with salt, for these very active elements
are being married to salt as to a fitting bride, and are fixed in its embrace.
J. Mayow also used the word combinetur or combinentur in speaking of the congressus
of different substances. The necessary similitude, relationship, kinship, or family
292 INORGANIC AND THEORETICAL CHEMISTRY
tie between reacting substances was taken for granted by J. C. Barchausen in his
Pyrosophia (Lugduni Batavorum, 1698), and the then chemical world generally.
Heretics, however, did insist that opposite natures are best suited for chemical
imion, and, following Empedocles, the principle was enforced by analogies drawn from
the theory of marriage. For example, in his book Elementa chemice (Lugduni Bata-
vorum, 1732), Hermann Boerhaave maintained that dissimilar substances show the
greatest tendency to combine with one another, and he metaphorically compared
affinity with love : amicitice si amor dicendus copulce cwpido — if love be called the desire
for marriage ; and in 1837, J. B. A. Dumas said that we must allow that there is some
truth in this poetic comparison. A. F. de Fourcroy (1801) ^ also emphasized the
fact that the concept of affinity is a generalization largely derived from observations
made on reactions between bodies of dissimilar natures, or between unlike
particles ; and the existence of this force is regarded as one of the first principles
of dynamical chemistry. There is, however, nothing to show that the force which
binds two dissimilar atoms to say chlorine and hydrogen is intrinsically different
from the force which binds a pair of hydrogen or a pair of chlorine atoms together.
Although unlike elements have a greater tendency to unite than like elements,
F. W. Clarke ^ has shown that there appears to be a preference for neighbouring
elements in the horizontal rows of the periodic table, rather than for those more
remote. Thus, silicon follows its neighbour aluminium in the magnitude of its
atomic weight, and the alumino-silicates form the most extensive class of stable
minerals ; similarly, phosphorus is nearer than arsenic to aluminium, and the
aluminium phosphates are more common than the arsenates ; while with copper
nearer to arsenic than to phosphorus, the arsenates are more common than the
phosphates. Another striking illustration is furnished by the compounds of oxygen,
sulphur, selenium, and tellurium. The oxides or oxidized salts are the most common
with the elements of low atomic weight ; from manganese and iron, the sulphides
are the most abundant ; while selenium and tellurium are more often united with
the metals with the larger atomic weights.
H. Boerhaave distinguished what he called cohesion between the parts of the
same substance from the affinity between the parts of different substances. Boer-
haave virtually used the term affinity to connote the tendency of different kinds of
matter to unite with one another ; and the term is therefore appHed to that pecuHar
selective force or form of energy which is the origin of all chemical changes. This
definition has nothing to say about the similarity or dissimilarity of the reacting
substances, and it makes no reference to the very real difficulty in distinguishing
clearly between chemical and physical changes. This definition has been repeated
with superficial variations of phraseology by most writers.
Although the law of gravitation has been styled " the most extensive generaliza-
tion to which the human intellect has ever attained," Isaac Newton did not unduly
speculate on the cause, but he employed the term gravitation to signify, in general,
any force by which bodies tend towards each other, whatsoever be the cause. He
said :
To show that I do not take gravity for an essential property of bodies, I have added
one question concerning its cause, choosing to pi'opose it by way of a question because I am
not yet satisfied about it for want of experiments.
Newton applied the gravitation concept to atoms, and in this sense he was the
founder of molecular as well as of celestial mechanics. The propensity of two bodies
to react chemically was attributed to the attraction of the particles of the one for
the particles of the other, pair by pair ; and conversely, when a compound of two
bodies is decomposed by a third body coming into the field of action, the particles
of the intruding body were supposed to attract the one and repel the other con-
stituent. The struggle between the three kinds of particles was supposed to be
decided by the resultants of two pairs of forces. Shortly after Isaac Newton had
published his views on chemical attraction, St. F. Geoffroy changed Newton's term
HYDROGEN 293
attraction to affi^iity, and attempted to make a table in which the powers which
different bodies possess of uniting with one another are represented by numbers.^
The same concept, if not the name, was used by J. R. Glauber (1648), F. de la Boe
Sylvius (1659), A. L. Lavoisier (1783),7 and others. For instance, R.' Boyle (1664)
used the terms coalition and association^ for the concept affinity. Both
T. Bergmann and G. L. L. BufEon,^ following Newton, tried the hypothesis
that the forces between the constituent particles of a body are the same in kind as
those which determine the relations of the heavenly bodies ; but both abandoned
the idea as impracticable, because it was believed that the atoms are so close
together that their shapes must interfere with the attractive force and make the
theorems to be solved inextricably complex.
The idea that the atoms were retained each to each by hooks or other mechanical
means is suggested inLucretius' poem, and was adopted byN.Lemery. lo R. J, Haiiy
supposed that different particles have different polygonal solid shapes which when
chemically compounded pack themselves together like so many solid bricks.
There have been many attempts to evade the use of the word affinity, and as
alternatives, the terms electric attraction, atomic gravitation, chemical activity,
chemical avidity, chemical energy, chemism, etc., have been variously suggested ;
but the original term chemical affinity is convenient, provided it be kept in its place
— in verbis non simus faciles. To say that " oxygen unites with hydrogen because
it has an affinity for it " explains nothing, but simply restates the fact in different
words. Many examples of similar pseudo-explanations might be given : "Hydrogen
burns because it is combustible," " morphine induces sleep because of its soporific
qualities," " arsenic causes death because it is a poison," " potassium carbonate
absorbs moisture from the air because it is deliquescent." Explanations and
definitions of this kind are so rife that the fault has been given a name : cir cuius
in definiendo. We quite recognize with Isaac Newton (1675) that " to tell us that
every species of things is endowed with an occult specific quality by which it acts
and produces manifest effects is to tell us nothing," yet is it hardly fair to say that
the term chemical affinity is a veil which covers our ignorance in obscure language.
What H. F. Link n wrote at the end of the eighteenth century — 1795 — might almost
have been written to-day :
Although the term affinity has sprung from the dark ages of chemistry, and appears at
first glance to be mystical and unprofitable, yet, it is certain that since the causes of all
chemical phenomena have been referred to one single cause, chemistry has made extra-
ordinary progress for which it has to thank the definition- — affinity is the cause of chemical
action.
With this understanding, chemical affinity can be conveniently regarded as " the
driving force of a chemical reaction." In 1887, W. E. Ayrton and J. Perry expressed
the idea that
„ ^. , ., Driving force
Reaction velocity =—.= — . ^
•^ Resistance
Consequently, if we could measure the chemical resistance offered by substances
to undergo chemical change, it would be possible to get a definite and quantitative
idea of chemical affinity from measurements on the velocity of a reaction. No real
advance can be made in the study of chemical affinity until a method of measurement
has been devised. As W. Whewell (1840) expressed it : " In all attempts to explain
the processes of nature, the proper course is first to measure the facts with precision,
and then to endeavour to understand their cause."
References.
1 Albertus Magnus, De rehus metallicis et mineralibus, Rouen, 1476.
^ J. J. Becher, Physica svhterranea, Lipsiae, 1669.
^ J. Mayow, Tractutus quinque medico -physici, Oxford, 1674.
* A. F. de Fourcroy, Systhne des connaissances chimiques, Paris, 1801.
^ F. W. Clarke, The Data of Geochemistry, Washington, 1916.
294 INORGANIC AND THEORETICAL CHEMISTRY
« I. Newton, OiJ/ici'5, London, 1704 ; W. V. Harcourt, Phil. Mag., (3), 28. 106, 478, 1846;
(3), 29. 185, 1846; St. F. GeofEroy, Mem. Acad., 202, 1718 ; 20, 1720.
' J. R. Glauber, Novi fumi philosophi, Franckfurt, 1648 ; F. de la Boe Sylvius, Opera omnia,
Paris, 1671 ; A. L. Lavoisier, Reflexions sur la phlogistique, Paris, 1783.
* H. Kopp, Oeschichte der Chemie, Braunschweig, 2. 306, 1844 ; F. A. Lange, Geschichte des
Materialismus, Leipzig, 2. 295, 1908 ; R. Boyle, Considerations and experiments touching the
origin of qualities and forms, Oxford, 1664.
^ I. Newton, Opticks, London, 1717 ; T. Bergmann, De attractionibus electivis, Upsala, 1775 ;
G. L. L. Bufifon, Epoques de la nature, Paris, 1778.
^" N. Lemery, Cours de chymie, Paris, 1675 ; R. J. Haiiv, Ann. Chim. Phys., (2), 14. 305,
1820
11 H. F. Link, Beitrdge zur Physih und Chemie, Rostock, 1795-7.
§ 4. The Measurement of the Affinity between the Acids and the Metals
I often say that if you can measure that of which you speak, and can express it by a
number, you know something of your subject ; but if you cannot measure it, your knowledge
is meagre and unsatisfactory.- — -Lord Kelvin.
Towards the end of the eighteenth century, T. Bergmann found it necessary
to issue a warning against " some chemists who consider thermometers and such-
like measuring instruments to be physical subtilties, superfluous and unnecessary
in a chemical laboratory," and he further emphasized his belief in the importance of
generally " so comparing an effect with its cause as to determine the exact quanti-
tative relation between the two." At this day, the resources of the physical labora-
tory are commandeered by the chemist in his quest after the quantitative relations
between causes and their effects.
Chemical reactions are not instantaneous processes, but are propagated with
finite measurable velocities which may range from the explosion wave travelling
through a mixture of hydrogen and oxygen at a speed not much less than 10,000 feet
per second ; or it may be so slow that years are needed to detect an appreciable
change. It is assumed that every elementary atom and ever)^ molecule is charged
with a definite amount of energy which is a measure and cause of its chemical
affinity ; chemical affinity is assumed to be the driving force of chemical reactions ;
and the speed of chemical reactions, other things being equal, is proportional to the
driving force. It will be obvious that if a ball be sent rolling with a velocity of
20 cm. per second, the force applied to the ball will be twice as great as would be
required to make the ball travel with a velocity of 10 cm. per second during the same
time. Neglecting friction, the intensities of the two forces are proportional to the
velocities which they impart to each unit of mass during the same time.
The relation between the speed o£ a chemical reaction and affinity. —
In an important book, Lehre von der Venvandtschaft (Dresden, 28, 1777), C. F. Wenzel
tried to determine the affinities of the metals for different acids by comparing
the rates at which the metals liberate gas from acids of different concentration.
He found that if an acid of a given concentration dissolves one unit of metal per
hour, an acid of half that concentration will take two hours to dissolve the same
amount of metal. The velocity of these reactions can be measured by finding the
amount of gas liberated per minute, or the amount of acid or of metal consumed,
say, every minute. Then, at any given moment :
Amount of gas Hberated_Acid consumed
y~ Time occupied ~~Time occupied
The affinity of a metal for an acid depends on the concentration of the acid. In
fine, the velocity of the chemical action at any instant is proportional to the
concentration o! the reacting substances. This is sometimes called Wilhelmy's
law, because L. Wilhelmy (1850) i demonstrated the generaHzation by measurements
on the speed of inversion of cane sugar. A comparison of the rate of dissolution
of, say, magnesium in hydrochloric acid of different concentrations, jjjN, ^iV, and
HYDROGEN
50
Minutes.
100
^N, where iV here denotes an equivalent weight of HCl per litre, brings out clearly
the increase of speed with increasing concentration. The slopes of the curves in
Fig. 4 are proportional to the speed of the attack..
If a solution be of such a concentration that it contains a gram-molecules of
acid per unit volume, then at the end of a certain time
t, X gram-molecules of the acid per unit volume will ,oo
have been consumed, and the solution will contain a—x ».
gram-molecules of the acid per unit volume. Hence, the c
velocity of the reaction will gradually slacken down. At ^
the beginning, the velocity V will be proportional to a ; '^ so
that is, V=ka, where A; is a constant of proportion ; and at "
the end of the time t, the velocity will be V^k{a—x). e
Hence, in C. F. Wenzel's experiment, when x=^a, ^
the reaction is only progressing half as fast as at the
beginning when ic=0. The speed of the reaction at
different times is illustrated by the slope of the curve ^'^- ^•-^^'°^^*^^^ °^,^^-
T7^. ^ , , , , . •' . , , • • t nesium in Hvarocnionc
m J^ig. 5, where the abscissa axis represents time, and Acid of Different Concen-
the ordinate axis, the velocity expressed in any con- trations.^
venient units — say, volume of gas evolved per minute.
The velocity of a reaction is not always quickest at the start. When some
reacting substances are brought into contact, a certain interval of time — called the
period of induction — elapses before
the reaction can proceed " full speed
ahead." With zinc and dilute hydro-
chloric acid, for example, the dotted line,
Fig. 5, shows how the speed gradually
increases, reaches a maximum, and then
gradually diminishes as described by Wil-
helmy's law.^ A similar period of induc-
tion has not been observed with mag-
nesium. During the period of induction,
therefore, some action takes place by
which the resistance to combination is Fig. 5
decreased, or a more favourable con-
dition for combination is inaugurated.
By measuring the rate at which hydrogen is liberated per minute per unit area
of the different metals on the same sample of dilute acid, it is possible to get a rough
idea of the relative affinities of the different metals for
that particular acid. The comparison of the effect of
dilute hydrochloric acid on zinc and magnesium — Fig. 6
— shows that the dissolution of magnesium gives a steeper
curve than does zinc. This means that magnesium dis-
solves faster than zinc. Experiments with other metals
show that with dilute hydrochloric acid, starting with
the most vigorous, this order is : Potassium, sodium,
calcium, magnesium, zinc, and iron.
The surface of the dissolving metal is supposed
to be constant. As near as can be determined, the p^^ q — Dissolution of Zinc
rate of attack, for any given concentration of acid, and Magnesium by Dilute
increases or decreases proportionately with the surface Hydrochloric Acid,
exposed to attack. Exact measurements are difficult
because so many disturbing influences are at work— local rise of temperature;
bubbles of gas protecting the surface of the metal from attack; variations
in the surface of the metal during the action ; etc. Still, the conclusion just
indicated is in harmony with a great deal of work on a variety of other simple
heterogeneous reactions. The generalization now under discussion is a special
"T
^
~"
«
«o
1
o
1
T^N. HvHrocHloric Acid.
o
j
1
r
V,
_,
r
u
\^.
1"
.._,
^
-i
^>S/f//
n.
_
.^
[a
n
/
2
tz
t
L
■"■
5
t=
B
8
Minutes
-The Reduction in the Speed of Chemical
Action with Time.
296
INORGANIC AND THEORETICAL CHEMISTRY
case of one still more general : The speed of a heterogeneous reaction at any
instant of time is directly proportional to the remoteness of the system
from its equihbrium condition, and to the surface of contact of the reacting
parts. M. Wildermann (1899) has shown that this generalization is generally
applicable to molecular reactions between two parts of a heterogeneous system —
e.g. the evaporation of liquids, the condensation of vapours, the solidification and
crystallization of liquids, etc. When zinc dissolves in acid, the rising bubbles of
gas keep the liquid constantly agitated, bringing fresh acid to the zinc plate, but
when a solid dissolves in a liquid, without disturbances of this kind, the rate of dis-
solution is nearly proportional to the rate at which the solid diffuses in the liquid ;
because the nearer the liquid is to the dissolving solid, the greater its concentration,
and the layer of liquid in the immediate vicinity of the solid, is virtually a saturated
solution.
By measuring the rates at which hydrogen is evolved with one metal and
different acids of equivalent concentration, an idea of the relative affinity of the
acids for the given metal can be obtained. For instance, acids containing 36 "5 grms.
of HCl (hydrochloric acid) per litre, 49 grms. of H2SO4 (sulphuric acid), and 60 grms.
of CH3COOH (acetic acid) per litre are chemically equivalent to one gram of
hydrogen. When these three acids — in equivalent concentrations — react with
magnesium ribbon (say, 0"05 grm.), the relative affinities appear to be in' the order
named :
Hydrochloric acid.
100
Sulphuric acid.
70
Acetic acid.
0-5
■ 100
50
1
h
i^
•rff
w. :
50
Minutes.
100
The first gives off most hydrogen in a given time, the latter least. Measurements
of the volume of gas (reduced to n.p.t.) evolved by the action of normal hydro-
chloric and acetic acids upon magnesium gave results
which furnished the curves shown in Fig. 7. The
difference in the speeds of attack by the two acids is
brought out clearly by the relative slopes of the two
curves. The swifter the reaction the steeper the curve.
The speed of the reaction of a given concentration
of hydrochloric acid or sulphuric acid on iron or zinc
was found by G. Lunge to be considerably reduced if
the viscosity of the medium be augmented by admixture
with glycerol, or gum arable. It is assumed that the
■p^^ rj rpj^g Speed of the explanation is partly due to the more tardy liberation
Dissolution of Magnesium of the gas bubbles from the surface of the metal in the
by Hydrochloric and more viscous medium, and this prevents contact between
Acetic Acids. the metal and the acid. That this is the whole ex-
planation is regarded as improbable because a mixture
of the acid with a little lamp-black (moistened with alcohol to make the acid
" wet " it) also acts in an analogous manner.
Returning to the law symbolized in the equation V=k{a — x), with an acid of unit
concentration, it follows that the initial velocity V=k. And k has accordingly been
called the affinity constant of the acid for the metal ; k represents the speed of
the reaction at the instant when the acid has unit concentration. The speed of a
chemical reaction is usually, not always, augmented by raising the temperature.
In illustration, with J2V-sulphuric acid, a rise of temperature from 12'8° to 35°
nearly doubled the speed of dissolution of zinc ; and a rise from 128° to 55° nearly
trebled the speed. The result of this discussion shows that the velocity of a
chemical reaction is proportional (1) to the " affinity constant " between the
reacting substances, and (2) to the concentration of the reacting substance ;
while (3) the velocity of a chemical reaction is augmented by raising the
temperature.
HYDROGEN 297
References.
1 L. Wilhelmy, Pogg. ^?i7?., 81. 413, 499, 1850.
2 T. Ericson-Auren, Ze.it. arwrg. Chem., 27. 209, 1901 ; M. Wildermann, Phil. Mag., (6), 18.
538, 1909 ; M. Tarle, Studien fiber den Zusammenhang zwischen der Reaktionsfdhigkeit und Dis-
soziation, Weida i. Th., 1912; G. Lunge, Chem. News, 35. 92, 1877.
§ 5. Opposing Reactions. Guldberg and Waage's Law
Chemical action is reciprocal, and its effect is the result of a mutual tendency to com-
bination.—C. L. Berthollet (1803).
In an aggregate of molecules of any compound, there is an exchange constantly going
on between the elements which are contained in it. — A. W. Williamson (1850).
Some of the earlier chemists i — e.g. St. F. Geoffroy (1718), and Torbern Bergmann,
in his De attractionihus electiviis (Upsala,1775) — argued that the result of a chemical
change must be in favour of that substance with the stronger affinity. Accordingly
St. F. Geoff roy compiled what he called Tables des differ ents rapports observes en
chimie enlre differ entes substances ; and T. Bergmann, afi&nity tables intended to
show the order in which the different substances would displace one another from
a given compound. It was argued that if A displaces B from one cotnpound, and B
displaces C from another compound, the order of the affinity of these three substances
is A, B, G. It was clearly recognized that this method of work does not give a
numerical measure of affinity, but it was thought that relative results were obtained.
The suggestion is certainly a good trial hypothesis. Let us compare it with the
facts.
We have seen that iron can displace hydrogen from its combination with oxygen ;
hence iron has a stronger affinity than hydrogen for oxygen. Similarly, we have
seen that hydrogen can displace iron from its combination with oxygen ; con-
sequently, hydrogen has a stronger affinity than iron for oxygen. These two
conclusions are contradictory ; both cannot be true. Therefore, the affinity hypo-
thesis must be either false, or soyne powerful perturbing influence must be at work.
In 1799, C. L. Berthollet 2 clearly recognized an important disturbing factor,
and described it in an heretical but prophetic work entitled Recherches sur les lois
de Vaffinite (Paris, 1801). Berthollet noticed large quantities of trona — sodium
carbonate — on the shores of the natron lakes of Egypt. He suggested that the
sodium chloride brought down by the rivers was decomposed by the calcium car-
bonate present on the banks of these lakes :
CaC03+2NaCl-CaCl2+Na2C03
Berthollet knew, quite well, that this reaction is the reverse of that which usually
obtained in the laboratory, for sodium carbonate, when added to calcium chloride,
precipitates calcium carbonate :
Na2C03+CaCl2=CaC03+2NaCl
but, added Berthollet, the large masses of calcium carbonate on the banks of these
lakes is able to " strengthen " the weak affinity of carbon dioxide for sodium, or of
chlorine for calcium. Here Berthollet brings the disturbing factor into bold relief :
Chemical action is conditioned not only by aflSnity but by the relative
concentrations of the reacting bodies — Berthollet 's law. Excessive concentra-
tion can compensate for a weakness of affinity. A chemical reaction can be
reversed by changing the concentrations of the reacting bodies. We must apply
Berthollet's hypothesis to the reaction under consideration — the action of iron
on steam.
The reaction'^ between iron and steam. — At the outset, it will be obvious
298
INOKGANIC AND THEORETICAL CHEMISTRY
that we have to deal with two opposing reactions : steam reacts with iron to produce
iron oxide and hydrogen :
3Fe+4H20=Fe304+4H2
and iron oxide and hydrogen react to produce steam and metallic iron :
Fe304+4H2=3Fe+4H20
J. L. Gay Lussac and H. V. Regnault ^ showed that these opposite effects are not
produced merely by a difference of temperature, for, at every degree of temperature
from the dullest to the brightest red heat, the action takes place sometimes in one
way and sometimes in the other. All depends on the relative proportions of
hydrogen and water vapour which are present. If hydrogen be in excess, iron
oxide is reduced and water vapour is formed ; while if the water vapour be in excess,
the iron is oxidized and hydrogen is formed. In either case a mixture of water
vapour and hydrogen is obtained. Reactions of this kind were regarded by the
older chemists as an example of the conflict of affinities, and they grouped such
reactions as a class exhibiting what they called reciprocal affinity — affinitates reciprocce,
C. M. Despretz found zinc, tin, cobalt, and nickel to act in the same manner as
iron, and H. V. Regnault found uranium and cadmium to do the same.
It is therefore clear that two antagonistic changes take place simultaneously in
the system. The result of the change will be determined by the fleeter reaction.
When steam is passed over red-hot iron, the hydrogen
does not get much chance, it is carried away into the
gas jar before it has had time to set up the reverse
change. Similarly, when hydrogen is passed over red-
hot iron oxide, the steam does not get a chance, for it
is carried away from the reduced iron by the stream
of hydrogen. In order to study the affinity relations
between these different substances, they should be
heated in closed vessels so that the products of the
reaction are not whisked away from the seat of the
T-.- o -r^ -i-,, . ^ ^ reaction as soon as they are formed. The result is
Fig. 8. — Equilibnum Curve of .1 • tj. •£ 2.1. x- j.
Steam and Hydrogen in the ^^^^ ^^^7 ^urious. It seems as if the reaction stops
presence of Iron. after a time. At any rate, if the temperature remains
constant, no further change can be detected, however
long the system be heated. In other words, the system assumes a state of
equilibrium. G. Preuner's experiments show that at 200° the system is in
equilibrium when the volume of the steam is to the volume of hydrogen nearly as
20 : 1. Otherwise expressed, for equilibrium at 200° :
Volume of hydrogen_ 1
Volume of steam "~20
If a mixture of one volume of hydrogen and twenty volumes of steam be passed over
iron filings or over iron oxide at 200°, no apparent change will occur, for the mixture,
after passing through the tube at 200°, will have the same composition as when it
entered, if no secondary actions occur. If more than this proportion of hydrogen
be present at 200°, some iron oxide will be reduced until the equilibrium ratio is
obtained. If the temperature be raised, the velocities of the two reactions are altered
in such a way that at 440° the volume of steam will be to that of hydrogen nearly
as 6 : 1, or as 1 : 0*17 ; and at 1500°, as 1:1. This means that if equal volumes
of steam and hydrogen be passed over iron filings or iron oxide at 1500°, no change
in the composition of the gaseous mixture will be perceptible. The results are
summarized in Fig. 8. The curve showing the percentage amount of, say, steam
in the system at different temperature divides the plane surface into two regions.
If the state of the system be described by a point in the region of oxidation, iron will
\m
>x:
:;:;:;
fc^
r-
■■;■:•■:
Is:
•x;
Reqion
\ \of 1
Req
on
;^:-:
Oxidation/
of :. '.
/
Heduchon?.
A
:■;:;:
;i;;i;
4
%i-
:;;■:
]:[>■■
y
•=i-:-
'■■i
m
::;:•:
^^^
0 50 100
Per Cent- Sf-ea/n
HYDROGEN 299
be oxidized ; and conversely, if the state of the system be described by a point
in the region of reduction, iron oxide will be reduced.
Chemical equilibrium is dynamic, not static— Let us now try to picture
what is taking place. Start with metallic iron and steam. At the outset when the
reaction is just starting, the velocity of decomposition of the steam will be greatest
because the system then contains the greatest amount of reacting substance, and
because the velocity of chemical action is proportional to the concentration of the sub-
stances taking part in the reaction. From this moment, the velocity of the reaction
gradually slows down as the concentration of the reacting steam becomes less and
less. On the other hand, the velocity of the reverse action will be zero at the com-
mencement, because none of the reacting hydrogen is then present. The speed of
the reverse change will become faster and faster as the product of the first-named
reaction — hydrogen — accumulates in the system. Ultimately, a point will be
reached where the velocities of the two opposing reactions will be equal. The one
will be balanced by the other. The reaction will appear to have stopped in spite
of the fact that more or less of the original substance still remains untransformed.
The system is then in a state of equilibrium. No further change will occur, however
long the substances be heated under the same physical conditions of temperature,
etc. Chemical changes of this kind are conveniently styled opposing or balanced
reactions, according as it is wished to emphasize the fact that the one reaction is
undoing the work of the other ; or that the speeds of the opposing reactions are
ultimately so nicely balanced that the composition of the system remains constant.
The idea of a dynamic and not a static equilibrium is such reactions was emphasized
by A. W. Williamson ^ about 1850, while studying the action of acids on alcohol.
He said :
An exchange is constantly going on between the elements of the molecules of a compoimd
so that each atom of hydrogen in the molecules of HCl present in a drop of hydrochloric
acid does not remain quietly in juxtaposition with the atom of chlorine with which it first
united, but, on the contrary, is constantly changing places with the other atoms of hydrogen,
or what is the same thing, changing its chlorine atom,
A. W. Williamson further adds that when a system appears to be in equilihrio,
that condition " is only kept up by the number of exchanges in one direction being
absolutely the same in each moment of time as those in the opposite direction."
This is a very clear explanation of M. Lieben's contention that when substances
are decomposed by heat and the products of decomposition are not removed as they
are formed, the decomposition is never complete ; and the small quantity of
the original substance, which always remains undecomposed, produces a kind of
** molecular equilibrium."
Terminology.- — Reversed pointers " ;^ " are conventionally used in place of the symbol
" = " for opposing reactions, so as to indicate that two reactions are proceeding simul-
taneously " from right to left " and " from left to right." Accordingly, the reaction imder
consideration is symbolized : 3Fe+4H20^Fe304+4H2. Opposing reactions are also
called incomplete or reversible reactions in contradistinction to irreversible or complete
reactions typified by the action of zinc on sulphuric acid, where the reaction is completed
in one direction and is not opposed by a counter reaction.
Concentration and active mass. — It is not difficult to see that the absolute
quantities of steam, hydrogen, iron, and iron oxide, in the reaction under considera-
tion, do not matter. The velocities of the two opposing reactions, and therefore
the distribution of the reacting substances, when in equilibrium, is determined
by the relative concentrations of the changing substances. This is conveniently
expressed by the number of gram-molecules of each present in unit volume. Thus
18 grams of water — H2O — per litre represents one gram-molecule ; 36 grams of
water per litre, two gram-molecules ; etc. The concentration of a reacting substance
is sometimes (inappropriately) called— after C. M. Guldberg and P. Waage (1864)
—its active mass or— after C. L. BerthoUet (1803) -its chemical mass. How-
ever, it is not mass but concentration which is the determining factor in chemical
300 INORGANIC AND THEORETICAL CHEMISTRY
equilibria, and hence, C. N. Lewis (1907) proposed to substitute the term activity
in place of active mass.
If the surface of the iron were doubled, it is true that twice as many molecules
of the black oxide, Fe304, might be formed in a given time by the decomposition
of the steam, but then twice as many molecules of Fe304 would be decomposed
by the hydrogen in the same time. Hence, the amount or the concentration
of the solid in a gaseous reaction can have no appreciable influence
on the equiUbrium ; although it may affect the speed at which the
state of. equilibrium is attained. In studying equilibria in gases and
liquids, anything which separates in the solid condition is often supposed to be
thrown out of the reacting system because the state of equilibrium is independent
of the concentration of the solid ; and a liquid which separates when studying
gaseous equilibria, is also supposed to be thrown out of the reacting system. The
vapour pressure of water, for instance, in presence of its own liquid is independent
of the amount of liquid water present. Suppose also that the back reaction between
hydrogen and black oxide of iron, Fe304, at the temperature of the experiment
furnishes black ferrous oxide, FeO, the reaction would then be symbolized :
SFeO+HgO^FegOi+Hg
Determinations of the ratio H2O : H2 would not give any information as to the
accuracy of the assumption. The principles of opposing reactions just outlined
are summarized : In a system of reacting bodies, the effect of
each substance is proportional to its concentration, and the total
effect is proportional to the product of the molecular concentra-
tions of the reacting substances. This is one statement of what
is called Guldberg and Waage's law of mass action, because the
ideas of Berthollet were expressed in this form by C. M. Guldberg
and P. Waage in an important memoir, Etudes sur les qffinites
chimiques (Christiania, 1867), published first in Norwegian in 1864.^
L. P. Cailletet (1869) ^ has shown that when, say, sodium amalgam
acts on water, the hydrogen exerts no back action because it passes
away from the seat of the reaction as soon as it is formed. Sup-
pose the reaction be conducted in a thick-walled vessel, B, in Fig. 9,
Fig. 9. capable of withstanding great pressure and fitted with a manometer,
A, in order to measure the pressure of the hydrogen as it accumulates
in the apparatus. The speed of the reaction gradually slackens and finally stops.
With dilute sulphuric acid and zinc, this occurred when the hydrogen was exerting
a pressure of nearly 20 atmospheres. The cessation of the reaction is not a mere
mechanical effect because an equal pressure exerted by an indifferent gas will not
do. The idea of preventing chemical action in this way is said to have been first
tested by C. Babbage in 1813, for he confined concentrated hydrochloric acid in a
hole 30 inches deep and 6 inches wide drilled in limestone rock (Chudley Rocks,
Devon) and plugged up the hole. He thought the pressure of the developed gas
would prevent the action of the acid on the limestone. No definite conclusion was
drawn from the experiment.
Still further, as N. N. Beketoff (1864) has shown, if salts like zinc sulphate,
copper acetate, silver sulphate, etc., be exposed to hydrogen under great pressures
— over 100 atmospheres — the metal is precipitated and the acid is re-formed.
Hence, the reaction Zn+H2S04=ZnS04+H2 is reversible when conducted in a
closed vessel.
It might be added that in some cases the back reaction does not proceed along
the same path as the forward reaction. Thus, hydrogen at 130 atmospheres pressure
(70°) precipitates cuprous oxide from copper acetate, and at 150 atmospheres
(120°) metallic copper. It might also be added that the work done by affinity
during, say, the reaction between zinc and dilute acid, may be roughly regarded
as proportional to the pressure exerted by the gas provided the reacting system
HYDROGEN 301
sufEers no change in volume. If v denotes the change in volume, and p the pressure
registered by the manometer when the reaction just stops, the work W done by
affinity at a constant temperature will be W—pv.
Chemical afi&nity. — To summarize the preceding discussion : chemical affinity
is a convenient term for the driving force which causes certain substances
to combine together and to remain united with one another. (1) Unlike gravitation,
chemical affinity seems to act only when the reacting substances are in contact
with one another ; or, as it is sometimes expressed, " when the substance are
brought within insensible distances of each other." (2) Unlike gravitation, chemical
affinity is a selective force, and it seems to act more intensely the more unlike the
substances are ; or, as it is sometimes expressed, " Like reacts with the unlike."
(3) The affinity of an element is not only definite as to the kind, but it is also definite
as to the quantity of the elements which enter into combination. In this again it
differs from gravitational attraction. The quantitative characteristics are described
by the laws of chemical combination. (4) The strength of the affinity varies with
changes in the conditions of temperature, pressure, fight, etc. This has not been
noticed with gravitational attraction, (5) The effects produced by chemical affinity
are modified by the relative concentrations — active masses — of the reacting
substances.
The ideas developed in this section were not so clear to the old workers, not
even to Berthollet himself, for Berthollet appears to have confused the incomplete-
ness of certain reactions with the law of multiple proportions. The confusion gave
him some strong arguments in the celebrated Berthollet v. Proust Controversy.
Proust did not know enough to clarify Berthollet's argument.
References.
1 St. F. Geoffroy, Mem. Acad., 202, 1718 ; 20, 1720 ; G. E. Stahl, Chymia rationalis, Leipzig*
106, 1720.
2 C. L. Berthollet, Mem. Nat. Inst., 3, 1799 ; Ann. Chim. Phi/s., (1), 36. 302, 1801 ; (1), 37.
225, 1801 ; (1), 38. 113, 1801 ; Essai de statique chimie, Paris, 1802; W. OstwahTs Klassiker,
74, 1896. '
3 J. L. Gay Lussac and H. V. Regnault, Ann. Chim. Phys., (2), 1. 33, 1816 ; (2), 62. 372,
1836; C. M. Despretz, ib., (2), 43. 222, 1830; H. St. C. Deville, Campt. Rend., 70. 1105, 1201,
1870 ; 71. 30, 1870 ; H. Debray, ib., 88. 1341, 1879 ; G. Preuner, Zeit. phys. Chem., 47. 385,
.904.
* A. W. Williamson, B. A. Rep., 65, 1850; Phil. Mag., (3), 37. 350, 1850; Journ. Chem.
Soc., 4. 229, 1852 ; Alembic Club Reprints, 16, 1902.
6 C. M. Guldberg and P. Waage, Forh. Viden. Sels. Christiania, 35. 92, 111, 1864; Jo7irn.
prakt. Chem., (2), 19. 69, 1879 ; W. Ostwald's Klassiker, 104, 1899.
« L. P. Cailletet, Compt. Rend., 68. 395, 1869; M. Berthelot, ib., 68. 536, 780, 810, 1869 ;
N. N. Beketoff, ib., 48. 442, 1859 ; W. Nernst and G. Tammann, Zeit. phys. Chem., 9. 1, 1892 ;
G. Quincke, Pogg. Ann., 160. 118, 1877 ; C. Babbage, 1813. By some inadvertence, the source of
Babbage's statement is not indicated on the writer's record card. A great search has been made
to locate the original reference, but without success.
§ 6. The Solubility of Hydrogen
Hydrogen is slightly soluble in water. About 1803, W. Henry i found that
100 c.c. of water absorb 1*56 c.c. of hydrogen ; J. Dalton found 2*5 c.c. ; and T. de
Saussure, 4*55 cc. N. Paul noted that the solubiHty is augmented by pressure, so that
100 c.c. of water can be made to absorb about 33 c.c. of hydrogen. Accordmg to
A. T. y Marti, water can be made by degrees to absorb more and more hydrogen, so
that in two years water will take up not quite its own volume of the gas. It was
suggested that hydrogen suboxide, H4O, is formed, and C. J. B. Karsten even supposed
this oxide to be formed by saturating cold water with hydrogen sulphide, and
removing the sulphur by certain metals. There are no satisfastory reasons for
supposing the existence of this oxide.
302 INORGANIC AND THEORETICAL CHEMISTRY
R. W. Bunsen (1855) 2 thought that the solubility of hydrogen in water is not
affected by variations of temperature between 0° and 24°, so that between these
temperatures one volume of water absorbs the equivalent of 00193 vol. of
hydrogen ; but his method of measurement was probably not sensitive enough to
detect the difference, since W. Timofejeff and others have shown that there is quite
an appreciable change between these temperatures. Let j3 denote the absorption
coefficient used by R. W. Bunsen ; it represents the volume of gas reduced to 0°
and 760 mm. which is absorbed by one volume of the solvent when the pressure of
the gas itself, without the partial pressure of the solvent, amoimts to 760 mm. ; and
let j3' denote the absorption coefficient when the total pressure of the gas and water
vapour is the barometric pressure, 760 mm. The solubiHty of a gas — symbohzed S —
can also be represented as the volume of gas absorbed by unit volume of the solvent
at the temperature of the experiment — consequently, the solubihty ^ of a gas at
d° is 1+0-00367^ times the coefficient of absorption /3, or ^-:j8(l +0-00367^). The
solubihty can also be represented as the weight oj of gas in grams dissolved by
100 grams of the solvent at the temperature of the measurement and a total
pressure 760 mm., where 760 mm. represents the partial pressure of the gas plus
the partial pressure of the solvent.
The solubility o£ hydrogen in water. — The values of the constants )3 and
OJ for the solution of hydrogen in water between 0° and 24° are :
a>
and W. Timofejeff represents the coefficient of absorption j3 at a temperature d
over the range 0° to 26°, by j8=0-0215286-0-03196(9+0-0517228^2. j'or the
range from 25° to 100°, the four solubihty coefficients are :
0'
4°
8°
12°
16°
20°
24*=
0-02153
0-02079
0-02010
0-01947
0-01889
0-01837
0-01791
0-03192
003185
0-03179
0-03173
003167
003162
003157
25°
30°
40°
50°
60°
80°
100°
i8
. 0-0175
0-0170
0-0164
0-0161
0-0160
0-0160
0-0160
?'
. 00171
0-0163
00153
0-0141
0-0129
0-0085
0-0000
S
. 00156
0-0145
0-0140
0-0131
00125
0-0113
0-0107
ta
. 003156
0-03147
0-03139
0-03129
0-03119
0-0470
0-0000
Although hydrogen, hke other gases, decreases in solubihty as the temperature
rises, no definite law has been discovered for the phenomenon. According to
C. Bohr and J. Bock the solubihty of hydrogen decreases gradually with rise of
temperature and from >S=0-0203 at 0° to /S=0'0155 at 90°, and then rises to 0-0166
at 100°. Hehum is the only other gas which gives any indication of a reversal in
the direction of the solubihty curve. The formulae representing the relation
between the temperature d and the solubihty /S, based on the series
S=a-\-hd-{-cd^-\- . . . — where a^h^ c^ . . . are constants whose numerical values are
derived from the observed data — are quite empirical. L. W. Winkler (1892) sought
to prove that the percentage decrease in the absorption coefficient or solubihty
is nearly proportional to the cube root of the molecular weight of the gas in question.
A comparison between the observed and calculated results for hydrogen, nitrogen,
oxygen, carbon monoxide, and nitric oxide was satisfactory ; but T. E. Thorpe and
J. W. Rodger (1894) showed that L. W. Winkler's rule is not generally vahd, though
they found that for the same gas, the decrease in the coefficient of absorption for
any interval of temperature is nearly proportional to the corresponding decrease
in the viscosity coefficient of the solvent.
According to K. Angstrom,^ the increase in the volume of the solvent which
occurs during the solution of a gas is proportional to the amount of gas absorbed ;
and with
Nitrogen, Air. Carbon monoxide. Oxygen. Hydrogen. Carbon dioxide.
Dilation . 0-04294 004346 0-04418 0-04474 0-04204 00023
The increase in volume with hydrogen is 00016 for chloroform ; 0*0017 for benzene ;
0-0017 for methyl alcohol ; 0-00152 for ethyl alcohol ; and 0-00184 for ether
HCl
HNO,
+H2SO4
iiV-acid .
0-0186
0-0188
0-0185
4iV-acid .
0-0160
00160
0-0141
HYDROGEN 303
The dilation due to absorption was found to be independent of the nature of the
liquid, and not in agreement with W. Ostwald's statement that " the volume of the
absorbed gas is almost exactly reduced to the volume of its molecules."
The solubility of hydrogen in aqueous solutions 0! acid, bases, and salts.— A
great many empirical observations have been made on the solubility of hydrogen in
aqueous solution of acids, bases, and salts.4 The values of S for hydrochloric,
nitric, sulphuric, acetic, chloroacetic, and propionic acids at 25° are :
CH3COOH CHjCICOOH C-HsCOOH
0-0192 . 0-0189 0-0017
0-0186 0-0180(2iV) 0-0016(lJiV)
A. Christofi found the solubility of hydrogen in 95*6 per cent, sulphuric acid to be
0-01097 between 17° and 20° ; in 61*62 per cent, sulphuric acid, 0'007181 ; in
35-82 per cent, acid, 0-009544 ; and in pure water, 0-02077. The solubiHty S of
hydrogen in a JiV-solution of potassium hydroxide is 0*0167 ; and in a normal
solution, 0*0142 ; for a JiV-solution of sodium hydroxide, 0-0165 ; for a normal
solution, 0*0139 ; and for a 4iV^-solution, 0*0055. W. Knopp, H. von Euler, and
others have suggested formulae for representing the efiect of a salt on the solubility
of hydrogen in water.
The coefficient of absorption, ^8, of a 1-037 per cent, solution of ammonium nitrate at
20° is 0-01872, and for a 11-55 per cent, solution, 0-01647 ; for a 4-73 per cent, solution of
potassium nitrate ^=0-01683 (15°), and a 21-46 per cent, solution j8=0-01180 (15°); a
5-57 per cent, solution of sodium nitrate has ^=0-01603 (15°), and a 37*43 per cent,
solution, ^=0-00578 (15°). A 3*83 per cent, solution of potassium chloride has
^=0-01667 (15°), and for a 22-92 per cent, solution, ^=0-00892 (15°); with a 1-25
per cent, solution of sodium chloride, ^=0-0191 (15°), 0-0177 (20°), and for a 23-84 per
cent, solution, ^=0-00595 (15°) ; a 3-48 per cent, solution of lithium chloride has
j8=0-01619 (15°), and a 14-63 per cent, solution, jS=0-0099 (15°). For a 3-29 per cent,
solution of barium chloride, ^=0-0185 (15°), 0-0172 (20°) ; a 7 per cent, solution has
^ = 0-0172 (15°), 0-0159 (20°) ; with a 3-47 per cent, solution of calcium chloride, j8=
0-01450 (15°), and a 26-34 per cent, solution, ^=0-00519 (15°). For a 4-58 per cent, solution
of sodium sulphate, ^=0-01519 (15°), and a 16-69 per cent, solution, ^8=0-00775. For a
4-97 per cent, solution of magnesium sulphate, j8=0-01501 (15°), and for a 23*76 per cent,
solution, ^=0-00499 (15°) ; for an 8*1 per cent, solution of zinc sulphate, ^=0-001446
(15°), and a 48-4 per cent, solution, ^ = 0-00510 (15°). A 22-82 per cent, solution of
potassium carbonate has ^=0-01628 (15°), and a 41-81 per cent, solution, ^=0-0016 (15°) ;
for a 2-1 per cent, solution of sodium carbonate, jS=0-01639 (15°) ; and a 11-52 per cent,
solution, j8=0-00839 (15°).
The solubility of hydrogen in organic liquids.— Hydrogen is much more
soluble in ethyl alcohol than it is in water. The solubiHty decreases the more
the alcohol is diluted with water down to about 28 per cent, alcohol, when the
coefficient of absorption increases ; there is therefore a minimum in the solubiUty
curve. Thus, at 20° the absorption coefficient ^ is :
Per cent, alcohol 0 9-09 28-57 33-33 50-0 66*67 98-8 99-7
^ . . . 0*0184 0-0133 0-0097 0-0108 0-0187 0-0237 0-0740 00740
W. Timofejefi (1890) found the absorption coefficient ^ of hydrogen in 9*88 per
cent, alcohol increased from 0*0676 at 0° ; to 0*0693 at 6*2° ; to 0*0705 at 13*4° ; and
to 0*0740 at 18*8°. L. Carius (1855) found rather lower numbers ; he represents
the coefficient of absorption j3 of hydrogen in alcohol at 6° by the expression :
^8=0-06925— 0-00014870— 0*000001^2^ when 0° lies between 0° and 25°. The
solubility S of hydrogen in some alcohols is —
Methyl alcohol
Ethyl alcohol
Iso-butyl alcohol
Amyl alcohol
20° 25°
0-0902 0*0945
20° 25°
0*0862 0-0894
20° 25°
0-0929 0-0976
20° 2^
0-0353 0-0301
The corresponding values for water are S=0*0200 (20°) and .S=0*0199 (25°). The
solubilities of hydrogen in a large number of organic compounds and m aqueous
304 INORGANIC AND THEORETICAL CHEMISTRY
solutions of organic compounds have been measured.^ In some cases it will be
observed that the solubility of the gas is greater at the higher temperature.
The solubility S of hydrogen in a 2*29 per cent, glycerol solution at 14° was found by
P. Drucker and E. Moles to be 0*01886, and in a 15-31 per cent, solution, 0-01765 ; while
at 25°, a 4-0 per cent, solution has a solubility of 0-0170 ; a 50*5 per cent, solution, 00089 ;
and a 95-0 per cent, solution, 0-0030. For a 2-63 per cent, solution of propionic acid at
20°, *Sf=0-0186 ; and for a 9-91 per cent, solution, 0-0165, which rises to 0-0209 at 5°. For
a solution of 89 grams of attiidopropionic acid, CH3.CH{NH2).COOH2, per litre, /S=0-0145
(20°) ; for a solution with 75 grms. of glycocol per litre, *S=0-0147 (20°) ; for a solution with
60-07 grams of urea per litre, 6^=0-0159 (20°) ; and for one with 59 grams of acetamide per
litre, <S=0'0167 (20°). The solubility of hydrogen in a 4-91 per cent, solution of chloral
hydrate at 20° is 0-0171, and in a 63-9 per cent, solution, 0*0122 ; in 16-67 per cent. For
sugar solutions at 15°, C. Midler found ^9^0-01479 ; and in 47-65 per cent, solutions,
^=0-008456 ; with 41*4, 80-8, and 166-62 grams oi glucose per litre, at 20°, S was respectively
0-0164, 0-0153, and 0-0141. The solubility S of hydrogen in aniline is 0-0303 (20°) and
0-0285 (25°) ; nitrobenzene, 0*0353 (20°) and 0-0371 (25°) ; toluene, 0-08384 (20°) and
0 08742 (25°) ; xylene, 0*07834 (20°) and 0*08185 (25°) ; carbon disulphide, 0*0336 (20°)
and 0*0375 (25°) ; acetone, 0*0703 (20°) and 0-0764 (25°) ; ethyl acetate, 0*0788 (20°) and
0-0852 (25°) ; amyl acetate, 0-0743 (20°) and 0-0774 (25°) ; and iso-butyl acetnie, 0-09287
(20°) and 0-09758 (25°). S. Gniewasz and A. Walfisz found the absorption coefficient of
hydrogen in petroleum at 10° is 0-0652, and at 20°, 0-0582 (water at 20°, 0-0193) ; G. Fahr
measiu-ed the solubility of hydrogen in the blood and serwn of different animals.
As a rule, the absorption coefficient of hydrogen in organic solvents decreases with
increasing concentrations of the solutions ; it is also noteworthy, that the solubiUty
of hydrogen in several organic solvents increases as the temperature rises — usually,
the solubility decreases as the temperature increases. The work of A. Christoff,
previously cited, shows that the solubility increases as the surface tension of the
solvent decreases.
The permeability of the metals to hydrogen.— In 1863, H. St. C.
Deville and L. Troost ^ showed that hydrogen gas can diffuse through red-hot
platinum or iron, but not through the cold metals, and they suggested that the
permeability was caused by the development of a kind of porosity of a greater degree
of minuteness than the porosity of graphite and earthenware ; they said that
this new porosity is entirely due to the expansive agency of heat opening up inter-
molecular spaces in the heated metal — la porosite resulte de la dilation que la chaleur
fait ejprouver aux espaces intermoleculaires . It has been estimated that about four
litres of hydrogen can pass through a square metre of palladium per minute ; and
about half a litre through a square metre of platinum in the same time. T. Graham
and many other investigators have investigated the pressure obtained by the
diffusion of hydrogen through platinum and palladium. According to A. Winkel-
mann, the quantity of hydrogen which diffuses through the metal is not proportional
to the pressure, for at low pressures the quantity which diffuses is relatively
larger than accords with this assumption. A. Winkelmann also investigated the
diffusion of hydrogen through iron and platinum ; and 0. W. Richardson, J. Nicol,
and T. Parnell also studied the diffusion of hydrogen through platinum. A. Sie verts
and P. Beckmann found hydrogen begins to diffuse through copper at 640° ;
platinum at 500° ; nickel at 450° ; iron at 300° ; and palladium at 240°. 0. W.
Richardson, J. Nicol, and T. Parnell found the speed of diffusion increased with a
rise of temperature, while the temperature coefficient at the same time decreased.
The speed of diffusion at a constant temperature is proportional to the square root
of the pressure. The diffusion is probably connected with the absorption of the gas
by the metal. The gas is absorbed on one side and given off on the other side
where the partial pressure of the hydrogen is smaller. It is suggested by A. Winkel-
mann 7 that the hydrogen is dissociated and only hydrogen in the atomic condition
can traverse the platinum. J. Schmidt raised objections to the atomic hypothesis,
but neither 0. W. Richardson nor A. Winkelmann accepted J. Schmidt's views.
A. Lessing found the velocity of diffusion of electrolytic hydrogen increases with
increasing potential. The diffusion of hydrogen through platinum has to be taken
HYDROGEN
305
into consideration in chemical analysis when certain substances are heated over the
gas flame in a platinum crucible. Reducing gases, from the coal gas, pass through
the walls of the crucible and exert a reducing action on substances being calcined—
e.g. manganese oxide.
According to G. Quincke (1877),8 hydrogen, oxygen, and nitrogen do not diffuse
through glass of lb mm. thickness at ordinary temperatures, and under a pressure
of 126 atmosphere, nor is there any sign of the permeability of glass up to its softening
temperature. M. Berthelot maintained that hydrogen can diffuse through the walls
of heated glass— ordinary and Jena. A. Sieverts and W. Krumbhaar (1910) also
found that unglazed porcelain is impervious to hydrogen at 1650°, and that it can
support a vacuum at 1400°. Glass surfaces absorb hydrogen. According to
P. Chappius, 1 sq. mm. of glass at normal pressure, on heating from 0°to 180° gives off
0-00027 CO. of hydrogen. Powdered quartz at 1100° can absorb 0*015 c.c. of hydro-
gen or helium per gram. Transparent quartz glass at ordinary temperatures and
pressures is impervious to hydrogen, oxygen, etc. ; but at 330°, the quartz is pervious
to hydrogen for pressures varying from 560 to 960 mm. No leakage was observed
with nitrogen or oxygen at pressure less than one atmosphere. Hydrogen leaked
through the tube at about 430°. At a constant temperature, the leakage increased
with increasing pressures as illustrated in Fig. 10. M. Bodenstein and F. Kranen-
disck found that 4*38x10-6 c.c. of hydro-
gen diffused through 1 sq. c.c. of quartz
glass 1 mm. thick per hour when there was
a difference of pressure of 1 atmosphere on
the two sides and the temperature was
732°; at 880°, 8-65x10-6 c.c. of gas
diffused under similar conditions.
The solubility of hydrogen in the
metals.— In 1866, the attention of T.
Graham 9 seems to have been arrested by
H. St. C. Deville and L. Troost's observa-
tion on the permeability of metals to gases.
T. Graham could not detect any signs of the
passage of oxygen, nitrogen, chlorine, steam,
hydrogen chloride, carbon dioxide, carbon
monoxide, methane, ethylene, hydrogen
sulphide, or ammonia through a septum of platinum I'l mm. thick at a full red
heat in an apparatus capable of detecting 0'2 c.c. per hour. He also found that
when certain metals are heated in hydrogen gas, more or less of the gas is
absorbed and retained as the metal cools, forming a kind of solid solution of the
gas in the metal.
T. Graham heated a small piece of the metal under investigation in a porcelain
tube glazed inside and out. The tube was heated to redness and exhausted. Hydrogen
was then allowed to pass over the heated metal, and the metal cooled in the same gas. The
tube was then evacuated, and afterwards heated, with the pump in action, until no more
gas was evolved. The gas was collected over mercury and measured. The volume and
weight of metal were also determined, and the results could be represented in any desired
way. He expressed his results as average volumes of gas absorbed per unit volume of
metal. He found :
4.3A
O
#f^
.C
/
X
^
p^"
nm*
—
(Q
o
/
-^
y
T^
P17
0)
3
//.
^
■^J
^//
V
560 mm.
Time
n Mil
luf-es.
'
0-217 -^
0 10 20 30 40 50 60 70
Fig. 10. — The Leakage of Hydrogen at
Different Pressures from Vessels of Quartz
Glass at 430".
Vols.
Vols.
Platinum (fused)
4-68
Copper (wrought)
. 0-31
Platinum (hammered)
3-03
Gold
. 0-4B
Palladiiun (foil) .
. 495-50
Silver
. 0-90
Copper (sponge)
0-60
Iron
. 415
An osmium-iridium alloy absorbed a trace of the gas. Palladium, therefore, of all
the metals, appears to possess the power of absorbing hydrogen in the highest
degree. The volume of gas absorbed by a metal depends on its condition, and
increases with the superficial area. The gas is retained by the metal very tenaciously,
VOL. I. X
306
INOKGANIC AND THEOKETICAL CHEMISTRY
and it can be recovered from the metal only by beating to redness in vacuo.
Hence, added T. Graham :
It appears necessary to recognize in palladium a new property, a power to absorb
hydrogen at a red heat, and to retain gas at a temperature under redness for an indefinite
time. It may be allowable to speak of this as a power to occlude (to shut up) hydrogen,
and the result as the occlusion of hydrogen by palladium.
The fixation of the gas was found to be more energetic if the metal under investiga-
tion were used as a negative electrode during electrolysis. For example, M. Thoma
has shown that if palladium be employed as negative electrode during the electrolysis
of acidulated water, the metal may become supersaturated with the gas and dissolve
over 935 times its volume of the gas ; the amount actually dissolved depends on the
strength of the current. The excess is quickly evolved when the current ceases.
In order to prepare by drogenized palladium, a clean piece of palladium foil about 16 sq. cm.
is used as cathode in a solution of palladium nitrate- — 2 grams of the salt per 100 c.c. of water.
The anode is also of palladium. After a current of about one ampere has passed for 20-30
miniates, the foil will be covered with a velvety-black film of palladium black. This plate
is then used as cathode in the electrolysis of dilute sulphiu-ic acid with a current of 0*8 to
rO amp. The pla]be will be saturated with hydrogen in about an hour. The foU is rapidly
washed with water, and dipped in absolute alcohol and ether. It can be preserved for
some weeks under air-free water.
G. Neumann and F. Streintz, like T. Graham, measured the volume of gas,
reduced to standard conditions, absorbed by one volume of metal. Some of
the latter's results are here included with those of the former :
Silver (foil)
Silver (powder)
Aluminium (sheet)
Cobalt (reduced)
Copper (wire)
Copper (reduced
Iron (wire)
Iron (wrought)
Iron (reduced)
Magnesium
Nickel (reduced)
Vols, absorbed
0-21
to
to
to
to
to
0-9
1-1
59-0
0-3
0-6
0-46
0-57
9-4
1-4
170
0-95
2-7
153
4-8
4-8
to 0-8
to 19-2
to 18-0
Gold (leaf)
Gold (precipitated)
Palladium (wrought)
Palladium (wire)
Palladium (sponge)
Palladium (fused)
Platinum (sponge)
Platinum (foil)
Lead (fused)
Zinc (electrolytic)
Antimony
1-4
37-0 to 46-0
3760 to 643
930-0
680-0 to 852
68-0 to 200
1-5 to 490
0-8 to 1-5
0-11 to 015
traces
nil
According to G. Neumann and F. Streintz, silver absorbs no hydrogen — but
others have reported the absorption of about one-fifth of its volume of the gas.
The numbers here given are not to be taken as absolute — perhaps not even com-
parable. Considerable differences have been reported by other observers. Thus,
according to A. Sieverts and his co-workers, hydrogen is not dissolved by cadmium.^
thallium, aluminium, zinc, lead, bismuth, tin, antimony, tungsten, silver, or gold ;
while coffer, nickel, and iron do dissolve the gas. H. R. Carveth and B. E. Curry
reported the occlusion of 250 vols, of hydrogen by electrically deposited chromium.
L. Troost and P. Hautefeuille obtained evidence of the occlusion of hydrogen by
manganese, and the presence of manganese increases the solubiHty of hydrogen in
iron ; on the other hand, E. Wedekind and T. Veit did not obtain much occlusion,
but they worked with a gas containing some oxygen. C. Winkler says that beryllium,
manqanese, yttrium, and zirconium take up hydrogen when their oxides, intimately
mixed with magnesium, are heated, but magnesium alone did not absorb the
gas. The experimental conditions do not appear to be such as would furnish
reliable evidence of the occlusion of gases by these elements. A. van Berghe and
E. L. Lederer did not find that molybdenum occluded hydrogen to any marked
degree ; A. Sieverts and his co-workers, and G. Neumann have studied the occlusion
of hydrogen by iron ; A. Sieverts and his co-workers, W. Ipatieff, and M. Mayer and
V. Altmayer have studied nickel ; A. Sieverts, cobalt and coffer. The experiments
of H. E. Roscoe, and of W. Muthmann, L. Weiss, and R. Riedelbauch make it
probable that hydrogen is occluded by vanadium ; and likewise H. von Bolton,
HYDROGEN
307
W. Muthmann, and A. Sieverts and their co-workers suggest that hydrogen is
occluded by tantalum. M. von Piriani says tantalum at a red heat absorbs 0'3 per
cent, of gas, and more at a higher temperature. There is no evidence of the
occlusion of any marked amount of hydrogen by mercury. T. Wilm said that
rhodium absorbs more hydrogen than does palladium, but E. Quenessen thought
this to be improbable, and A. Sieverts and E. Jurisch found rhodium absorbs very
little hydrogen. F. Rother, and R. Finkener and F. Fisher show that while ordinary
iridium does not occlude much hydrogen, under the stimulus of a prolonged cathodic
bombardment it can take up 800 vols, of hydrogen. According to A. Ledeber,
this, however, is doubtful evidence of occlusion. A. Gut bier and co-workers found
pure iridium occludes 140 vols, of hydrogen at 20°. C. Winkler considers it unlikely
that titanium occludes hydrogen. L. Cailletet and E. CoUardeau say that ruthenium
can take up gases during electrolysis, and T. Graham says that an osmium-iridium
alloy absorbs no hydrogen when heated in the gas. It may therefore be said that
the two elements — ruthenium and rhodium — as well as of scandium, gallium, and
indium, have not been investigated. A. Sieverts and E. Bergner obtained a small
occlusion of hydrogen with an impure form of uranium. The work of C. Winkler,
0 10 20 30 4.0 50 60|70 80
^^" 0-5/,ro6.
C 30
-t) 20
jy 10
«»
^
0 10 20 30 40 SO 60 70 80
\
\
\
0
■^
^
^
20 30 40 50
Figs. 11 to 13. — Absorption of Hydiogen by Palladium Alloys.
C. Matignon, W. Muthmann, and H. H. Zhukoff show unmistakable evidence of the
occlusion of hydrogen by ceriu7n and lanthanum, for those metals form phases of
variable composition in which the hydrogen may reach high concentrations.
C. Winkler and C. Matignon found evidence of the absorption of hydrogen
by thorium, and C. Matignon by samarium, neodymium, and praseodymium.
Nothing definite is known of the behaviour of europium, gadolinium, terbium,
holmium, erbium, thulium, and ytterbium.
T. Graham and A. J. Berry have measured the solubility of hydrogen in
palladium-gold alloys. A. J. Berry found that the decrease in the amount of occluded
hydrogen is a simple function of the proportions of palladium in the alloy, but the
occluding power of the alloy vanishes when the proportion of palladium falls below
25 per cent., roughly 0-5 gram-molecules of gold. A. Sieverts, E. Jurisch^ and
A. Metz extended these observations at different temperatures between 138° and
820° ; they also employed palladium-silver and palladium-platinum alloys. The
results with the alloys are indicated in Figs. 11-13. All proportions of platmum
diminish the solubility of hydrogen ; hydrogen is virtually insoluble m silver, but
the addition of silver to palladium raises the solubility until a maximum is reached
with 40 per cent, of silver— at 130° this alloy dissolves four times as much hydrogen
308
INORGANIC AND THEORETICAL CHEMISTRY
1600
=>I200
0
as pure palladium — the solubility diminishes with increasing proportions of silver,
and becomes zero with alloys containing over 70 per cent, of silver. The curves
for different temperatures are illustrated in Fig. 14.
Different results are obtained with different samples of metal ; this is mainly
due to differences in the purity of the metal, and to the method employed in its
preparation — e.g. cobalt in the form of ingots absorbs virtually no hydrogen, while
the metal reduced from the oxide at a low temperature absorbs relatively large
volumes of the gas, and if reduced at a high temperature much less gas is absorbed.
In a general way, the amount of gas absorbed depends upon the surface area which
the metal presents to the gas. Thus the colloidal palladium prepared with sodium
protalbate by C. Paal and C. Amberger absorbs 300 to 400 vols, of hydrogen, and
the solution prepared by C. Paal and J. Gerum absorbed 1000 to 3000 vols. The
solubility of hydrogen in the metals increases proportionally with the temperature
up to the melting point when there is an abrupt increase which again increases
proportionally with the temperature. With pal-
ladium, however, the solubility does not depend
on the temperature, and unlike the other metals
there is an abrupt decrease in solubility (nearly
one-half) as the metal melts. The dissolved gas
is mainly rejected as the metal cools, and that
which is retained can be recovered on heating in
vacuo. If the metal is heated in an atmosphere
0-25 0-50 0-75 1^0 ^^ ^^^ ^^^> morc and more hydrogen is retained
s of Hydrosen per 100 ^rms.of Inn. by the cooHug mctal the higher the temperature
Fig. 14.-The Effect of Temperature *o which it has been heated. The amount of
on the Absorption of Hydrogen by hydrogen retained by iron, containing 0'04 per
Iron. cent, of carbon, heated to different temperatures,
is indicated in Fig. 14. The amount of gas
retained by a metal also increases with the pressure. The relation between the
amount of hydrogen absorbed by the molten metal and the pressure does not
usually follow Henry's law. Thus 100 grms. of molten copper (1123°) dissolve •
Pressure {p) . . 1046 883 606 281 mm.
Absorbed hydrogen (m) 0-745 0-680 0-549 0-380 mgrm.
pljm .... 43-5 43-7 45-1 44-2
showing that the quantity of hydrogen absorbed is not proportional to the pressure
as it would be if it followed Henry's law — it is proportional to the square root of
the pressure. Palladium increases in volume during the absorption, but its general
appearance and properties — ^thermal and electrical conductivity, tenacity, etc. — are
not much altered, although a considerable amount of heat is evolved during the
absorption^^4370 calories per gram of gas.
Potassium and sodium were found by L. Troost and P. Hautefeuille to absorb
hydrogen ; H. Moissan found 126 vols, of the gas were absorbed between 200° and
400° ; and 237 vols, by sodium between 300° and 421°. C. Matignon found lithium
and thallium absorbed no hydrogen. K. A. Hofmann, 0. Ehrhart, and 0. Schneider
also found that osmium tetroxide absorbs hydrogen. F. Soddy patented the use of
calcium as an absorbent for gases in the production of high vacua. D. P. Smith
found that the elements which occlude hydrogen occupy a definite position in the
periodic table — Cap. VI— and that the capacity of a metal to occlude hydrogen is
confined to those with a strong magnetic susceptibility, for the elements which have
a specific magnetic susceptibility exceeding 0"9xlO"~^ at ordinary temperatures,
occlude relatively large proportions of hydrogen, while those with a smaller
magnetic susceptibility do not occlude hydrogen — copper, rhodium, and thorium
are possible exceptions.
It was once thought that the palladium formed a chemical compound — PdgH —
with the hydrogen, but this has not been accepted as a full explanation
HYDROGEN 309
T. Graham lo thought that the gas hydrogen, during absorption, condensed to a solid
metal which alloyed with the palladium. He gave the name hydrogenium to this
hypothetical metal in order to emphasize its supposed metallic nature. It was argued
that the metallic character of the occluded hydrogen is shown by the fact that if a
plate of palladium be charged electrolytically with hydrogen, and subsequently
immersed in a solution of copper sulphate, the metal is soon covered with a film of
metalUc copper: CuS04+2Hpaiiadium=H2S04+Cu. A similar phenomenon
occurs when the plate is immersed in a solution of salts of gold, platinum, silver, or
mercury, but not in solutions of salts of lead, iron, zinc, or magnesium. Solid
hydrogen, however, was found by J. Dewar to have rather the properties of a non-
metal than of a metal — its specific gravity, for instance, is 0'076, which is but one-
eighth that (0-6 to 0'7) calculated for the absorbed hydrogen in palladium. The
relation between the absorbed hydrogen and the metal is not perfectly clear.
Not only hydrogen, but several other gases are also absorbed by metals during
their preparation or purification in furnaces, etc. For instance, W. Heald found
that hydrogen is absorbed by many metals during their sublimation in that gas.
Hence, when metals are heated to bright redness in glazed porcelain tubes exhausted
by means of an air-pump, gases of various kinds are given off — e.g. aluminium gives
off hydrogen ; magnesium, hydrogen and carbon monoxide ; zinc and electrolytic
copper give off hydrogen, carbon dioxide and monoxide ; tin and platinum give
methane and nitrogen in addition to the gases just mentioned. L. Kahlenberg
and H. Schlundt n found that when metallic sodium and mercury react with one
another much heat is evolved, and 3*24 c.c. of hydrogen is evolved per gram of
sodium. Allowance must always be made for the presence of occluded gases in
metals, etc., which have not been heated to redness in vacuo ; otherwise wrong
inferences may be drawn. Indeed, at one time, H. Davy worked with the hypo-
thesis that the elements are compounds of hydrogen with an unknown base in
different proportions, and at first he seemed to succeed in getting relatively large
quantities of hydrogen from sulphur, selenium, and carbon, but he got no unknown
base. He soon recognized that the hydrogen he obtained was mechanically absorbed
by these elements and was not a product of the decomposition of the elements,
sulphur, selenium, and carbon. Alleged transmutations of one gas into another
have been traced to similar phenomena in vacuum tubes. Many rocks and minerals
also give off gases when heated under similar conditions, showing that they, too,
have occluded gases. The meteoric iron of Lenarto containing about 91 per cent,
of iron was reported by T. Graham (1867) to have yielded 2-86 times its volume
of occluded gas. Since, under ordinary atmospheric pressures, iron absorbs only
half its volume of gas, it was inferred that the meteorite must have come from an
atmosphere containing hydrogen under a far greater pressure than our atmosphere,
a deduction confirmed by spectroscopic observations on the dense hydrogen atmo-
spheres of the sun and fixed stars.
The absorption of hydrogen or other gases in vacuum tubes has been explained
as a result of the absorption of the gas by the disintegrated particles of the
cathode by L. Vegard,i2 S. Brodetsky and B. Hodgson, and F. Soddy and T. D.
Mackenzie ; of the occlusion of the gas in the cathode by R. Riecke ; of the chemical
action of the gas and the cathode by K. Mey ; of the chemical or mechanical
action of the gas on the anode by C. A. Skinner, B. Hodgson, and V. L. Chrisler ;
to the chemical action of the gas on the glass by R. S. Willows, who found that the
absorption was greatest with soda glass, less with lead glass, and least with Jena
glass ; of the occlusion of gas in the glass by A. A. C. Swinton ; and by S. E. Hill,
of the chemical action produced by the formation of active nitrogen.
The permeability of indiaxubber to gases.— According to T. Graham,
indiarubber absorbs about 0*0113 times its volume of hydrogen. In 1786,
J. Priestley ^3 noticed that indiarubber was permeable to gases ; and J. K. Mitchell,
in 1831, found that carbon dioxide diffused through a rubber membrane faster
than hydrogen, and hydrogen, in turn, passed through more quickly than oxygen.
310 INORGANIC AND THEORETICAL CHEMISTRY
T. Graham measured the rates of diffusion of gases through rubber, and found that
the rate at which the gases pass through rubber is not dependent on the densities
of the gases. Equal volumes of the following gases penetrate rubber in the relative
periods of time here indicated (nitrogen unity) :
Carbon dioxide.
Hydrogen.
Oxygen.
Methane.
Carbon monoxide.
Nitrogen,
13-585
5-500
2-556
2-148
1113
1-000
so that 2*556 volumes of oxygen penetrate the rubber in the same time as one volume
of nitrogen. J. Dewar found with a membrane 001 mm. thick, and at atm.
pressure, at 15°, the number of c.c. of gas which diffused per day per sq. cm. :
Air
Na
CO
He
A
02
Hz
CO2
20
1-38
1-88
3-5
2-56
40
11-2
280
No relation can be detected between the chemical composition or physical
properties of a gas and its diffusibility through rubber. The speed increases
proportionally with a rise of temperature. By plotting the logarithm of the rate
against temperature, straight lines are obtained ; these lines show a distinct break
at 0^, suggesting that water is in some way involved ; with carbon dioxide there
is also a more pronounced break at —37°. H. Kayser found that the quantity of
hydrogen which passes in unit time through a sq. cm. of surface and 1 cm. thick,
when there is a difference of pressure of one atmosphere on the two faces of the rubber,
can be represented by (0'000158^ — 0*000537^2) c.c. per minute for temperatures B
between 9° and 33°. The effect of passing air through a vessel with rubber walls, a
vacuum on the outside of the rubber, can be obtained by multiplying these numbers
by the partial pressure of the gases. With air, P. Margis (1882) found that the
mixture which diffused through the rubber walls contained 40 per cent, of oxygen ;
when this product was again passed through the apparatus a mixture containing 60
per cent, of oxygen was obtained ; a third passage gave a mixture with 80 per cent.
of oxygen, and after a fourth passage, the mixture contained 95 per cent, of oxygen.
The phenomenon appears to depend on an absorption or occlusion of the gas by the
rubber, and the subsequent evolution of the gas on the side under reduced pressure.
G. Hiifner says that he found grey vulcanized rubber absorbed no measurable amount
of hydrogen between —2° and 13° at 760 mm. G. Austerwell and J. B. L. Juhle
studied the diffusion of hydrogen through the walls of gas balloons.
According to J. Hunter,^* cocoanut charcoal absorbs 4*4 times its volume of
hydrogen reduced to 0° and 760 mm. The amount absorbed is proportional to the
pressure (temperature constant), and inversely proportional to the temperature
(pressure constant) ; and, according to H. Kayser, if f denotes the pressure, and
V the volume of hydrogen absorbed per c.c. of charcoal, i;=6*036— 1*55 log ^ at 0° ;
J. Dewar found that charcoal absorbed 4 c.c. of hydrogen, and at —185°, 135 c.c.
(reduced to n.p.t.) ; and J. L. Baerwald (1906) found that the volume of hydrogen
absorbed at different temperatures by one volume of charcoal when the volume of
gas is reduced to 0° and 760 mm.
10" 0" -10° -50° -100° -150° -185°
Hydrogen absorbed 3 4 4 9 24 76 135 vols.
Neither platinum nor palladium shows this remarkable increase in absorptive power
at low temperatures. J. Dewar used the energetic absorption of hydrogen by wood
charcoal to separate this gas from neon, helium, etc., which are absorbed to a far
smaller degree, and this is in accord with the rule that the lower the boiling point,
or the more volatile or less condensable the gas, the less is it absorbed by wood
charcoal. J. W. McBain investigated the absorption of hydrogen by charcoal
and found at the temperature of liquid air and 19 mm. pressure, the solubihty is
4*1 c.c. of gas per gram. He also assumes that the absorbed gas is in the atomic
condition. J. W. McBain further showed that the process of occlusion is of a
dual character. If the surface of charcoal is first supersaturated by a short ex-
posure to hydrogen at a high pressure, and a portion of the hydrogen be then
HYDROGEN 311
removed, there is first a rise of pressure owing to the escape of the gas condensed
|0r adsorbed on the surface, the pressure then slowly falls owing to the slow diffusion
if the gas into the interior where it is absorbed or occluded to form a soUd solution.
Te proposes to include both processes — rapid surface adsorption, and slow internal
7Sorj)tion — by the general term sorption.
References.
1 W. Henry, Phil. Trans., 93. 29, 274, 1803 ; J. Dalton, Mem. Manchester, Lit. Phil. Soc.,t.
284, 1789 ; 5. 11, 1805 ; T. de Saussure, Gilbert's Ann., 47. 163, 1814 ; C. de Marty, Ann. Chim.
Phys., (1), 61. 271, 1807 ; C. J. B. Karsten, Jahresb. Pharm., 472, 1820; N. Paul, Tech. Reposit.
6. 16, 1824; A. T. y Marti, Ann. Chim. Phys., (1), 6. 271, 1790; Jmrn. Phys., 52. 176,
1801.
2 R, W. Bunsen, Liebig's Ann., 93. 1,1855; 95. 1, 1855; Gasametrische Methoden, Braun-
schweig, 136, 1857 ; W. Timofejeflf, Zeit. phys. Chem., 6. 141, 1890 ; G. Just, ib., 37. 342, 1901 ;
G. Gefifeken, ib., 49. 257, 1904 ; K. Drucker and E. Moles, ib., 75. 405, 1910 ; G. Hiifner, ib.,
57. 511, 1906; L. W. Winkler, ib., 9. 171, 1892; 55. 344, 1906; Ber., 24. 89, 1891 ; C. Bohr
and J. Bock, Wied. Ann., 44. 318, 1891 ; C. Bohr, ib., 62. 644, 1897 ; T. E. Thorpe and J. W.
Rodger, Journ. Chem. Soc, 65. 782, 1894.
3 K. Angstrom. Wied. Ann., 15. 297, 1882 ; 33. 223, 1888 ; J. J. Mackenzie and E. L. Nichols,
ib., 3. 134, 1878 ; E. L. Nichols and A. W. Wheeler, Phil. Mag., (5), 11. 113, 1881 ; W. Ostwald,
Stoichiometrie. Leipzig, 356, 1885.
* P. Steiner, Weid. Ann., 52. 275, 1894; G. Gefifeken, Zeit. phys. Chem., 49. 257, 1906;
L. Braun, ib., 33. 721, 1900 ; W. Knopp, ib., 48. 97, 1904 ; V. Gordon, ib., 18. 14, 1895 ; G. Hufner,
ib., 59. 416, 1907 ; A. Christofif, ib., 55. 622, 1906 ; H. von Euler, Arkiv. Kem. Min. Geol., 1.
143, 1904.
s L. Carius, Liebig's Ann., 94. 131, 1855 ; G. Fahr, Journ. Physiol, 43. 417, 1912 ; K. Drucker
and E. Moles, Zeit. phys. Chem., 75. 405, 1910; G. Gefifeken, ib., 49. 257, 1904 ; G. Just, ib., 37.
359, 1901 ; V. Gordon, ib., 18. 14, 1895 ; S. Gniewasz and A. Walfisz, ib., 1. 70, 1887 ; L. Braun, ib.,
33. 721, 1900 ; W. Knopp, ib., 48. 97, 1904 ; C. MuUer, ib., 81. 483, 1912 : G. Hufner, ib.,57. 611,
1906 ; P. Steiner, Wied. Ann., 52. 275, 1894 ; 0. Lubarsch, ib., 37. 525, 1889 ; H. Henkel, Beitrdge
zur Kenntnis der physikalisch-chemischen Eigenschaften verdunnter Glycerinlosungen, Berlin, 1905.
« H. St. C. Deville and L. Troost, Compt. Rend., 56. 977, 1863; 57. 894,' 1863; H. St. C.
Deville, ib., 59. 102, 1864; L. P. CaiUetet, ib., 58. 327, 1057, 1864; 60. 344, 1865; 66.
847, 1868.
' A. Winkelmann, Ann. PhysiL, (4), 6. 104, 1901 ; (4), 8. 388, 1902 ; (4), 16. 773, 1905 ;
(4), 17. 589, 1905; (4), 19. 1045, 1906; A. Lessing, Ber., 40. 569, 1907; S. W. S.
Schmidt, ib., (4), 13. 747, 1904; E. Dorn, Phys. Zeit., 7. 312, 1906; G. Charpy and
S. Bonnerot, Compt. Rend., 154. 592, 1912 ; 156. 394, 1913 ; 0. W. Richardson, J. Nicol,
and T. Parnell, Phil. Mag., (6), 7. 266, '904 ; W. Ramsay, ib., (5), 38. 206, 1894 ; T. Graham,
ib., (4), 32. 401, 503, 1866; Proc. Roy. Soc, 15. 502, 1866; 17. 212, 500, 1869; 0. W.
Richardson, Proc. Cambridge Phil. Soc., 13. 27, 1905 ; A. Sieverts and P. Beckmann, Zeit. phys.
Chem., 60. 129, 1907 ; W. W. RandaU, Amer. Chem. Journ., 19. 682, 1897 ; T>. Tsakalotos, Proc.
Chem. Soc, 24. 208, 1908.
8 G. Quincke, Pogg. Ann., 160. 118, 1877 ; A. Bartoli, Gazz. Chim. Ital, 14. 544, 1885 ; E. C.
Mayer, Phys. Rev., (2), 6. 283, 1915; A. Sieverts and W. Krumbhaar, Ber., 43. 893, 1910;
P. Chappius, Wied. Ann., 8. 671, 1870 ; P. Villard, Cmnpt. Rend., 130. 1752, 1900 ; G. Belloc, ib.,
140. 1253, 1905; M. Berthelot, ib., 140. 817, 1159, 1286, 1905; A. Jaquerod and F. L. Perrot,
Archiv. Science Geneve, 18. 613, 1904; 20. 454, 1905; Compt. Rend., 139. 789, 1904; 0. W.
Richardson and R. C. Ditto, Phil. Mag., (6), 22. 704, 1911 ; E. C. Mayer, Phys. Rev., (2), 6.
283, 1915 ; M. Bodenstein and F. Kranendisck, Nernsfs Festschrift, 99, 1902.
9 T. Graham, Phil. Mag., (4), 32. 401, 503, 1866 ; Proc. Roy. Soc, 15. 505, 1866 ; 16. 422,
1868 ; 17. 212, 500, 1869 ; Journ. Chem. Soc, 22. 419, 1869 ; G. Neumann and F. Stremtz,
Monatsh., 12. 642, 1891 ; Wied. Ann., 46. 431, 1892 ; A. Sieverts and W. Krumbhaar,
Ber., 43. 893, 1910; Zeit. phys. Chem., 74. 277, 1910; A. Sieverts and P. Beckmann, ib., 60.
129, 1907 ; A. Sieverts and J. Hagenacker, ib., 68. 115, 1909 ; Ber., 42. 338, 1909 ; M. Mayer
and V. Altmayer, ib., 41. 3062, 1908 ; G. Kriiss and L. F. Nilson, ib., 20. 1691, 1887 ; E. Wede-
kind and T. Veit, ib., 41. 3771, 1908; C. Winkler, ib., 24. 873, 1966, 1891 ; J. C. Poggendorf,
Pogg. Ann., 136. 483, 1869 ; K. Lisenko, Ber., 5. 29, 1872 ; M. Berthelot, Compt. Rend., 94.
1377, 1882; Ann. Chim. P/iy^., (5), 30. 519, 1883 ; A. Berliner, Wied. Anyi., 35. 791, 1888:
M. Thoma, Zeit. phys. Chem., 3. 69, 1889 ; L. Mond, W. Ramsay, and J. Shields, Proc. Roy. Soc,
62. 290, 1898; C. Paal and C. Amberger, Ber., 38. 301, 1394, 1905; 43. 243, 1910 ; C.
Amberger and J. Gerum, ib., 41. 808, 1908; C. Amberger, 6., 38. 1398, 1905;^. V^ohler,
ib., 9. 1713, 1876; H. Schifif, ib., 18. 1727, 1885; T. Wihn, ib., 25. 217, 1892; W Hempel,
ib., 12. 636, 1879; A. de Hemptienne, Bull. Acad. Bdgique, (3), 36. 155, 1898; L. Smith,
Amer. Chem., 5. 213, 1875; E. Root, Sitzber. Akad. Berlin, 217, 18/6; \\. Beetz,
Wied. Ann., 5. 1, 1878; W. Ipatieff, Journ. prakt. Chem., (2), 77. 513, 1908; W. Muthmann,
L. Weiss, and R. Riedelbauch, Liebig's Ann., 355, 91, 1907; W. Muthmann, K. Kraft,
312 INOKGANIC AND THEORETICAL CHEMISTRY
and E. Bauer, ib., 325. 263, 281, 1902; H. E. Roscoe, Phil. Trans., 159. 691,
1869 ; H. R. Carveth and B. E. Curry, Journ. Phys. Chem., 9. 364, 1905 ; D. P. Smith, ih., 23.
186, 1919 ; A. van den Berghe, Zeit/anorg. Chem., 11. 397, 1896 ; A. Sieverts, E. Jurisch, and
A. Metz, ib., 92. 329, 1915 ; G. Tammann, ib., 107. 89, 1919 ; E. Jurisch, Studien ueber die Loslich-
keit von Gasen in festen Metallen und Legierung, Leipzig, 1912 ; H. H. Zhukoff, Journ. Eussian
Phys. Chem. Soc, 45. 2073, 1911 ; F. Rother, Verh. Sachs. Ge.s. Wiss., 64. 5, 1912 ; R. Finkener
and F. Fischer, Ber. Internal. Cong. App. Chem., 2. 30, 1903 ; M. Thoma, Zeit. phys. Chem., 3.
69, 1889 ; C. Hoitsema, ib., 17. 1, 1895 ; G. Wolff, ib., 87. 575, 1914 ; A. Holt, E. C. Edgar,
and J. B. Firth, ib., 82. 513, 1913 ; ib., 83. 507, 1913 ; F. Halla, ib., 86. 496, 1914 ; A. Tchirikoff,
Btdl. Soc. Chim., (2), 38. 171, 1882 ; W. Nernst and A. Leasing, Nachr. Gottingen, 146, 1902 ;
P. A. Favre, Compt. Rend., 77. 649, 1873 ; A Moutier, ib., 79. 1224, 1874 ; L. Troost and
P. Hautefeuille, ib., 78. 686, 1874; M. von Pirani and A. R. Mayer, Zeit. Elektrochem., 16. 444,
1910 ; J. Dewar, Arch. Sciences Genkve, 50. 207, 1874 ; F. Mohr, Ber., 4. 239, 1871 ; H. Moissan,
Bev. Gin. Chim. Pure Appl, (5), 6. 48, 277, 1903 ; A. Gutbier, H. Gebhart, and B. Ottenstein,
Ber., 46. 1453, 1913 ; J. H. Andrew and A. Holt, Proc. Boy. Soc, 89. A, 170, 1913 ; A. Holt,
ib., 90. A, 226, 1914 ; 91. A, 148, 1915 ; Proc. Chem. Soc, 19. 187, 1903 ; K. A. Hofmann,
O. Ehrhart, and 0. Schneider, Ber., 46. 1657, 1913 , A. Sieverts and E. Jurisch, ib., 45. 221, 1912 ;
T. Wilm, ib., 14. 629, 1881 ; A. J. Berry, Journ. Chem. Soc, 99. 463, 1911 ; A. E. Freemann,
Journ. Amer. Chem. Soc, 35. 927, 1913 ; J. Eggert, Zeit Elektrochem., 20. 370, 1914 ; L. Quenessen,
Compt. Bend., 139. 795, i904 ; C. Matignon, ib., 131. 891, 1900 ; L. Cailletet and E. Colardeau,
ib., 119. 833, 1894 ; L. Troost and P. HautefeuiUe, ib., 76. 482, 562, 1873 ; 78. 807, 968, 1874 ;
J. Moutier, ib., 78. 1242, 1874; E. Hughes, Chem. Ztg., 3. 38, 1880; T. W. Richards and C. E.
Behr, Zeit. phys. Chem., 58. 301, 1907 ; E. Heyn, ib., 58. 760, 1907 ; H. Wedding and T. Fischer,
Stahl Eisen, 23. 1268, 1903 ; G. Neumann, ib., 34, 252, 1914 ; A. Sieverts, J. Hagenacker, and
W. Krumbhaar, ib., 29. 1249, 1909 ; A. Ledebur, ih., 7. 681, 1887 ; R. Bottger, Dingier' s Journ.,
201. 80, 1871 ; Journ. prakt. Chem., (2), 9. 193, 1874; J. L. W. Thudichum and H. Hake, Journ.
Chem. Soc, 30. 251, 1876 ; 0. Litzenmayer, Ber., 11. 306, 1878 ; G. S. Johnsen, Chem, News, 37.
271, 1878; A. Ijeduc, Compt. Bend., 135. 1332, 1902; ib., 136. 1254, 1903; A. Loret, ib., 107.
733, 1888; H. Caron, ib., 63. 1129, 1866; A. Sieverts and W. Krumbhaar, Zeit. phys. Chem.,
74. 277, 1910; A. Sieverts, ib., 60. 129, 1907; A Sieverts and J. Hagenacker, Wallach-
Festschrift, 631, 1909; Zeit. phys. Chem., 68. 115, 1909; A. Colson, Compt. Bend., 130. 330,
1900 ; J. Trowbridge, Amer. Journ. Science, (4), 27. 245, 1909 ; J. Shields, Chem.. News, 65.
195, 1892; A. Sieverts and E. Bergner, ib., 44. 2394, 1911; A. Sieverts, Zeit. phys.
Chem., 60. 129, 1907 ; 77. 591, 1911 ; 88. 103, 451, 1914 ; 74. 277, 1910 ; Zeit. Elektrochem.,
16. 707, 1910 ; H. von Bolton, ib., 11. 50, 1905 ; 13. 148, 1907 ; W. Heald, Phys. Zeit., 8. 659,
1907 ; G. Gehlhoff, Verh. deut. phys. Ges., 13. 271, 1911 ; R. H. de Forcrand, Compt. Bend., 140.
990,1905; C. Matignon, i6., 131. 891, 1900; M. von Pirani, Zeit. Elektrochem., 11. 555, 1905;
K. A. Hofmann, 0. Ehrhart, and O. Schneider, Ber., 46. 1657, 1913 ; P. Beckmann. Ann.
Physik, (4), 46. 481, 1915; H. Koch, ib., (4), 54. 1, 1917; F. Soddy, Proc Boy. Soc, 78. A,
429, 1906; German Pat. D.B.P., 179526, 1906; A. Gutbier, B. Ottenstein, and G. L. Weise,
Ber., 52. 1366, 1919; A. Gutbier and O. Maisch, ib., 52. 1368-, 1919; E. B. Maxted, Jowm.
Chem. Soc, 115. 1050, 1919.
i» T. Graham, Proc. Roy. Soc, 17. 212, 500, 1869 ; 0. Low, Journ. prakt. Chem., (2), 1. 307,
1870 ; C. A. Seely, Chem. News, 21. 265, 1870 ; J. Dewar, ib., 84. 281, 293, 1901.
" L. Kahlenberg and H. Schlundt, Journ. Phys. Chem., 9. 257, 1905 ; W. Heald, Phys. Bev.,
24. 269, 1907.
12 L. Vegard, Phil. Mag., (6), 18, 465, 1909 ; (6), 32. 239, 1916 ; S. Brodetsky and B. Hodgson,
ib., (6), 31. 478, 1916 ; (6), 32. 239, 1916 ; R. S. Willows, ib., (6), 1. 503, 1901 ; C. A. Skinner,
ib., (6), 12. 481, 1906 ; Phys. Bev., (1), 21. 1, 169, 1905 ; Phys. Zeit., 6. 610, 1905 ; F. Soddy and
T. D. Mackenzie, Proc Boy. Soc, 80. A, 92, 1908; A. A. C. Swinton, ib., 79. A, 134, 1907; 81. A,
453, 1908; K. Mey, Ann. Physik., (4), 11. 127, 1903; R. Riecke, ib., (4), 15. 1003, 1904; Wied.
Ann., 3. 414, 1899 ; B. Hodgson, Phys. Zeit., 13. 595, 1912 ; V. L. Chrisler, ib., 10. 745, 1909 ;
S. E. Hill, Proc, Phys. Soc, 25. 35, 1912; B. Moore and J. W. MeUor, Trans. Cer. Soc,
7. 1, 1908.
1' J. Priestley, Experiments and Observations relating to the Different Branches of Natural
Philosophy, Birmingham, 1786 ; J. K. Mitchell, Journ. Boy. Inst., 2. 101, 307, 1831 ; T. Graham
Phil. Trans., 156. 399, 1866 ; S. von Wroblewsky, Pogg. Ann., 158. 539, 1876 ; Wied. Ann., 2
481, 1877 ; 8. 29, 1879 ; G. Hiifner, ib., 34. 1, 1888 ; H. Kayser, ib., 43. 544, 1891 ; J. Stefan
Sitzher. Akad. Wien, 77. 371, 1878; M. Peyron, Compt. Rend., 13. 820, 1841 ; P. Margis, Deut. Ind
Ztg., 23. 314, 1882 ; A. Revchler, Bull. Soc Chim., (3), 9. A, 404, 1893 ; Lord Rayleigh, Phil
Mag., (5), 49. 220, 1900; L. Grunmach, Phys. Zeit., 6. 795, 1905; G. Austerwell, Co7npt. Bend,
154. 196, 1902; J. B. L. Juhle, ib., 154. 423. 1902; J. Dewar, Proc Boy. Inst., 21. 543,
813, 1918.
1^ J. Hunter, Journ. Chem. Soc, (2), 10. 649, 1872 ; H. Kayser, Wied. Ami., 12. 526, 1881
J. Dewar, Chem. News, 97. 16, 1908 ; Compt. Bend., 139. 261, 1904 ; L. Joulin, ib., 90. 741, 1880
J. B. Firth, ib., 158. 121, 1914 ; G. Claude, ib., 158. 861, 1914 ; J. L. Baerwald, Ueber die Absorp
tion von Gasen durch Holzkohle bei tiefen Temperaturen, Freiburg, 1906; A. Titon, Zeit. phys
Chem., 74. 641,1910; M. J. Burgess and R. V. Wheeler, Journ. Chem. Soc, 99. 649, 1911
J. W. McBain, Phil. Mag., {6), 18. 916, 1909.
HYDEOGEN 313
§ 7. The Physical Properties o£ Hydrogen
Hydrogen was once used as the standard for the atomic weights because it is
the lightest element known. It is so much lighter than air that it escapes very
quickly from a jar with its mouth upwards, and slowly from a vessel with its mouth
downwards. For the weight of a litre of hydrogen at 0° and 760 mm. the
data by H. Cavendish correspond with 0'092 grm. ; A. L. Lavoisier, 0*0769 grm. ;
T. Thomson, 00693 grm. ; J. B. Biot and F. J. Arago, 0*0732 grm. ; P. L.
Dulong and J. J. Berzelius, 0-0688 grm. ; and J. B. A. Dumas and J. B. J. D.
Boussingault, 0*0695 grm. These determinations are merely of historical interest.
Later and more accurate determinations by H. V. Eegnault gave 0*0896 at Paris,
and when this value was corrected for the difference in volume between an exhausted
and full globe, J. M. Crafts found that H. V. Regnault's value should be 0*08988 at
Paris. J. P. Cooke's value at 45° and sea level is 0*089864 ; A. Leduc's, 0*08982 ;
Lord Kayleigh's, 0*089979 ; E. W. Morley's, 0*089873 ; and J. Thomsen's, 0*089947.
The numbers i thus range from 0*0896 to 0*090032 for the weight of a litre of
hydrogen at 0° and 760 mm. pressure, at latitude 45° and sea level — the best repre-
sentative value is taken to be 0*08985 grm. For the density of hydrogen (air
unity), H. V. Regnault gave 0*0692 grm. (air unity), and with J. M. Craft's correction,
0*06949. J. P. Cooke gave 0*06958 ; A. Leduc, 006947 ; Lord Rayleigh, 0*06960.
The data for the density of hydrogen, air unity, thus vary from 0*06927 to 0*06960,
and the best representative value is taken to be 0*0694 (air unity). Hydrogen is
14*37, say 14J, times lighter than air ; 11,160 times lighter than water ; and
151,700 times lighter than mercury. According to J. Dewar, the density of hydro-
gen at its boiling point, —252*5°, is 0*55, air unity ; and according to V. Meyer,
the density does not alter at high temperatures. Liquid hydrogen has a specific
gravity about J^th that of water — and the variation of the specific gravity with
the absolute temperature T, according to J. Dewar, is such that the specific gravity
at T is 0*04136-0*000247T ; it is 0*0700 at —252*5° ; 0*0754 at —258*3° ; and
0*0763 at —259*9°. According to H. K. Onnes and C. A. CrommeUn, soUd hydrogen
had a specific gravity of 00763 at —259*9°, and 0*08077 at —262°. The cfOntraction
on freezing is about 4*8 per cent, of the liquid volume. The atomic volume of
liquid hydrogen is therefore 14*3 ; and according to J. Dewar, the molecular volume
of hydrogen at absolute zero is 24*18 (extrapolation).
The weight of a hydrogen atom was estimated by R. D. Kleeman 2 to be
1*56 X 10-24 grm., and J. Perrin estimated the mass of the hydrogen atom to be
1*4x10-24 grm. The mean diameter of the molecule of hydrogen is 2*68x10-8
cm. ; the volume of the molecule, 10-25 c.c. ; the mean free path of the
molecule is 18*3x10-6 to 17*8 xlO-^ cm. ; the number of molecules per c.c is
2*75x1019; the collision frequency is 92*8x108, or 1*64x1029 per c.c. per second;
and the molecular velocity 169,400 cm. per second. The value of J. D. van der
Waals' a=0*00042, and of his &=0-00088, and J. J. van Laar has discussed the
variability of these magnitudes.
PubUshed data ^ for the viscosity of hydrogen gas at 60°, vary from
82*2x10-6 to 85*74x10-6 C. G. S. units— the mean may be taken as the best
representative value. According to P. Breitenbach, at 15° the viscosity is 88*9
XlO-6; at 99*2°, 182*4x10-6; and at 302°, 139*2x10-6; K. L. Yen gives
7^=0000088216 with an accuracy of 0*15 per cent, at 23° and 760 nmi. The
viscosity at a temperature 23'' is 0*00008821610*15 per cent. According to
W. Kopsch, the viscosity decreases from 83*7x10-6 at —0*1° to 80*2x10-6 at
—17*8°; to 71*0x10-6 at —60*2°; to 37*42x10-6 at -194*9°. According to
J. E. Verschaffelt, the viscosity of the saturated vapour of hydrogen at 20*4° K.
and 769 mm. pressure is 0*000010 ; and of the liquid, 0*000130. The coefficient C
in W. Sutherland's equation for the relation of viscosity with temperature
7y=7yo{(273+r)/(T+C)i(T/273)^, has the value 127; Lord Rayleigh gave 128*2;
H. Markowsky, 138 ; and P. Kleint, 136. A gas experiences frictional effects not
314 INOKGANIC AND THEORETICAL CHEMISTRY
only when two of its layers flow past one another with different speeds — internal
friction or viscosity — but also when it streams along the surface of a fixed body
or of a body which moves with it — external friction. The gas does not adhere
firmly to the solid, but slips along it. If y be the coefficient of slip, r] the viscosity,
and € the external friction, y=rj/€. For hydrogen, A. Kundt and E. Warburg,
and 0. E. Meyer, give y=0'0000186. Liquid hydrogen has a surface tension j^th
that of water, or 0*2 of that of liquid air. According to P. L. Dulong,* the
velocity of sound in hydrogen gas is 1269*5 metres per second at 0°, and according
to I. B. Zoch, 1286-362 metres per second at 0°.
The value of the product pv for hydrogen has beenstudied by S. von Wroblewsky ^
up to 70 atm. ; by E. H. Amagat up to 3000 atm. ; and by W. J. de Haas up to
30,000 atm. pressure. The results show that the volume of the gas at high pressures
is greater than is indicated by Boyle's law. Arbitrarily assuming that the product
pv is unity at 0° and one atm. pressure, then.
Pressure .
1
500
1000
1500
2000
2500
2800 atm,
Volume .
1-00000
0-002713
0-001725
0-001380
0-001194
0-001078
0-001024
pv .
1-0000
1-3565
1-7250
2-0700
2-3890
2-6950
2-8686
According to H. K. Onnes and H. H. F. Hyndmann the compressibiUty of hydrogen
at 20° is given by equation ^v=l-07258+0-000667/t^+0-00000099/?;2. A. Jaquerod
and 0. Scheuer give the compressibility —(d{pv)ldp)lpv between 400° and 800° as
—0-00052 ; A. Leduc gives —0-00064 ; D. Berthelot, —0-00060 ; P. Chappius,
-0-00058 ; and Lord Rayleigh, -0-00053. S. von Wroblewsky and W. J. de
Haas found the compressibility of hydrogen does not follow Boyle's law at high
temperatures. It falls from 0-000408 at 1000 atm. to 0-000158 at 30,000 atm.
pressure. According to P. A. Guye, the value of d{pv)lpv.dv. is +0'00052 from 0 up
to 1 atm. pressure, and 0-00069 from 40 to 70 cm. pressure — temperature 0°. Accord-
ing to L. Cailletet, the compressibility at 15° falls regularly between 60 and 505 atm. ;
while E. H. Amagat noted that it follows Boyle's law up to 250°, at 3 to 6 mm.
pressure ; J. A. Siljestrom found its elasticity higher than is required by Boyle's law,
only at pressures below one atm. ; but at very low pressures, F. Fucks, E. Budde,
and C. Puschl obtained negUgibly small deviations. These results were confirmed
by Lord Rayleigh, who found that the product pv at low pressures, 3 to 6 mm., is
in conformity with Boyle's law for hydrogen.
The effect of variations of temperature and pressure on the coefficient of
thermal expansion of hydrogen is, for small pressures.
Pressure .
. 0-0077
0-025
0-47
093
11-2
76-4
100 mm.
Temperature
. 16°-132°
15°-132°
12°-185°
—
— .
— .
0°-100°
ax 102 .
. 0-3328
0-3623
0-3656
0-37002
0-36548
0-36504
0-36626
According to P. von Jolly,^ the coefficient of thermal expansion (pressure
constant) is a=0-0036562 +0-0000010001 ^ ; and according to M. W. Travers,
G. Senter, and A. Jaquerod, the pressure coefficient (temperature constant) is
j3=0-00366255, or very nearly 1/273-03 per degree. A. Leduc gives a=0-0O3662,
and j3=0003664 from 0° to 100°. A. W. Witkowsky has calculated the coefficient
of thermal expansion (pressure constant) from —212° up to 100°. Hydrogen has
a thermal conductivity about seven times larger than air.7 The great heat
conductivity of hydrogen was noticed by J. Priestley as early as 1781 ; and F. C.
Achard in 1783. According to N. N. Beketoff, the conductivity expressed in terms
of the number of calories transmitted per second through a layer 1 cm. thick
per sq. cm. of surface when the difference of temperature at the two sides is 1°, is
0000327 (0°;, 0-0003693 (100°), and at 6°, the conductivity is 0*000327 (1+0-00175^) ;
S. Weber gives 0*0004165. L. Graetz's values for the heat conductivity of hydrogen
at 0° and at 100° are respectively 0-0003190 and 0-0003693. The increase of the con-
ductivity with temperature 6° is represented by 0-0003190 (1+0-0060). According
to J. Janssen, the ratio of the cooling velocity of hydrogen to that of air is 1000 : 7-459
HYDROGEN 315
—the calculated value is 1:7:1. According to P. A. Eckerlein, the thermal con-
ductivity falls from 0-0003186 at 0°, to 00002393 at —59°, and to 00001 175 at -150°.
The specific heat of hydrogens at constant volume, Ct,, is 2*4: when referred
to an equal weight of water, and 099 referred to an equal volume of air. The
specific heat at constant pressure, Cp, is 0-2438 between —28° and 9° ; and the
molecular heat Cj,, at 16° is 3-403 ; at —76°, 3-157 ; and at —181°, 2-644 ; from 20°
to 50°, 3-4212 ; and from 20° to 100°, 34226. The molecular heat, Cp, rises from
3-402 at atmospheric pressure to 3*788 at 30 atm. pressure. A. Eucken gives
0^=3-20 for hydrogen at a concentration of 2*67 gram- molecules per litre at
35° K., and C^,=3-14 at 45° K. ; for hydrogen at a concentration 223 gram-
molecules per litre Oj,=3-32 at 35° K., and 3-28 at 45° K. G. Vieille estimates the
molecular heat at constant volume and ordinary temperatures as 4*8 ; at
3100°, 6-30 ; at 3600°, 7-30 ; and at 4400°, 8-10. According to W. H. Keesom and
H. K. Onnes (1918), the atomic heat of liquid hydrogen at 14-82° K. is 1-75 and 2-26
at 20-11° K. ; the atomic heat of the solid at 12-55° K. is 0-64. A. Eucken gives
for the molecular heat of liquid hydrogen Oj,=3'95 at 17-4° K., and 4-70 at 21-3° K.
For temperatures between 11° and 95°, and pressures jp up to 34 atm., S. Lussana
gives for the molecular heat C2,=3-4025+0013300(j9— 1). W. Nernst and H. von
Wartenburg give 0^=4-68+0-00026^, where T denotes the absolute temperature.
M. Pier's value for Cp between 0° and 2350° is 0^=4 '700 +0-0004^ ; and G. N.
Lewis and M. Randall's value, Cp=6-50+0-0009jr, for hydrogen molecules.
According to A. Eucken, the specific heat of diatomic hydrogen below 60°
K. is the same as for monatomic gases, viz. 2-98. According to R. Clausius,
the ratio of the two specific heats is 13852 ; 0. Lummer and E. Pringsheim
give 1-4084; K. Scheel and W. Heuse, 1-407; J. Jamin and F. Richard, 1-41 ;
M. C. Shields gives 1-4018 at 18° ; and W. C. Rontgen, 1-3852. According to
A. Eucken, the ratio of the two specific heats at —180° is 1-604 ; according to
K. Scheel and W. Heuse, 1-595 ; and according to M. C. Shields, 1-592 at -191°. This
makes it appear as if the hydrogen molecule entirely loses its two degrees of rotational
freedom at low temperatures. According to J. Dewar, the atomic heat of liquid
hydrogen is 6-4, and this is higher than that of any other known liquid. The atomic
heat is therefore in conformity with Dulong and Petit's rule. N. N. Beketofi gives
the atomic heat of hydrogen absorbed in palladium as 5*88. G. N. Lewis and
G. E. Gibson estimate the entropy of hydrogen gas at 25° to be 29*4 per gram-
molecule, when the increase of entropy from absolute zero to the melting point T
is (f)^jCpd log r=0-5, from the solid to the liquid at the melting point 32/15=2-13 ;
from the melting point to the boiling point 1-22 ; from the liquid to the gas at the
boiling point, 218/20-5=10-73 ; and of the gas from its boiling point to 298° K., 14-80.
In 1877, L. P. Cailletet noticed the formation of a mist when hydrogen at a
pressure of about 280 atm. is suddenly released; in 1884, S. vonWroblewsky obtained
signs of liquefaction when hydrogen at a pressure of 190 atm. and cooled by boUing
nitrogen is suddenly relieved. In 1884-5, K. Olszewsky reported that he obtained
colourless drops of liquid hydrogen by a similar process. J. Dewar (1895) first
obtained sufficient liquid hydrogen to show a definite meniscus by applying the regene-
rative process to the gas cooled to —205°. Just below the critical temperature,
—241°, a pressure of about 15 atm. will liquefy the gas ; above the critical tem-
perature no pressure, however great, will liquefy the gas. The critical pressure
is 20 atm., and the critical volume 0-00264. Liquid hydrogen is clear and
colourless, thus resemblint^ water ; it has a sharp meniscus, and a high refractive
index and dispersion. Its boiling point is -252-77°, or 205° K. ; P. G. Cath and
H. K. Onnes 9 give 20-39° K., and they found for the vapour pressure of hquid
hydrogen, at T°K., between 24*59° K. and 32-93° K., T log ^= -56-605 +3-8015^
— 0*10458 jr2-|-0-003321T3-000005102r4 atm.; while, according to M.W.Travers,G.
Senter, and A. Jaquerod, the vapour pressure is, on the absolute scale of temperatures^
Temp, (abs.) . 2041° 19-93° 1941° 18-82° 18-15° 17-36° 19-37° 1493°
Vapour pres. . 800 700 600 500 400 300 200 100 mm.
316 INORGANIC AND THEORETICAL CHEMISTRY
Hydrogen solidifies to a transparent mass like ice with a foamy surface, when the
liquid is evaporated rapidly in a partial vacuum. The unique temperature, called
the triple point, where the liquid, soHd, and vapour are all in equihbrium, is near
— 259° and 55 mm. pressure. The white sohd is crystalUne and, according to
J. Dewar, its melting point is —2592^. The data concerning the change of state
of hydrogen can be symbolized :
-259-2° --252-77»
Hydrogen ^j.^ - Hydrogen^^^^ ^ Hydrogen^^^
W. Wahl foimd the velocity of crystalUzation of hydrogen to be very great at
about 20° K. He obtained isotropic crystals belonging to the cubic system — either
trisoctahedrons or hexoctahedrons. Needle-like branches grow at right angles
to a fully developed crystal face. The latent heat of vaporization of hquid
hydrogen is nearly 218 cals. at its boiling point. A. Eucken gives 229 cals. ; W. H.
Keesom, 222 cals. ; and W. H. Keesom and H. K. Onnes, 212 cals. J. Dewar's
value for the latent heat of fusion of the solid is 15 to 16 cals.^ per gram-atom.
M. CrouUebois 10 gives 1-000137 as the mean index of refraction of hydrogen
for white Ught. J. Koch gives for light of wave-lengths X=2S0- 2 fifju, 1-0001594 ;
A=354-4/>t/x, 1-0001449 ; X=4t3b' 8 fjufju, 10001488 ; A=546-l/xjLt, 1-0001397 ;
A=670-8/x/x, 1-0001385 ; and for the ultra-red rays A=6709-4jLt)Lt, 1-0001361 ; and
A=8678-4/x/>t, 1-00013611 at 0° and 760 mm. Analogous observations have been
made by E. Ketteler, L. Lorenz, E. Mascart, E. Perreau, K. Scheel, and by C. and M.
Cuthbertson. J. W. Briihl, H. Landolt, and F. Eisenlohr have calculated the
atomic refraction by Gladstone and Dale's formula, and J. H. Gladstone gives
for the A line 1-29 ; J. W. Briihl and J. Traube, and H. Landolt, by Lorenz and
Lorentz's formula, give 1-02 for the ^-line. The dispersive power of hydrogen from
M. Croullebois' datum is 0-1814 (air unity) ; while W. Ramsay and M. W. Travers
give 0-4733. K. Hermann gives for the relative dispersion V~^=(F—C)I{ijlD-^),
when F=65-9. L. Natanson, R. Ladenburg, and S. Loria have also studied the
dispersion of hydrogen. J. W. Briihl gives for the atomic dispersion of hydrogen
Ry~Ra=0'036. C. and M. Cuthbertson find that b of Cauchy's dispersion formula,
/A— l=«(l+6/A2), rises gradually between the red and violet. Better results are
obtained with fji—l=CI{no^—n^), where the constants 0=1-692x1027, and % is
12409 Xl027_and for oxygen it is 3-397x1027, and for nitrogen 5*0345x1027.
Hence, these three elements agree with P. Drude's rule 0/1;= constant, when v
denotes the positive valency of the atom.
The magneto-optic rotation of a body refers to the angle through which a
ray of polarized light is rotated when the light is passed through the body in a
direction parallel to the lines of magnetic force. This phenomenon with glass was
discovered by M. Faraday 11 in 1845 ; it was afterwards noticed that the amount of
rotation depends on the nature of the substance, on its physical condition, on the
strength of the magnetic field, and on the wave-length of the polarized light.
A. Kundt and W. C. Rontgen noticed the phenomenon with gases in 1879. The
rotation per cm. per unit magnetic field is called E. Verdet's constant, for
E. Verdet showed, in 1853, that with the same medium and magnet, the
rotation is directly proportional to the intensity of the magnetic field. The
product of E. Verdet's constant with the molecular volume — or M/D — ^is called
the molecular rotation. The molecular rotations of many substances have
been found to be additive, for they are the sum of constants for the constituents of
the molecules. For a pressure 85 kgrms. per sq. cm., and 9*5°, and light of wave-
length A, E. Verdet's constant 12 is 0-00007585A-i-f 0-00002295 A" 3 for values of
A between Oi23fjL and 0-684/i..
If a continuous discharge be passed through a Geissler's tube containing hydrogen
at a pressure of 0-05 to 3*00 mm., there is a white glow in the capillary, and strata,
alternately pale pink and pale blue, appear about the electrodes. The white glow,
in the spectroscope, appears as a multitude of lines of varying intensity, and hence
HYDKOGEN 317
is called the white spectrum of hydrogen. If a condenser of large capacity is intro-
duced into the circuit, the oscillatory discharge changes the colour of the glow
from white to deep red. Most of the spectral lines are obHterated and the so-called
four-line spectrum or the red spectrmn of hydrogen is obtained which is comparable
with the blue spectrum of argon. The four-line spectrum of hydrogen i^ is
conveniently observed in a Geissler's tube, with the gas at about a millimetre
pressure, and through which a discharge from an induction coil is passing ; the
four lines, shown in Fig. 15, correspond with H^ in the red with a wave-length
6564-97 ; H^ in the greenish-blue 4862-93 ; Hy in the indigo-blue 4341-90 ; and
H^ in the violet 4103-10 ; these lines correspond respectively with Fraunhofer's
dark lines, C, F. G, and h. According to L. Janicki (1906), the red line is really a
double one. If the oscillatory discharge is suitably damped, the obHterated lines
preceding the four-lines spectrum gradually reappear. F. Emich considers it
possible to detect up to 7x10"^^ milligram of hydrogen by means of its spectrum
in a vacuum tube. H. W. Vogel and A. Paalzoff have photographed a great many
other lines chiefly in the violet and ultra-violet spectrum of hydrogen, and com-
pared them with analogous lines in the spectra of the sun and stars. A. WiiUner
and J. Pliicker and W. Hittorf showed that hydrogen furnishes a second and yet a
third spectrum, which A. Schuster and A. V. Angstrom ascribed to the presence of
acetylene or sulphur. G. Salet also failed to verify the observation with pure
hydrogen. Several observers, however, have verified the existence of the white
spectrum in which the usual four hydrogen lines are present though not specially
prominent. According to J. Trowbridge and T. W. Kichards, hydrogen con-
6564-97 4862-93 434.1-90 4103-0
[liiijmmiiii^^
20 30 40 50 60 70 80 90 100 UO 120 130 140 150 160
Fig. 15.— The Four-line Spectrum of Hydrogen in a Geissler's Tube.
taining a trace of water vapour readily gives the red four-line spectrum with a
continuous discharge, while the perfectly dry gas gives the white spectrum, and a
very great strength of current is then required to produce the red four-line spectrum.
B. Hasselberg postulates that the difference in the hydrogen spectra are due to
dissociation. He says :
As the explanation of the displacement of a spectrum by a new one with rise of tempera-
ture, and the first spectrum (of hydrogen) must be ascribed to a more complicated arrange-
ment of molecules or to a compound of the body with itself. Since, according to the
investigations of Wiedemann, in the case of hydrogen a continual rise in temperature
produces first a gradual diminution of the spectrum above described, and then upon reaching
a certain limit its almost sudden disappearance, these considerations lead us to the view
proposed as a second alternative by Angstrom, according to which the spectrum belongs to
a compound of hydrogen with itself. The heat-equivalent found by Wiedemann for the
quantity of energy necessary to transform this spectrum into that consisting of the three
characteristic brfght lines, would therefore be nothing else than the thermal equivalent of
the corresponding work of dissociation. This hypothesis furnishes an easy explanation of
the fact that in the spectra of the sun and most stars only the characteristic Hnes of this gas
appear as bright lines or absorption lines, as the case may be detected. The reason is to be
found in the enormous temperatures existing in these bodies.
J. N. Lockyer also believes dissociation occurs in flame and spark ; and A. Dufour
pointed out that the gas in a Geissler's tube is under conditions specially favourable
for dissociation, for the pressure is low and the temperature high. '
E. Frankland and J. N. Lockyer found that with a feeble current, or by changing
the pressure and temperature, the spectrum of hydrogen can be reduced to a single
i^-line. With increasing pressures the spectrum of hydrogen approaches a con-
tinuous one.
318 INORGANIC AND THEORETICAL CHEMISTRY
Tabls II. — Bai^bcbr's Sebees of Lines in the Arc Spectrum of Hydrogen.
HUne.
Wave-length, X.
Difference.
n
Calculated.
Observed.
HaOrC
3
6564-96
6564-97
+0-01
H^ or F
4
4862-93
4862-93
— .
HyOrG
5
4341-90
4342-00
-0 1
Hj OTh
6
410310
4103-11
+0-01
HeOrH
7
3971-4
3971-4
. —
H^ or a
8
3890-3
3890-3
—
H^or^
9
3836-7
3836-8
+0-1
He or y
H. or 8
10
3899-2
3799-2
11
3771-9
3771-9
—
Hk or €
12
3751-4
3751-3
-0-1
Hxor^
13
3735-6
3735-3
-0-3
H|u or 7]
Hvor^
14
3723-2
3722-8
-0-4
15
3713-2
3712-9
-0-3
In 1885, J. J. Balmer discovered a remarkable relation between the vibration-
periods or the wave-lengths of the spectral lines of hydrogen ; if the different lines
in the spark spectrum of hydrogen be numbered consecutively, starting with the
Ha-^ne as number 3, the next 4, 5, 6 . . ., the wave-length of the nth. line is given by
A=3647-2
or A-=3647-2
5—22
J. S. Ames' measurements are compared in Table II with values calculated by
means of J. J. Balmer's formula. The agreement between the calculated and
observed wave-lengths is very good, and hence the preceding formula may be taken
to represent closely the various kinds of elastic vibration which prevail among the
vibrating particles which produce Balmer's series of lines in the spark spectrum of
hydrogen. According to R. W. Wood, the different lines of the principal Balmer's
series and their accompanying channelled spectra are probably produced by
different entities — either by atoms which have lost 1, 2, 3, 4, . . . electrons, or
by aggregates or complexes of 1, 2, 3, 4, . . . atoms. In either case, it seems
probable that the members would break up in continuously increasing numbers.
W. E. Curtis' measurements led him to make a slight modification in Balmer's
formula. W. Ritz and F. Pascheni* found that there is another series of lines in the
ultra-red spectrum of hydrogen — the so-called Paschen's series, represented by
X=an^/{n^~3^), where a is a constant; and the so-called Lynman's series represented
by X=an^l{n'^—1^), in the ultra-violet.
F. Croze, 1^ and F. Paschen and E. Back have measured the Zeeman effect for the
primary lines of the hydrogen spectrum, viz. Ha, H^, Hy, and Hs. J. Stark and
co-workers have shown that certain hydrogen lines of the primary and secondary
series are resolved into linearly polarized components under the influence of a strong
electrical field of 1300 volts per cm. The longitudinal components are not polarized
so that the effect of the electrical field — Stark effect — is in marked contrast with
the Zeeman effect, for in the latter case the components are circularly polarized.
With still stronger electrical fields, 104,000 volts per cm., J. Stark found a separation
into further components. T. Takamine, N. Kokubu, and U. Yoshida have also
investigated the Stark effect on the hydrogen lines with a field strength of 15,000
volts per cm., and found that besides the primary lines, eleven lines in the region
below A=4000 were affected. C. Fabry and H. Buisson have made estimates of
the mass of the particles' which emit the first and second hydrogen spectra, and found
that is the same as that of an hydrogen atom. B. Reismann found hydrogen shows
both its spectra at the anode and cathode of a Geissler's tube excited by a direct
HYDROGEN 319
current discharge. H. L. P. Jolly has measured the distribution of energy in the
spectrum of hydrogen.
J. Tyndall could find no absorption of invisible heat radiations by hydrogen at
atm. pressure. Neither J. Janssen (1885) ^^ nor V. Schumann could find an absorption
spectrum for hydrogen gas ; W. Barmeister found hydrogen gas has no infra-red
absorption bands. J. Dewar (1894) did not find any in the Uquefied gas. Hydrogen
thus appears to be the most transparent of all known bodies ; even the Schumann
rays powerfully absorbed by other gases are freely transmitted by hydrogen. In
1907, A. Pfliiger, and R. Ladenburg and S. Loria showed that an absorption
spectrum can be obtained while the gas is in a state of luminescence during its
excitation by the discharge. Several others have since studied the absorption
spectrum of hydrogen.
A. L. Lavoisier and P. S. de Laplace^^ noticed that the hydrogen which is liberated
by the action of sulphuric acid on iron is positively electrified ; and in the case of
zinc, W. Hankel showed that when the gas is positively electrified, the sulphuric
or hydrochloric acid and the metal are negatively electrified. J. S. Townsend
showed that the electrification is not due to the spray mechanically carried by the
gas but is produced during the bubbling of the gas through the acjd. J. Enright
showed that this is but a special case of electrification by chemical action. J.Eranck,i®
R. Pohl, and W. B. Haines have shown that free electrons exist in hydrogen gas at
atmospheric pressures. Hydrogen gas, according to E. Villari, resists the passage of
an electric spark less than nitrogen, oxygen, or carbon dioxide. The discharge
tension in gases is influenced by numerous factors — temperature, pressure, form
of electrodes, the character of the spark, etc. — with spherical electrodes at 0'08 cm.
apart, the discharge potential is 2*4 kilovolts for hydrogen, 3'9 for air, 3*7 for
carbon monoxide, 4'9 for nitrogen, and 3*4 for oxygen ; with the electrodes 0'5 cm.
apart, the numbers are respectively 9*7, 17*5, 15'8, 18*0, and 15*6 kilovolts.
The difference between the potential of a plane and of a point placed at right
angles to the plane, needed for the passage of electricity, is called the minimum
potential. The observed minimum potentials depend on the sharpness of the point.
W. C. Rontgeni9 found 1296 and 1174 volts respectively for hydrogen at 205 and
110 mm. pressure ; and F. Tamm found that the decrease with high pressures is
small; but more rapid with low pressures. J. Precht obtained for a -f- point 2135 volts
and for a — point 1550 volts in hydrogen at 760 mm. E. Warburg, H. Sieveking, and
F. Tamm measured the relation between the current and potential. A. L. Hughes
and A. A. Dixon found the ionizing potential is dependent on the least energy
necessary to ionize the molecules of a gas by the impact of electrons, and amounts
to 10*2 volts for hydrogen ; J. Franck and G. Hertz found 11 volts, F. S. Goucher,
10'25 volts; and the value calculated by K. T. Compton's formula F=0194(Z— l)~i
is 11*8 volts, where V denotes the ionizing potential, and K the specific inductive
capacity. F. M. Bishop obtained 11 volts, and found the result independent of the
pressure. W. J. Paloff obtained a similar result. F. M. Bishop also found a second
type of ionization at 15*8 volts. Neither value is in accord with Bohr's theory,
which requires for the ionization potential 10*2 volts for the first line of longest
wave-length, and 13*6 volts for the shortest wave-lengths.
The ionization of hydrogen gas by Rontgen rays has been studied by R. K.
McClung, A. S. Eve, J. A. Crowther, G. Shearer, H. Donaldson, N. Campbell and
C. G. Barkla, and A. J. Philpot ; by radium radiations by A. S. Eve, L. Wertenstein,
W. Seitz and N. Campbell ; by a-rays by C. G. Darwin, W. Duane and G. L. Wendt,
T. S. Taylor, E. Rutherford and J. M.Nuttall, E. Marsden, and R. D. Kleeman ;
by j8-rays, by H. W. Schmidt ; by y-rays, by T. H. Laby and G. W. C. Kaye and
R. D. Kleeman ; by radium bromide by H. Baker ; by radiations from polonium
by T. S. Taylor ; by radiations from actinium by R. D. Kleeman ; by light from a
Geissler's tube by H. G. Cannegieter ; by canal rays by R. Seeliger ; by collision
by E. S. Bishop and W. J. Pawloff ; and by spraying by L. Block. The ionization
of the hydrogen flame has been studied by M. de Broglie. The action of hydrogen
320 INORGANIC AND THEORETICAL CHEMISTRY
on the electric discharge between various metals has been studied by A. Thiel and
E. Breuning, J. N. Pring, 0. W. Richardson, H. A. Wilson, A. Becker, and C. Sheard.
The electric discharge in hydrogen or in hydrogen mixed with other gases has been
studied by E. M. Wellisch, K. E. F. Schmidt, J. Trowbridge, K. Fredenhagen,
A. P. Chattock, and A. M. Tyndall. The mean values of J. Zeleny's, J. Franck's,
R. Pohl's, and A. P. Chattock's determinations of the velocities of the positive and
negative hydrogen ions when the electric discharge in a field of 1 volt per cm. are
respectively 606 and 7'69 cm. per second. J. Townsend gives for the diffusion
coeflGlcient of the positive and negative ions respectively 0123 and 0190 per sq. cm.
per second. J. Townsend, and R. K. McClung give respectively dnjdt^ZO^Orfi,
and dnldt= —2940^2, where n denotes the concentration of the ions, and dn/dt
the velocity of combination of the ions to form ordinary molecules. H. A. Erickson
and P. Phillips have studied the effect of temperature on this reaction. F. W.
Aston estimates that the minimum energy required for the ionization of the hydrogen
atom in a gas is l"7xlO~ii erg. W. B. Haines has investigated the mobilities
of the positive and negative hydrogen ions. Unlike all other elements yet investi-
gated, J. J. Thomson was never able to impart more than one charge of electricity
to the hydrogen atom.
According to the ionization hypothesis, the acids are more or less ionized in
aqueous solution, and they all furnish in common hydrogen ions which act as
carriers of positive electricity. The characteristic properties of acids are assumed
to be the characteristic properties of H"-ions ; during electrolysis the positively
charged hydrogen cations are discharged at the cathode, and the negatively charged
anions are discharged at the anode. The electric charge carried by a gaseous ion
is the same as that carried by a H'-ion during electrolysis ; the charge per ion in the
former case is approximately 4 X 10~io units, although in the latter case the observed
numbers vary between lXlO~io and 6xlO~io units. Again, the value of e/m,
where e denotes the charge and m the mass of the ion, is the same for all gaseous
ions, and approximate to 10^, while the value of the ratio e/m for the hydrogen ion
in solution is 10^. Consequently, the relation between the masses m of the gaseous
ion and of the hydrogen ion in solution is as 1 : 1700. The gaseous ion with its
negative charge was first called a corpuscle by J. J. Thomson, but the term electron
is now in general use. The positively charged hydrogen ion in solution is assumed
to be a hydrogen atom which has lost one of its negatively charged electrons;
during the scission of the molecule in the process of ionization the electron lost by
the hydrogen atom in forming the H'-ion remains attached to the other atom or
radicle and so imparts a negative charge to the anion.
Solutions containing equivalent quantities of the different acids do not neces-
sarily contain the same quantities of H*-ions, for a portion of the acid may not have
suffered ionization. There is a state of equilibrium, HAr=^H"-f A', in which anion,
cation, and un-ionized molecules are present. One or more of these three entities
may be more or less hydrated in the solution. The degree of ionization is dependent
on the concentration of the solution and on the temperature. The equivalent
conductivity at infinite dilution A^, when ionization is complete, is the sum of the
conductivities of the H'-ion, namely, v', and of the anion, v', so that X^=v'-\-v\
or v=X^—v'. The value of the molecular conductivity (jl changes with the tempe-
rature. According to F. Kohlrausch,20 the conductivity of the hydrogen ion at 6°
is r=318-l+0-0154(^-18)-0-000033(^-18)20 reciprocal ohms. W. Ostwald and
R. Luther say that F. Kohlrausch's value ?;-=352 (25°) is too high and give
r=347 (25°) ; A. A. Noyes and G. V. Sammet say that F. Kohlrausch's value
is too low, and give v—SQi'd (25°). V. Rothmund and K. Drucker's value is
tJ-=338.
The hydrogen electrode is used in measuring the concentration of H*-ions in a
solution ; it is based on the definite difference of potential which exists between a
platinum or palladium electrode, saturated with hydrogen gas at a given pressure,
and immersed in a solution of definite acidity or alkalinity. The particular forms oi
HYDROGEN 321
the cells are indicated in laboratory manuals. The e.m.f., E, of the cell H2Pt |
Solution I I Electrolyte | Solution II | H2Pt is given by
^= — lege Sil; or ^=00001987 logio El volts
« m h LHj2
where € represents the farad, 96,540 coulombs, and R the gas constant, 9'316 joules.
If solution I be OOliV-HCl, and solution II, O'OOliV-HCl, and the temperature
be 18°, £'—0058 volt, nearly. The hydrogen electrode with a normal solution of
H-ions may be used as a standard, and the hydrogen concentration of a given
solution is determined by measuring the e.m.f. of the combination. Suppose it is
0-5 volt at 18°. Consequently, 0-5=0-058(logio 1— log [H'lr) or the required
concentration of the H*-ions is 2-4x10-9. The total temperature coefficient of the
hydrogen electrode is small, and is made up of a number of factors — concentration
of dissolved gases, pressure of water vapour or of other gases, etc. Pressure
raises the positive potential of the hydrogen. If the pressure of the hydrogen
is p atm., the correction to be added at 18° is E—Ep=—^ of 0*058 log p. If the
pressure be 740 mm., the correction is J of 0*058 log (740/7 60) =0 00033 volt.
The fluctuations of the barometer can usually be neglected.
The hydrogen ions often have a catalytic action accelerating the speed of some
reactions, and retarding the speed of others. The catalytic action of the H"-ions
on the speed of the inversion of cane sugar, the hydrolysis of methyl acetate, etc.,
has been used to estimate their concentration. The discharge of hydrogen ions
furnishes hydrogen molecules 2H' =H2. W. Ostwald estimates the heat of ionization
to be H2=2H — 11 Cal. ; and K. Fajans, the heat of hydration of gaseous H'-ions
to be 362 kgrm. cal. per gram-ion. The free energy of formation of a hydrogen ion
is found by measuring the difference of potential E between hydrogen gas at a
pressure p atm. and a solution of H-ions of concentration G gram-ions per litre at the
absolute temperature T, since E=EQ-\-RT{\og C— J log^), where R, the gas constant,
is 0"861 XlO-* ; and Eq is the potential when 0 and p are unity. Measurements
by N. T. M. Wilsmore and others give £'o=— 0*283 volt (25°) with the normal
calomel electrode zero. The value of Eq thus measures the tendency of hydrogen
to form ions, or half the intensity of the electro-affinity. W. Nernst took the
potential of the hydrogen electrode — platinum or platinized palladium saturated
with hydrogen — as zero because of its position in a series of the potentials of the
different elements. W. Ostwald objected to this because of the large variation which
occurs with changes of pressure — one millivolt per 0*0345 atm. The relatively
small electro-affinity of the hydrogen ion is correlated with its great tendency to
form complex ions. Thus, with ammonia it forms NH4 ions, and similarly with
the various amines and oxonium compounds ; Hkewise also with the anions of the
acid salts HSO2', HSO3', HCO3', HC2O4', etc.
G. W. Osann 21 obtained a gas which he called ahtiven Wasserstoff oder Ozonwasser-
stoff by the electrolysis of a mixture of distilled Nordhausen sulphuric acid and water.
He claimed that the gas is a far more active reducing agent than ordinary hydrogen ;
but neither J. Lowenthal nor G. Magnus could confirm G. W. Osann's conclusions,
and they attributed his results to the contamination of his hydrogen with some
sulphur dioxide. W. Duane and G. L. Wendt exposed hydrogen, of as high a degree
of purity as they could prepare, to the intense bombardment of a-radiations, and
found the gas became more chemically active* at ordinary temperatures, for it then
combined directly with sulphur to form hydrogen sulphide ; with phosphorus to
form phosphine ; with arsenic to form arsine ; and with nitrogen to form ammonia.
It reacts with mercury forming yellow crystals — possibly mercury hydride — which
resist attack by water and weak alkali lye, but dissolve in hydrochloric and nitric
acids ; when gently warmed, the yellow crystals form globules of mercury. The
activated hydrogen reduces potassium permanganate solutions forming manganese
dioxide ; it does not bleach methyl violet or indigo carmine. W. Duane and
VOL. I. Y
322 INORGANIC AND THEORETICAL CHEMISTRY
G. L. Wendt suggest that the activated gas is related to normal hydrogen as ozone
is related to oxygen — a kind of ozonohydrogen, H3. They do not beheve that it is
monatomic hydrogen because its formation is attended by a contraction in volume.
The activation of the hydrogen is not due to the formation of ions because it is not
destroyed by the passage of the gas through glass wool and an intense electrostatic
field ; but it is removed by passing the gas through a tube immersed in liquid air.
The activated hydrogen is not stable for its life is measured in minutes. J. J.
Thomson found that in a discharge tube containing hydrogen, there are present
charged atoms, charged molecules, and sometimes a constituent with three times
the mass of ordinary hydrogen atoms. The potential used was of the order 20,000
volts.
A. J. Dempster found that with a potential of 800 volts hydrogen is ionized by
detaching a single elementary charge from the molecule, but the gas is not dis-
sociated ; the positive molecules so formed, however, can dissociate the gas forming
the complex H3. This constituent is not stable, and is not present when there is
no dissociation of the hydrogen molecule.
Liquid hydrogen is electrically non-conducting. The dielectric constant, K, of
hydrogen 22 was found by A. Occhialini to be given by (Z— 1)/(Z+2)Z)=90154 X 10" ^
between 94 and 196 atm. pressure ; and by extrapolation Z=l '0002705 at ordinary-
pressures. Other determinations of the dielectric constant of hydrogen gas at
atmospheric pressures and 0°, give 1 '000264 (vacuum unity), and at 20°, 1 '000273
— since the dielectric constant of air at 0° is 1*000590, the dielectric constant
of hydrogen at 0° is 0'999674 (air unity). At 20° and 20 atm. pressure, the
dielectric constant of hydrogen is 1 "00500 ; under 60 atm. pressure 1 '01460 ; and
under 100 atm. pressure, 1-02378. At —191°, and 760 mm., H. Riegger found
1-000928. Hydrogen is diamagnetic. The magnetic susceptibility of hydrogen
at 1 to 40 atm. pressure and 16° is below 0'008 X 10" ^^ volume units. 23 According
to S. Henrichsen, the atomic magnetism of hydrogen in organic compounds is 9.
References.
1 T. Thomson, Ann. Phil, 16. 161, 1820; Compt. Rend., 12. 1048, 1841 ; J. B. Biot and,
F. J. Arago, Mem. Inst., It. 7, 1787 ; P. L. Dulong and J. J. Berzelius, Ann. Chim. Phys., (2),
15. 386, 1820 ; J. B. A. Dumas and J. B. J. D. Boussingault, (3), 3. 275, 1841 ; Compt. Bend., 12.
1013, 1841 ; H. V. Regnault, ib., 20. 975, 1845 ; J. M. Crafts, ib., 106. 1662, 1888 ; 11. 509, 1890 ;
A. Leduc, ib., 113. 186, 1891 ; Lord Rayleigh, Proc. Roy. Soc, 53. 134, 1893; E. W. Morley,
On the Density of Oxygen and Hydrogen, and on the Ratio of their Atomic Weights, Washington,
1895 ; Zeit. phys. Chem., 20. 242, 1896 ; J. Thomsen, Zeit. anorg. Chem., 12. 6, 1896 ; J. P. Cooke,
Am^r. Chem. Journ., 11. 509, 1890 ; J. Dewar, Proc. Chem. Soc, 24. 146, 1898 ; Proc. Rcy. Soc,
73. 251, 1904; V. Meyer, Ber., 13. 2019, 1880.
2 F. Kleint, Verh. deut. phys. Ges., 3. 146, 1905; E. Dorn, Wied. Ann., 13. 378, 1881 ;
J. Bosler. Compt. Rend., 146. 686, 1908 ; J. Perrin, ib., 147. 594, 1908 ; W. Sutherland. Phil. Mag.,
(5), 36. 507, 1893 ; (6), 17. 320, 1909 ; (6), 19. 25, 1910 ; T. Svedberg, Zeit. phys. Chem., 67.
105, 1909 ; J. H. Jeans, The Dynamical Theory of Gases, Cambridge, 1916 ; R. D. Kleeman.
Proc. Cambridge Phil. Soc, 17. 149, 1913 ; Phil. Mag., (6), 19. 783, 1910 ; A. E. Haas, Phys. Zeit.,
11. 531, 1910; W. Altberg, Ann. Physik, (4), 37. 849, 1912; H. Erfle, ib., (4), 23. 594, 1907 ;
H. Sirk, ib., (4), 25. 894, 1908 ; M. Reinganum, ib., (4), 10. 334, 1903 ; (4), 28. 142, 1908 ;
M. Knudsen, ib., (4), 35. 389, 1911 ; J. Robinson, ib., (4), 31. 769, 1910; J. Dewar, Proc Roy. Soc,
79. A, 529, 1907 ; Compt. Rend., lA^. 110, 1907; P. Debye, Phys. Zeit., 11. 1115, 1910;
W. Altberg, Journ. Russ. Phys. Chem. Ges., 44. 431, 1912 ; N. Bohr, Phil. Mag., (6), 26. 1, 1913 ;
E. Rutherford, ib., (6), 27. 488, 1914 ; J. W. Nicholson, ib., (6), 22. 864, 19lf ; ib., (6), 27. 541,
1914; J. Langmuir, Journ. Amer. Chem. Soc, 34. 860, 1912; Phil. Mag., (6), 27. 188, 1914;
R. Fiirstinau, Phys. Zeit., 9. 849, 1908; J. H. Jeans, Phil. Mag., (6), 8. 692, 1904; A. P. N.
Franchimont, Rec Trav. Chim. Pays-Bas, 1. 275, 1882; J. J. van Laar, Proc. Acad. Amsterdam,
20. 750, 1195, 1918.
3 0. E. Meyer, Pogg. Ann., 148. 14, 1871 ; 0. E. Meyer and F. Springmuhl, ib., 148. 626,
1873 ; H. Buff, ib., 158. 177, 1876 ; W. Kopsch, U^er die Koeffizienten der inneren Reibung von
Wasserstoff und Argon bei niederen Temperaturen, Halle, 1909; J. Puluj, Sitzber. Akad. Wien,.
78. 279, 1879; 79. 97, 743, 1879; F. Kleint, Verh. deut. phys. Ges., 3. 146, 1905; W. Crookes,
Proc Roy. Soc, 31. 446, 1881 ; Lord Rayleigh, ib., 62. 112, 1897 ; J. H. T. Roberts, Phil. Mag.,,
(6), 23. 250, 1912 ; C. Barus, Amer. Journ. Science, (3), 35. 407, 1888 ; H. Markowsky, Ann.
Physik, (4), 14. 742, 1904 ; A. Bestelmeyer, ib., (4), 15. 423, 1904 ; K. Schmitt, ib., (4), 30.
HYDROGEN 323
393, 1909 ; P. Breitenbach, ib., (4), 5. 166, 1901 ; M. Knudsen, ib., (4), 28. 75, 1908 • H Voce!
ib., (4), 43. 1235, 1914; J. E. Verschaffelt, Proc. Akad. Amsterdam, 20. 986," 1918*
J. E. Verschaffelt and C. Nicaise, ib., 19. 1084, 1917 ; E. C. Bingham and J. P. Harrison'
Zeit. phys. Chem., 66. 1, 1909 ; E. C. Bingham, Amer. Chem. Journ., 43. 287, 1910; K. L. Yen'
Phil. Mag., (6), 38. 582, 1919; A. Kundt and E. Warburg, Pogg. Ann., 155. 337, 1875; O. E.'
Meyer, Die Jcinctische Theorie der Gase, Breslau, 152, 1877; W. Sutherland. Phil. Maa '(5) 36
507, 1893. ^ ' ^^'
4 P. L. Dulong, Ann. Chim. Phys., (2), 41. 113, 1829 ; I. B. Zoch, Pogg. Ann., 128. 497, 1866.
5 S. von Wroblewsky, Monatsh., 9. 1067, 1888 ; E. H. Amagat, Compt. Rend., 75 479 'l872 •
90. 995, 1880 ; 95. 281, 1882 ; 107. 522, 1888 ; Ann. Chim. Phys., (5), 25. 456, 1883 ; L. P. CaiUetet'
Compt. Rend., 70. 1131, 1870; J. A. Siljestrom, Pogg. Ann., 151. 451, 573, 1874; F. Fucks'
Wied. Ann., 35. 430, 1888 ; E. Budde, Journ. prakt. Chem., (2), 9. 30. 1874 ; C. Pusch'l Monatsh '
8. 374, 1887 ; 9. 39, 1888 ; W. J. de Haas. Proc. Acad. Amsterdam, 19. 1468, 1917 ; H. K.'
Onnes and C. A. Crommelin, ib., 19. 245, 1917 ; P. A. Guye, Mem. Soc. Phys. 'Oeneve' 1 977
1909 ; Compt. Rend., 144. 976, 1907 ; S. J. C. Schalkwijk, Comm. Phys. Lab. Leiden, 70, 1902 ;
H. K. Onnes and C. Braak, ib., 97. 99, 1907 ; H. K. Onnes and H. H. F. Hyndma'nn 'ib 78*
23, 1902 ; W. J. de Haas, ib., 127. 3, 1912 ; 0. Sackur, Zeit. Elektrochem., 20. 563, 1914 •
A. Jaquerod and 0. Scheuer, Mem, Soc. Phys. Geneve, 35. 659, 1908; T>. Berthelot Compt Rend '
126. 954, 1898 ; 145. 180, 1907 ; A. Leduc, ib., 123. 743, 1896 ; A. Leduc and P. Sacerdote ib'
125. 297, 1897 ; Lord Eayleigh, Proc. Roy. Soc, 69. 448, 1892 ; 196. 205, 1901 ; Phil Trans'
380. A, 351, 1905.
6 P. von Jolly, Pogg. Ann. Jubelbd., 82, 1874 ; A. W. Witkowsky, Bull. Acad. Cracow, 305,
1905; A. Leduc, Compt. Rend., 148. 1173, 1909; M. W. Travers, G. Senter, and A Jaquerod
Proc. Roy. Soc, 70. 484, 1903.
' J. Stefan, Sitzber. Akad. Wien, 65. 45, 1872 ; J. Priestley, Experiments and Observations
relating to Various Branches of Natural Philosophy, Birmingham, 378, 1781 ; F. C. Achard
Ber. Akad. Berlin, 84, 1785; N. N. Beketoff, Ber., 12. 686, 1879; A. Winkebnann, Pogg Ann'
157. 457, 1876 ; 159. 177, 1876 ; Wied. Ann., 1. 63, 1877 ; 44. 177, 429, 1891 ; L. Graetz, ib'.,
14. 232, 1881 ; J. Janssen, Wied. Ann. Biebl., 701, 1879 ; A. Schleiermacher, ib., 34. 623 1888 •
P. A. Eckerlein, Ann. Physik, (4), 3. 120, 1900; S. Weber, ib., (4), 54. 325, 1917; L Graetz'
Wied. Ann., 14. 232, 1881 ; C. Christiansen, ib., 14. 23, 1881 ; 19. 282, 1883; A. Kundt and'
E. Warburg, Pogg. Ann., 156. 177, 1875; S. Weber, Ann. Physik, (4), 54. 325, 437, 1917.
8 P. Vieille, Compt. Rend., 96. 1218, 1358, 1883 ; M. Berthelot and P. Vieille, ib., 98. 770,
852, 1884 ; E. Mallard and H. le Chatelier, ib., 93. 1014, 1881 ; J. Jamin and F. Richard, ib 71 '
336, 1870 ; W. C. Rontgen, Pogg. Ann., 148. 580, 1873 ; F. Haber and L. Bruner, Zeit. Elektro-
chem., 12. 78, 1906 ; W. Nernst and H. von Wartenburg, Zeit. phys. Chem., 56. 543, 1906 ;
L. Holborn and P. Henning, Ann. Physik., (4\ 78. 739, 1905; G. N. Lewis andM. Randall, Journ.
Amer. Chem. Soc, 36. 1969, 1914; 34, 1128, 1912; M. Pier, Zeit. Elektrochem., 15. 536, 1909;
N. Bjerrum, ib., 18. 101, 1912 ; K. Scheele and W. Heuse, ib., (4), 37. 79, 1912 ; (4), 40. 473'
1913 ; W. Escher, ib., (4), 42. 761, 1913 ; A. Piecken, Sitzber. Akad. Berlin, 13. 44, 1913 ; H. V.'
Regnault, Ann. Chim. Phys., (2), 73. 1, 1840 ; E. Wiedemann, PM. Mag., (5), 2. 81, 1876 ; Pogg
Ann., 157. 1, 1876 ; M. C. Shields, Phys. Rev., (2), 10, 535, 1917 ; 0. Lummer and E. Prin^sheim,
TFieti. ^wn,, 64. 536, 1898; S. Lussana, iVwow Cmewto, (3), 36. 5, 70, 130, 1894; (4), 2. 327,'
1895 ; (4), 3. 92, 1896 ; W. H. Keesom and H. K. Onnes, Proc Acad. Amsterdam, 20. 1000, 1918 ,'
J. Joly, Proc. Roy. Soc, 41. 352, 1886; Phil. Trans., 182. A, 73, 1892; 185. A, 943, 1894;
G. N. Lewis and G. E. Gibson, Journ. Amer. Chem. Soc, 39. 2554, 1917; A. Eucken, Ber. deut!
phys. Ges., 18. 4, 1916.
9 J. Dewar, Journ. Chem. Soc, 73. 528, 1898 ; Chem. News, 84. 281, 293, 1901 ; 91. 216,
1905; Proc Roy. Soc, 73.251,1904; 64.231,1898; Compt. Rend., 129.451,1899; K. Olszew-
sky, Phil. Mag., (5), 39. 188, 1895 ; (5), 40. 202, 1895 ; M. W. Travers and A. Jaquerod, Chem.
News, 86. 61, 1902 ; Zeit. phys. Chem., 49. 224, 1904 ; H. K. Onnes and C. A. Crommelin, Proc.
Acad. Amsterdam, 16. 245, 1913; A. Eucken, Ber. deut. phys. Ges., 18. 4, 1916; W. H. Keesom,
Comm. Phys. Lab. Leiden, 137, 1913 ; W. H. Keesom and H. K. Onnes, ib., 137, 1913 ; P. G.
Cath and H. K. Onnes, Proc Acad. Amsterdam, 20. 1155, 1160, 1918; W. Wahl, Proc Roy. Soc,
88. A, 61, 1913.
^0 M. Croullebois, Ann. Chim. Phys., (4), 20. 136, 1870; J. H. Gladstone. Proc Roy. Soc,
31. 327, 1881 ; W. Ramsay and M. W Travers, ib., 62. 227, 1897 ; I. Traube, Ber., 30. 39, 1897;
L. Natanson, Zeit. phys. Chem., 61. 321, 1908; R. Ladenburg, ^n?i. Physik., (4), 38. 249, 1912;
R. Ladenburg and S. Loria, Ber. deut. phys. Ges., 6. 858, 1908; Phys. Zeit., 9. 875, 1908; J. W.
Bruhl, Ber., 13. 119, 1880; J. Traube, ib., 30. 39, 1897; H. Landolt, ib., 15. 1031, 1882; Pogg.
Ann., 122. 545, 1864; 123. 595, 1864; F. Eisenlohr, Zeit. phys. Chem., 75. 585, 1910; J, W.
Briihl, ib., 1. 307, 1887 ; 7. 1, 140, 1891 ; J. H. Gladstone, Proc Roy. Soc, 31. 327, 1881 ; 18.
49, 1892 ; Phil. Trans., 159. 13, 1869 ; E. Ketteler, Farhenzerstreuung der Case, Bonn, 1865 ;
L. Lorenz, Wied. Ann., 11. 70, 1880 ; J. Koch, Arkiv. Mat. Astr. Fysik., 8. 20, 1912; Ann. Physik.,
(4), 17. 658, 1905; E. Mascart, Ann. £cole Norm. Sup., 6. 1, 1877; F. Perreau, Ann. Chim..
Phys., (7), 7. 289, 1896 ; K. Seheel, Verh. deut. phys. Ges., 9. 24, 1907 ; C. and M. Cuthbertson,
Proc Roy. Soc, 83. A, 164, 1909; K. Kermann, Ber. deut. phys. Ges., 6. 211, 476, 1908;
H. Rohmann, Ann. Physik., (4), 34. 979, 1911 ; S. H. Siertsemaand W. J. de Haas, Proc. Acad.
Amsterdam, 20. 624, 1918; Phys. Zeit., 14. 568, 1913.
11 M. Faraday, Phil. Trans., 136. 1, 1846; Experitnental Researches in Electricity, London,
324 INORGANIC AND THEORETICAL CHEMISTRY
8. 1, 1855 ; E. Verdet, Ann. Chim. Phys., (3), 41. 370, 1854 ; (3), 52. 129, 1858 ; A. Kiindt and
W. C. Rontgen, Wied. Ann., 6. 332, 1879; W. H. Perkin, Journ. Chem. Sot., 69. 1237, 1896 ;
55. 706, 1889 ; 65. 59, 1893 ; 51. 363, 1887 ; 45. 501, 1884.
12 L. H. Siertseraa, Arch. Nierl, (2), 6. 830, 1901 ; Proc. Akad. Amsterdam, 7. 294, 1899.
1' A. V. Angstrom, Recherches svr le spectre solaire, Upaala, 1868 ; H. W. Vogel, Ber. Akad.
Berlin,, 586, 1879 ; A. Paalzoff, ih., 192, 1880 ; J. J. Balmer, Wied. Ann., 25. 80, 1885 ; 60. 380,
1897 ; J. S. Ames, Phil. Mag., (5), 30. 33, 1890 ; J. Trowbridge and T. W. Richards, ih., (5),
43. 135, 1897 ; L. Janicki, Ann. Physik., (4), 19. 36, 1906 ; F. Emich, Sitzher. Akad. Wien, 109.
411, 1900; J. Trowbridge and T. W. Richards, Amer. Journ. Science, (4), 3. 117, 1899; Phil. Mag.,
(5), 43. 135, 1899 ; J. Trowbridge, ib., (6), 5. 153, 1903 ; T. W. Richards, Amer. Chem. Journ.,
21. 172, 1899 ; A. WiiUner, Pogg. Ann., 135, 497, 1868 ; 137. 337, 1869 ; 147. 321, 1872 ; Wied.
Ann., 14. 355, 1881 ; J. Pliicker and W. Hittorf, Proc. Roy. Soc., 13. 153, 1864; A. Schuster,
Arch. Sciences Geneve, 45. 414, 1872 ; E. Villari, ih., 44. 84, 1872 ; A. V. Angstrom, Pogg. Ann.,
144. 300, 1871; B. Hasselberg, Bull. Acad. St. Petersburg, 27. 97, 1881; 30. 14, 1885; Phil.
Mag., (5), 17. 351, 1884; G. Salet, Ann. Chim. Phys., (4), 28. 5, 1873 ; A. Griinwald, Monatsh.,
8. 344, 1887 ; 13. Ill, 1892 ; R. S. Hutton, Phil. Mag., (5), 46. 338, 1898 ; A. Sundell, ih., (5),
24. 98, 1887; G. M. Seabrooke, ih., (4), 43. 155. 1872; C. Fiavez, Gmnpt. Rend., 92. 521, 1881;
D. van Monckhoven, ih., 95. 378, 1882; L. Cailletet, ih., 74. 1282, 1872; E. Frankland and
J. N. Lockyer, ih., 68. 420, 1869 ; Phil. Trans., 163. 650, 1873; A. Dufour, Ann. Chim. Phys.,
(8), 9. 431, 1906; R. W. Wood, Physical Optics, New York, 537, 579, 1911 ; T. Takaraine and
U. Yoshida, Mem. Coll. Science, Kyoto, 2. 137, 1917 ; E. Wiedemann, Wied. Ann., 10, 202, 1880.
" F. Paschen, Ann. Physik., (4), 27. 565, 1908 ; W. Ritz, Phys. Zeit., 9. 521, 1908 ; T. Lynman,
ib., 13. 583, 1912; W. E. Curtis, Proc. Boy. Soc, 90. A, 605, 1914.
15 F. Paschen and E. Back, Ann. Physik, (4), 39. 897, 1912 ; J. Stark, Sitzher. Akad. Berlin,
932, 1913 ; Nachr. Gott.. 427, 1914 ; J. Stark and G. Wendt, Ann. Physik, (4), 43. 983, 1914 ;
J. Stark, ih., 48. 183, 1915 ; J. Stark and H. Kirschbaum, ih., (4), 43. 991, 1017, 1914 ; T. Takamin
and N. Kokubu, Mem. Coll. Science Kyoto, 3. 271, 1919 ; T. Takamine and U. Yoshida, ih., 2.
137, 1917 ; C. Fabry and H. Buisson, Compt. Rend., 154. 1500, 1912 ; B. Reismann. Zeit. wiss.
Phot., 13. 269, 1914 ; H. L. P. Jolly, Phil. Mag., (6), 26. 801, 1913.
i« J. Janssen. Compt. Rend., 101. 649, 1885 ; R. Ladenburg, Ber. deut. phys. Ges., 12. 5, 1911 ;
6. 550, 1908 ; A. Pfliiger, Ann. Physik, (4), 24. 515, 1907 ; H. Barwald, Phys. Zeit., 11. 145,
1910 ; Verh. deut. phys. Qes., 12. 159, 1910 ; W. Jungjohann, Zeit. wiss. Phot., 9. 84, 105, 141,
1910; H. Rubens and H. von Wartenburg, Phys. Zeit., 12. 1080, 1911 ; J. Niviere, Bull. Soc.
Chim., 13. 958, 1913 ; E. C. C. Baly, Proc. Roy. Soc, 90. A, 409, 1914 ; J. Dewar, Chem. News,
70. 115, 1894 ; V. Schumann, Ann. Physik, (4), 4. 642, 1901 ; W. Burmeister, Verh. deut. phys.
Ces., 15. 589, 1913; R. Ladenberg and S. Loria, ih., 10. 858, 1908; J. Tyndall, Proc Roy. Soc,
35. 129, 1883.
17 A. L. Lavoisier and P. S. de Laplace, Mem. Acad., 359, 1780; 387, 1784; J. Enright,
Phil. Mag., (5), 29. 56, 1890 ; J. S. Townsend, Proc Cambridge Phil Soc, 9. 345, 1897 ; Phil.
Mag., (5), 45. 125, 1898 ; W. Hankel, Wied. Ann., 22. 387, 1884.
18 W. Dnane and G. T,. Wendt, Phys. Rev., (2), 10- 116, 1917: E. Villari, Arch. Sciences
Geneve, 44. 85, 1872 : J. E. Almv, Phil Mag., (6), 16. 456, 1908 ; N. Campbell, ih., (6), 26. 912,
1913 ; A. S. Eve, ih., 8. 610, 1904 ; H. Donaldson, ih., (6), 22. 729, 1911 ; C. G. Barkla and J. A.
Philpot, ib., (6), 25. 832, 1913 ; C. G. Darwin, ih., (6), 23. 901, 1912 ; ib., (6), 27. 499, 1914 ; F. S.
Taylor, ih., (6), 26. 402, 1913 ; E. Rutherford and J. M. Nuttall, ib., (6), 26. 702, 1913 ; J. J.
Thomson, ih., (6), 24. 668, 1912 ; E. Marsden, ib., (6), 27. 824, 1914 ; T. H. Laby and G. W. C.
Kaye,i&., (6), 16. 879, 1908; C. Sheard, ih., (6), 28. 170, 1914; D. N. Mallik, ib., (6), 16. 531, 1908;
A. P. Chattock and A. M. Tvndall, ib., (6), 16. 24, 1908 ; ib., (6), 19. 449, 1910 ; A. P. Chattock,
ih., (5), 48. 401, 1899 ; (6), 1. 79, 1901 ; W. B. Haines, ih., (6), 30, 503, 1915 ; (6), 31. 339,
1916 ; G. Shearer, ib., (6), 30. 644, 1915 ; R. K. McClung, ih., (6), 3. 283, 1902 ; (6), 8. 357, 1904 ;
Proc. Cambridge Phil Soc, 12. 375, 1904 ; E. M. Wellisch, Proc Roy. Soc, 82. A, 500, 1919 ;
Radium. 6. 241, 1909 ; K. E. F. Schmidt, Phys. Zeit, 8. 617, 1907 ; J. Trowbridge, Amer. Journ.
Science, (4). 29. 341, 1910; C. Fredenhagen, Phys. Zeit., 15. 19, 1914; J. A Crowther, Phil.
Mag., (6), 14. 653, 1908 ; Proc Cambridge Phil Soc, 15. 39, 1909 ; Proc Roy. Soc, 82. A, 103,
1909 ; J. Franck, Ber. deut phys. Ges., 5. 194, 1907 ; J. Franck and R. Pohl, ih., 11. 397, 1909 ;
L. Wertenstein, Cmnpt. Rend., 151. 469, 1910 ; ih., 155. 449, 1912; Radium, 7. 225, 1910;
W. Seitz, Phys. Zeit., 13. 476, 1912; R. D. Kleeman, Proc Roy. Soc, 79. A, 220, 1907;
ib., 82. A, 350, 1909; P. Philips, ih., 83. A, 246, 1910; H. W. Schmidt, Phys. Zeit, 10.
929, 1909 ; Verh. deut. phys. Ges., 11. 605, 1909 ; H. B. Baker, Proc Roy. Inst., 19. 791, 1910 ;
T. S. Taylor, Amer. Journ. Science, (4), 31. 249, 1911 ; Phil Mag., (6), 21. 571, 1911 ;
H. A. Erickson, ih., (6), 18, 328, 1909 ; A. Occhialini, Atti Accad. Lincei, (5), 22. ii, 482, 1913 ;
Nuovo Cimento, (6), 7. 108, 1914 ; R. D. Kleeman, Proc Roy. Soc, 83. A, 530, 1910 ; H. C.
Cannegieter, Proc Acad. Amsterdam, 19. 1331, 1911 ; R. Seeliger, Phys. Zeit.. 12. 839, 1911 ;
E. S. Bishop, ib., 12. 1148, 1911 ; W. J. Pawloff, Proc Roy. Soc, 90. A, 694, 1910 ; F. W. Aston,
ih., 80. A. 45, 1907 ; M. de Broglie, Radium, 8. 106, 1911 ; A. Thiel and E. Breuning, Festschr.
Versamm. Naturforsch. Aerzte Milnster, 148, 1912; Zeit anorg. Chem., 83. 329, 1913; J. N.
Pring, Zeit Electroehem. , 19. 2.55, 1913; O. W. Richardson, Proc Cambridqe Phil Soc, 13. 192,
1906 ; H. A. Wilson, Proc Roy. Soc, 80. A, 379, 1908 ; ib., 82. A, 71, 1909 ; A. Becker, Ann.
Physik., (4), 29. 909, 1909 ; ib., 31. 98, 1909 ; A. Coehn and H. Mozer, Ann. Physik., (4), 43.
108, 1913 ; A. M. TyndaU, Compt Rend., 152. 1375, 1911 ; E. H. Riesenfeld, Zeit Electroehem.,
HYDROGEN 325
17. 725, 1911 ; E. H. Amagat, Cmnpt. Rend., 154. 909, 1912; J. Townsend, Phil. Trans., 193.
129, 1900 ; 195. A, 259, 1900 ; J. Zeleny, ib., 195. A, 193, 1900 ; J. Franck and R. Pohl, Vtrh.
dent. phys. Ges., 9. 69, 1907.
»» R. F. Earhart, Phys. Rev., (1), 33. 188, 1912 ; (2), 1. 85, 1913; E. H. Williams, ib., (1), 31.
216, 1910 ; J. Zeleny, ib., (1), 25, 305, 1907 ; (1), 26. 129, 1908 ; W. Kaufmann, 0611. Nachr., 243,
1899 ; W. C. Rontgen, ib., 390, 1878 ; E. Rieche, Ann. Physik, (4), 4. 292, 1901 ; F. Tamm, ib.,
(4), 6. 259, 1901 ; H. Sieveking, ib., (4), 1. 299, 1900 ; E. Warburg, ib., (4), 2. 295, 1900 ; Wied.
Ann., 67. 69, 1899 ; J. Precht, ib., 49. 150, 1893 ; E. Villari, AM Accad. Lincei, 7. 297, 1883;
J. E. Almy, Phil. Mag., (6), 16. 456, 1908 ; G. M. Hobbs, ib., (6), 10. 617, 1905 ; A. L. Hughes
and A. A. Dixon, Phys. Rev., (2), 10. 495, 1917 ; F. M. Bishop, ib., (2) 10. 244, 1917 ; B. Davis
and F. S. Goucher, «6., (2), 10. 101, 1917 ; K. T. Compton, ib., (2), 8. 412, 1916 ; F. S. Goucher,
ib., (2), 8. 561, 1916 ; J. Franck and G. Hertz, Verh. deut. phys. Oes., 15. 34, 1913 ; W. B. Haines,
Phil. Mag, (6), 30. 503, 1915 ; (6), 31. 339, 1916 ; W. J. Paloff, Proc. Roy. Soc, 90. A, 694, 1910.
2" F. Kohlrausch, Sitzber. Akad. Berlin, 1026, 1901 ; W. Ostwald and R. Luther, Hand- und
Hillfsbuch zur Ausfuhrung physiko-chemischer Messungen, Leipzig, 1902 ; A. A. Noyes and G. V.
Sammet, Zeit. phys. Chem., 43. 49, 1903 ; V. Rothmund and K. Drucker, ib., 46. 827, 1903 ;
N. T. M. Wilsmore, ib., 35. 302, 1900 ; W. Ostwald, ib., 35. 333, 1900 ; Lehrbuch der allgemeinen
CAemje, Leipzig, 2. i, 952, 1903; W. Nernst, Ber., 30. 1557. 1897; L. Michaelis, Die Wasser-
stoffionenkonzentration, Berlin, 1914. For a bibliography of the concentration of the hydrogen
ion up to 1919, see C. L. A. Schmidt and D. R. Hoagland, Pub. Physiol. Univ. California, 2.
23, 1919; K. Fajans, Ber. deut. phys. Ges., 21. 549, 1919.
21 G. W. Osann, Journ. prakt. Chem., (1), 58. 385, 1853; (1), 61. 500, 1854; (1), 66. 102,
1855; (1), 69. 1, 1856; (1), 71. 355, 1857; (1), 78. 93, 1859; (1), 81. 20, 1860; (1), 92. 210,
1864; Pogg. Ann., 95. 311, 1855; 97. 327, 1856; 98. 181, 1856; 106. 326, 1859; Wilrzburg
Nat. Zeit., 4. 7, 19, 1864 ; J. Lowenthal, Journ. prakt. Chem., (1), 73. 116, 1858 ; G. Magnus,
Pogg. Ann., 104. 555, 1858 ; W. Duane and G. L. Wendt, Phys. Rev., (2), 10. 116, 1917 ; A. J.
Dempster, Phil. Mag., (6), 31. 438, 1916 ; J. J. Thomson, Rays of Positive Electricity, London,
116. 1913.
22 L. Boltzmann, Pogg. Ann., 155. 403. 1875; Sitzber. Akad. Wien, 69. 196, 1872;
K. Klemencic, ib., 91. 712, 1885 ; K. Tangl, Ann. Physik., (4), 23 559, 1907 ; (4), 26. 59, 1908 ;
H. Rohmann, ib., (4), 34. 979, 1911 ; A. Occhialmi, AtH Accad. Lincei, (5), 22. ii, 482, 1903 ;
E. Oxley, Proc. Roy. Soc, 95. A, 58, 1918; H. Riegger, Ann. Physik, (4), 59. 753, 1919.
23 G. Qumcke, Wied. Ann., 24. 347, 1885 ; 34. 401, 1888 ; S. Henrichsen, ib., 34. 180, 1888 ;
R Bernstein, Untersuchung fiber den Magnetismus einiger Gase und Ddmpfe, Halle, 1909.
§ 8. The Chemical Properties of Hydrogen
Although the combustibility of hydrogen is one of its most characteristic
properties, perfectly dry hydrogen ignites with difficulty, if at all, when mixed with
perfectly dry oxygen. According to H. B. Baker (1902), i the dried mixture may
be heated to the melting point of silver — 960*5° — without appreciable combination.
Note, however, that moisture is a product of the reaction. Many other combustible
substances, if perfectly dried, do not burn when moisture is rigorously excluded.
The moisture is here said to act as a catalytic agent — Kara, down ; Xvtn, I
loosen. Finely divided platinum and many other metals will cause a mixture
of hydrogen and oxygen, at ordinary temperatures, to explode ; and if a jet of
hydrogen, in air, impinges on finely divided platinum, the metal becomes hotter
and hotter, and finally ignites the gas.
" Toy " automatic cigar lighters are made so that by turning the tap of a little hydrogen
generator^ — ^not unlike the Kipp's apparatus in principle — a jet of hydrogen can be directed
on a piece of spongy platinum when a "light " is desired. The platinum becomes hotter
and hotter, and finally ignites the jet of hydrogen. The flRme is extinguished by turning the
stopcock, and the apparatus is ready for another ignition when the jet of hydrogen is again
turned on to the platinum. This is the principle of the self-lighting lamp designed by
J. W. Dobereiner in 1822. Impurities in the hydrogen gas, however, appear to " poison
the platinum, for the apparatus soon ceases to be effective.
Hydrogen and oxygen, so far as we can tell, may remain an indefinite time in
contact with one another at atmospheric temperatures without showing any sign
of chemical action. Some say that the gases do react, but very, very slowly. By
measuring the diminution in the speed of the reaction from, say, 600° to 500° to
400° to 300°, and assuming that the rate of diminution of the speed of the reaction
326 INORGANIC AND THEORETICAL CHEMISTRY
follows the same law — reduction of the speed by one half per 10° reduction of
temperature — it has been estimated that no appreciable amount of hydrogen
and oxygen will have combined if a mixture of these gases be allowed to
stand at ordinary temperatures 1,000000,000000 years. The student has the
option of accepting or rejecting statements Hke these. They can neither be proved
nor disproved. The risks which attend this mode of reasoning about natural
processes have already been indicated. As P. Duhem has said (1910) : " It comes
to the same thing experimentally whether we say that the velocity of a reaction
is absolutely null, or that it is so small that there is no way of detecting it."
Shortly after H. Cavendish's work on hydrogen gas, there were suggestions made
for using hydrogen lamps for heating purposes ; thus, J. Priestley stated that
oxygen could be used for producing a very high temperature if fed into the hydrogen
flame by means of a suitable bellows ; and F. L. Ehrmann, in his Versuch einer
Schmelzkunst mit Hiilfe der Feuerluft (Strassburg, 1785), described the effect of a
hydrogen flame, strengthened by admixture with oxygen, upon many substances.
About 1801, R. Hare ^ devised an oxyhydrogen blowpipe, which was fully described
in his Memoir of the supply and application of the hlowpi^pe (Philadelphia, 1802),
pubHshed by order of the Chemical Society of Philadelphia. The oxyhydrogen jiame
is one of the hottest gas flames known ; by its means Robert Hare melted barytes,
alimiina, and siUca and obtained products resembling white enamel. Magnesia and
platinum were melted ; and, added R. Hare, " had I sufiS.cient confidence in my own
judgment, I should declare that silver, gold, and platinum were thrown into a state of
ebullition by exposure on carbon to the gaseous flame." When a stick of quickhme is
placed at the tip of the flame from a mixture of hydrogen and oxygen burning from
a special jet to avoid risk of explosion, the lime does not melt, but it becomes white
hot and glows with an intense white light known as Drummond's light ^ or the
lime-light or the calcium light. T. Drummond said that the lime-light is of such
dazzHng whiteness that it is plainly visible sixty-eight miles away. If zirconia be
used in place of lime, the zircon light is obtained. According to T. Drummond, the
light given by zirconia is less powerful than that of lime ; and that by magnesia is
only half as intense. C. H. PfafE says that if the light of a wax candle be unity, the
light emitted by a cylinder of lime one-fifth the diameter of the flame of a candle is
153 when heated by the oxyhydrogen flame ; 76 by the ether-oxygen flame ; 69 by
the alcohol-oxygen flame ; and 19 by the oxygen-coal gas flame. A. Pleischl and
M. A. Gaudin studied the flame from oxygen and turpentine, and oxygen and oil
gas. The oxyhydrogen flame is used for the autogenous welding of metals, for
soldering platinum, for making vessels of fused quartz, etc.
Hydrogen unites chemically, directly or indirectly, with most of the non-metallic
and with many of the metallic elements — more particularly the alkali and alkaline
earth metals. The binary compounds of the metals with hydrogen are usually
called hydrides. A. L. Lavoisier, in his Traite elementaire de chimie (Paris, 1.
116, 1789), foresaw the probabiHty of the formation of hydrides, for he said:
As combustible substances have in general a great affinity for oxygen, they ought
likewise to attract, or tend to combine with each other ; quae sunt eadem uni tertio, sunt
eadem inter se ; and the axiom is found to be true. Almost all metals, for instance, are
capable of \m.iting with each other, and of forming what, in common language, are called
alloys. Sulphur, phosphorus, and carbon readily iinite with metals. Hydrogen is like-
wise capable of combining with many combustible substances. It is worthy of being
examined whether hydrogen in its concrete state, uncombined with caloric, be susceptible
of combination with sulphur, phosphorus, and the metals. There is nothing that we know
of which, a priori, should render these suppositions impossible ; for combustible bodies
being in general susceptible of combination with each other, there is no evident reason for
hydrogen being an exception to the rule. However, no direct experiment yet establishes
either the possibility or impossibility of the union. Iron and zinc are the most likely of
all metals for entering into combination with hydrogen ; but, as these have the property
of decomposing water, and as it is very difficult to get them entirely free from moisture in
chemical experiments, it is hardly possible to determine whether the small portions of
hydrogen gas obtained in certain experiments with these metals were previously combined
HYDROGEN 327
with the metal in the state of solid, or if they were produced by the decomposition of a
minute quantity of water. The more care we take to prevent the presence of water in
these experiments, the less is the quantity of hydrogen produced ; and when very
accurate precautions are taken, even that quantity becomes insensible.
The hydrides of the non-metals are usually, not always, more stable than the
hydrides of the metals. Water can be regarded as an oxygen hydride, as well as a
hydrogen oxide. Hydrogen gas reacts directly with fluorine, and this at tempera-
tures as low as —210°, when the fluorine is liquid, or even at still lower temperatures
where the fluorine is solid, and the hydrogen liquid. * The product of the reaction
is hydrogen fluoride, HF ; chlorine and bromine unite with hydrogen in light but
not in darkness, forming in the one case hydrogen chloride, HCl, and in the other,
hydrogen bromide, HBr. Iodine commences to unite with hydrogen at about 200°
forming hydrogen iodide, HI. Aqueous solutions of the hydrides of fluorine,
chlorine, bromine, and iodine are well-known acids — respectively called hydro-
fluoric, hydrochloric, hydrobromic, and hydriodic acid. Sulphur or selenium
reacts with hydrogen at about 250°, and tellurium at 400°, forming respectively
hydrogen sulphide, H2S, hydrogen selenide, H2Se, and hydrogen telluride, H2Te.
Hydrogen does not unite with nitrogen by direct heating, but it does So when
stimulated by electric sparks or the silent electrical discharge — ammonia gas,
NH3, is the product of the action. Ammonia, NH3, is a nitrogen trihydride ; it
is a well-known base. Similar compounds of phosphorus — ^phosphine, PH3 —
arsenic — arsine, ASH3 — and antimony — stibine, SbH3 — also have basic properties ;
they are obtained indirectly, since direct union, if it occurs at all, is so insignificant
when these elements are heated together that it is not at all certain if combination
has occurred at all. Carbon and hydrogen do not react at ordinary temperatures.
When an electric arc is formed between carbon electrodes in an atmosphere of
hydrogen, the two elements unite forming acetylene, C2H2, and traces of other hydro-
carbons— e.g. methane, CH4. Carbon begins to react with hydrogen, forming
methane CH4, at about 1200°. Carbon forms an extensive series of hydrides usually
called hydrocarbons — e.g. methane, CH4 ; ethylene, C2H4 ; acetylene, C2H2 ;
napthalene, CioHg ; anthracene, C14H10 ; etc. The direct formation of the siHcon
or boron hydrides by heating the two elements together has not been satisfactorily
demonstrated. Lithium metal burns in hydrogen gas forming lithium hydride,
LiH. The alkali and alkahne earth metals unite directly with hydrogen when
heated over 300° to form hydrides — e.g. potassium and sodium hydrides — KH and
NaH respectively ; calcium hydride, CaH2, is the active agent in hydrolith.
The hydrides of neodymium, praseodymium, cerium, yttrium, samarium, thorium,
lanthanum, iron, cuprous and cupric copper, and silver have also been reported
—the last three are said to have been formed by the reducing action of hypo-
phosphorous acid on solutions of the salts of the respective elements. The alleged
compounds are probably of the nature of the so-called palladium-hydrogen
alloy — -with occluded hydrogen.
The diflerent products of the action of hydrogen on the metals may be arranged
in three classes : (1) Compounds of the non-metals and metalloids — e.g. stibine,
SbHg — in which hydrogen seems to play the role of a positive univalent element,
while the other element behaves as a negative or non-metallic element. As a
rule, these compounds are gaseous at ordinary temperatures and pressures, or are
very volatile. (2) Compounds of the alkali and alkaline earth metals. They are
transparent and crystalline ; volatile without decomposition in an atmosphere of
hydrogen at comparatively low temperatures. They are probably hydrides proper,
being related to the chlorides and nitrides in that the hydrogen plays the role of
a non-metallic univalent element. (3) UnUke the two preceding types, these
products are metallic, and form soUd phases whose composition varies with external
conditions. The typical example is the palladium-hydrogen alloy.
D. P. Smith 5 has shown that if A. Werner's periodic arrangement of the elements
be employed, the metals which form the compounds of the first class are confined
328
INOEGANIC AND THEORETICAL CHEMISTRY
at the extreme right ; and in Table III they are marked ofi with heavy lines. The
compounds of the second class are confined to a group on the extreme left. The
central part of the table is occupied by (i) metals whose relation is unknown ;
(ii) metals which occlude appreciable amounts of hydrogen represented by symbols
in clarendon type ; and (iii) metals which do not occlude anything but a relatively
small amount of hydrogen — represented by symbols in brackets. The dot attached
to the symbol means that the available evidence is conflicting or doubtful. The
occluding elements thus appear to form a central group with subgroups on the left
and right.
Table
III. — A.- Werner's Periodic Table
Elements to
modified tc
Hydrogen.
SHOW THE Relation
OF
THE
H
Mn.
Fe
Bu
(Os)
Co
(Bh)
(Ir).
Ni
Pd
Pt
Cu
(Ag)
(Au)
(Be).
(Mg).
(Zn)
(Cd)
(Hg)
B
(Al)
Ga
C
Si
Ge
N
P
As
0
s
Se
Te
F
CI
Br
I
He
Li
Cr.
Ne
Na
A
K
Ca
Sc
Ti
V
Kr
Bb
Sr
(Y).
(Zr).
Nb
(Mo)
In
(Tl)
Sn
(Pb)
Sb
(Bi)
X
Cs
Ba
£a
Rare earth
metals
—
Ta
(W)
_
—
ThU
—
—
—
.^
—
—
—
—
—
—
,—
The action of hydrogen on oxides and salt solutions. — Hydrogen
gas reduces a great many metal oxides forming the metal and water : MO+H2
=H20H-M. The oxides of silver and palladium are reduced by hydrogen gas in the
cold. F. Wohler ^ found black palladious oxide to be reduced by hydrogen with
incandescence. The oxides of copper, lead, cadmium, tin, iron, cobalt, nickel,
antimony, etc., must be heated before reduction occurs. If the mixture of hydrogen
and metal oxide were confined in a closed vessel, the reaction would come to a stand-
still, but when the oxide is heated in a stream of hydrogen gas, the water vapour is
whisked away from the seat of the reaction before it has time to set up the back
reaction. The reduction of cupric oxide, CuO, to the extent of 1"7 per cent, can be
detected after 15 minutes' exposure to a stream of hydrogen between 87° and 90° ;
7 per cent., at 100° ; and the reduction is complete in 15 minutes at 150°. The
results differ with the physical condition of the oxide. Thus, according to
F. Glaser, yellow mercuric oxide shows signs of reduction at 75°, and th^ red oxide at
140°. The reduction of ferric oxide was similarly detected after 15 minutes' exposure
at 220° ; magnetic oxide at 290°, and ferrous oxide at 305°. The reduction of
manganese dioxide could be detected at about 145°, pyrolusite at about 190° ;
manganoso-manganic oxide, Mn304, at about 255° ; and with manganous oxide
no action could be detected at 600°-1300°. The temperatures at which
a reduction can be detected with lead dioxide is 140° ; lead monoxide, 190°-
195° ; cobalt sesquioxide, 110° ; cobalt monoxide, 165° ; nickel sesquioxide,
70° ; and nickel monoxide, 225°. In view of the relatively easy reduction
of, say, lead dioxide, Pb02, at 140°, and the more difficult reduction of the
monoxide, PbO, at 190°, it is highly probable that the higher oxide is first
reduced to the lower oxide, and a higher temperature is then needed for the sub-
sequent reduction of the monoxide to metal. This, for example, is the case with
Mn02->Mn304->MnO->Mn
145° 255° 1300°
and FegOa-^FegOi-^FeO-^Fe
220° 290° 305'^
HYDROGEN 329
The oxide to be reduced may be placed in a porcelain boat which is heated in a tube
through which a current of hydrogen is passing ; or it may be heated in a crucible
fitted with a perforated lid and tube through which a slow current of hydrogen
passes. This form of crucible — called a Rose's crucible — is used in analytical work.
For example, in certain analytical processes, cobalt is precipitated as hydroxide,
and after drying, is ignited in a Rose's crucible with a stream of hydrogen, and
finally weighed as metal.
According to M. Berthelot, the displacement of the oxygen from oxides, sulphur
from sulphides, etc., is readily effected when the reaction generates heat, but not
if the reaction absorbs heat. The greater the amount of heat developed, the more
readily will the reaction occur. Contrast the heat developed during the formation
of the three oxides : silver oxide, 7 Cals. ; lead oxide, 51 Cals. ; and zinc oxide,
85*4 Cals. The heat developed during the formation of steam is nearly 58 Cals.
Hence, with silver oxide, there will be an evolution of 51 Cals. during the reduction
to silver ; similarly, with lead oxide there will be an evolution of 7 Cals. ; but with
zinc oxide, on the contrary, there is no energy to spare— rather is there a deficiency
of 58 — 854 = — 27'4 Cals. This is supposed to explain how silver oxide can be
reduced at ordinary temperatures by hydrogen, while zinc oxide requires a high
temperature :
Silver oxide,
Copper oxide,
Lead oxide,
Iron oxide,
Zinc oxide,
AggO.
PbO.
FeO.
ZnO.
Heat of reduction
+ 51
+ 21
+ 7
+ 8-4
-27-4
Reduction begins at
0°
90°
190°
305"
—
The rule is not rigidly exact ; for one thing it takes no account of the variation
in the thermal value of the reaction with temperature. According to F. Glaser,
cadmium oxide shows signs of reduction at 282°, and zinc oxide at 454°.
Many metal chlorides and other salts, etc., are also reduced by hydrogen. Thus,
iron, silver, and palladium chlorides are readily reduced by hydrogen gas forming
hydrogen chloride and the metal ; antimony sulphide is reduced to antimony ; etc.
According to C. Brunner, when hydrogen is passed through solutions of salts of
platinum, or palladium, the metal is slowly precipitated ; with silver salts, the
reaction is far from complete since but a small fraction of the total silver is pre-
cipitated ; iridium salts are scarcely affected ; and gold and mercury salts are not
reduced. B. Renault (1873) and H. Pellet (1873) claimed that silver salts are not
reduced by thoroughly purified hydrogen and that the alleged reduction is a
secondary effect due to the presence of arsine,. silane, or other impurities ; but
J. W. Russell, N. N. Beketoff, and A. R. Leeds 'showed that reduction does take place
with the pure gas. If the temperature or the pressure be augmented, the reaction
may be completed. Thus, at the temperature of the water-bath, platinum, palla-
dium, rhodium, and iridium are completely precipitated ; and at 200 atm. pressure,
silver is completely precipitated. E. Schobig says the reduction also occurs in
darkness. Similar remarks apply to solutions of salts of nickel, cobalt, lead, bis-
muth, etc. The deposition of nickel is complete at 200° and 180 atm. pressure.
E. Reichardt investigated the colorations produced by hydrogen on paper treated
with silver nitrate ; J. B. Senderens, the influence of hydrogen on hot solutions of
silver nitrate ; J. W. Russell, the precipitation of platinum, palladium, and gold
from solutions of their salts, the reduction of cupric nitrate to nitrite, and the
formation of a basic salt with mercurous nitrate ; and F. C. Phillips, the influence
of hydrogen on solutions of platinum, and palladium chlorides. N. N. Beketoff,
J. Lowenthal, and W. Ipatjeff have studied the precipitation of silver, mercury,
copper, zinc, cadmium, lead, etc. Y^giV-solutions of silver and mercury salts give
the metals at room temperatures and 200 atm. pressure ; copper sulphate gives no
metal at 600 atm. ; y^jiV-copper nitrate gives copper at 200° and 600 atm. Nickel
salts behave similarly. Cobalt, lead, iron, and bismuth salts give the metals only
at high temperatures and pressures.
Hydrogen gas is not oxidized by {i.e. it does not reduce) solutions of ferric
330 INORGANIC AND THEORETICAL CHEMISTRY
chloride, potassium ferricyanide, nitric acid (specific gravity 1"42), chromic acid, or
aqua regia. R. Bunsen, however, stated that in darkness ferric chloride is reduced
by hydrogen to ferrous chloride. J. Milbauer studied the oxidation of hydrogen by
sulphuric acid, H2S04+H2=2H20+S02, under the action of different catalytic
agents at different temperatures. With pure hydrogen, there is no appreciable
effect at ordinary temperatures, but the action is quite marked at 140°. Potassium
permanganate is reported ^ to be gradually reduced in acid, neutral, or alkaline
solutions at ordinary temperatures. A. C. Vournasos found that the nascent
hydrogen liberated by heating dry sodium formate reacts with several elements
which do not combine directly with free hydrogen. For example, phosphine is
obtained when phosphorus vapour is passed over sodium formate melted at 200° ;
with a mixture of phosphorus and four times its weight of sodium formate, at
400° ; and with a mixture of sodium formate and neutral sodium phosphite, or
anhydrous disodium phosphate. Arsine is likewise obtained from sodium arsenite ;
stibine from sodium antimonide ; hydrogen sulphide from sulphur, from sodium
sulphide, or from mercury, lead, or tin sulphide ; nitrides give ammonia ; cyanides
give hydrogen cyanide ; alkali carbides give acetylene ; and silicon chloride or
sulphide give siHcon hydride.
Hydrogen is a far more vigorous reducing agent if it acts in the presence of finely-
divided metals — nickel, platinum, etc. — than when alone. Colloidal platinum or
palladium is more effective than even the finely-divided platinum or palladium
black. A solution of ferric chloride in the presence of platinum is reduced by
hydrogen gas. There is a continuous catalytic action of the platinum on the gas
and the layer of solution adhering to the metal. F. Lehmann found that osmium
dioxide acts as a catalytic agent on the reduction of oleic acid and liquid oils by
hydrogen. The finely-divided metals alone reduce neutral solutions of potassium
permanganate, and accordingly decolorize dilute solutions, and the solution at the
same time becomes alkaline, owing probably to the formation of potassium hydroxide
and a brown manganic oxide — Mn(0H)4, or possibly Mn0(0H)2 — in symbols :
2KMn04H-3H2+2H20->2KOH+2Mn(OH)4. Finely-divided gold, platinum, silver,
arsenic, antimony, tungsten, and all the common metals reduce dilute neutral
solutions of permanganate. Mercury also reduces permanganate, and D. Borar
(1911) represents the reaction by the equation: 2KMn04+3Hg+H20->2K0H
+2Mn02+3HgO. Whatever be the mechanism of the reaction, the observed fact
is that hydrogen can do its work much more quickly in the presence of platinum
black ; or, if the alternative statement be preferred, that the finely-divided metal
can do its work more quickly in the presence of hydrogen. The point is illustrated
by A. Smith's lecture experiment :
Three test tubes are filled with dilute acidified potassium permanganate solution. Zinc
dust added to the one generates hydrogen and causes decolorization ; a little platinum
black is added to the second, and hydrogen gas is led through this and the third solution.
The contact action of the platinum enables the hydrogen quickly to reduce the per-
manganate, while the third portion remains unaltered.
F. Kuhlmann showed that the presence of certain metals can also induce a reaction
between nitric oxide and hydrogen with the formation of ammonia ; ^ and
P. Sabatier and J. B. Senderens have studied the reaction between hydrogen and
many organic compounds — thus, carbon monoxide is transformed into methane ;
ethylene or acetylene is also hydrogenized to methane ; aldehyde or ketone to
alcohol ; benzene to hydrobenzene ; etc.
In 1839, T. de Saussure ^ observed that various organic substances (peas, corn,
humus) in the act of decomposition may excite the combination of hydrogen and
oxygen, and that mixtures of these gases " behave with fermenting substances
the same as with platinum." Several micro-organisms are also reported by
A. J. Nabokich and F. A. Lebedeif, N. Bronislaff, J. Nikitinsky, and B. Niklewsky
to possess the property of oxidizing hydrogen.
HYDROGEN 331
The nascent state — status nascens. — Hydrogen is not particularly active,
chemically, at ordinary temperatures, but when the gas is exposed to an elevated
temperature, or to an electrical discharge, the bonds which hold the atoms together
appear to be relaxed, and hydrogen is then a potent agent, for its reactivity is great.
The hydrogen which is generated by many exothermal reactions, at the moment
of its birth, in statu nascendi, appears to be in a specially active state, for it can in-
augurate many reactions which gaseous hydrogen cannot invoke. For instance,
hydrogen gas can be passed into an acidified solution of ferric chloride or into water
in which silver chloride is suspended without producing any appreciable change,
but if metallic zinc be placed in the acidified solution, the brisk evolution of hydrogen
is soon attended by the reduction of the ferric to ferrous chloride in the one case,
and of the silver chloride to metallic silver in the other. Hydrogenized palladium
or platinum ^^ can also do chemical work which ordinary hydrogen cannot do, for
instance, it can reduce a solution of ferric to ferrous chloride as illustrated by the
symbols : FeCl3+Hpaiiadium=I'eCl2H-HCl ; it can reduce solutions of chlorates
to chlorides ; nitrates to nitrites and ammonia ; mercuric to mercurous chloride ;
f erricyanides to f errocyanides ; sulphurous acid to sulphur and hydrogen sulphide ;
indigo-blue to indigo-white ; nitrobenzene, C6H5NO2, to aniline, C6H5NH2, and it
can unite with chlorine, iodine, and oxygen in the dark at ordinary temperatures.
Charcoal saturated with hydrogen can also reduce chlorates to chlorides, f erri-
cyanides to ferrocyanides, but not nitrates to nitrites. The main sources of nascent
hydrogen are the amalgams of sodium, magnesium, and aluminium : the copper-
zinc couple with water ; or the metals zinc, tin, and iron with dilute acids, or,
maybe, with alkaline solutions. Hydrogen iodide in a sealed tube at 150°-275°
gives nascent hydrogen which is particularly active on account of the elevated
temperature.il According to M. Berthelot, it reduces organic compounds to the
hydrocarbon stage in which the molecule contains the maximum amount of hydrogen
consistent with the quadrivalency of carbon, e.g. ethyl iodide and ethyl alcohol
form ethane ; glycerol forms propane ; etc.
It has been suggested, without proof, that the hydrogen in palladium is in
the atomic condition, for atoms are supposed to be more chemically active than
molecules because some preliminary work has to be done in order to split the mole-
cules into atoms before the reaction can occur, whereas atoms are ready to react
immediately. This hypothesis was suggested by A. Laurent in 1846. He ascribed
the greater activity of the elements in their nascent condition to their atomic con-
dition. He reasoned that in the molecules (HH) and (BrBr) the affinity of bromine
for bromine and of hydrogen for hydrogen is sufficient to prevent the one combining
with the other, whereas if atomic hydrogen be in contact with atomic bromine
combination sets in without the need for a preliminary rupture of the molecules of
the two elements. Consequently it was inferred that the atoms of hydrogen in their
nascent state do the work of reduction before they have spent part of their energy
in grouping themselves into molecules. P. A. Favre and J. T. Silbermann (1846)
adduced physical evidence in favour of A. Laurent's theory of the nascent state,
and showed that carbon burning in an atmosphere of nitrous oxide, N2O, develops
considerably more heat than when it burns in oxygen because the energy required
to decompose the molecules of ordinary oxygen is greater than the energy required
to decompose the molecules of nitrous oxide. G. Bodlander also explains the
greater activity of the hydrogen occluded in metals as due to its being there present
in the atomic condition.
D. Tommasi (1898) 12 has emphasized the fact that this explanation does not
make clear why nascent hydrogen from zinc and sulphuric acid is able to reduce
chlorates or bromates to chlorides or bromides respectively, while hydrogen from
sodium amalgam will not do so ; nor does it explain why zinc and hydrochloric acid
will reduce salts of vanadium pentoxide, V2O5, to the dioxide, V2O2, while magnesium
and hydrochloric acid furnishes the trioxide, V2O3, under similar conditions. Nitro-
benzene, CeHgNOg, can be reduced to the base aniUne, CeHgNHg, in acidic but not in
332 INORGANIC AND THEORETICAL CHEMISTRY
alkaline solutions — say by sodium amalgam and water. The base hydroxylamine is
formed by the reduction of nitric acid in acidic solutions, and hyponitrous acid by
reduction in alkaline solutions. It might be expected that if the work of reduction is
really performed by nascent hydrogen, this agent should possess the same properties
from whatever source it is derived ; whereas the fact is that the reducing power o£
nascent hydrogen varies according to the nature of the chemical reaction which
gives it birth. Hence, the greater activity of nascent hydrogen is probably caused
by some influence other than the supposed atomic condition of the element in
statu nascendi.
M. Berthelot considers that the disengagement of hydrogen in these reactions
is but the result of a secondary reaction, and that the main reaction is the formation
of a complex with the substance to be reduced. He calls these complexes les
systhnes reducteurs. In reducing potassium chlorate by zinc and sulphuric acid,
the system Zn : H2SO4 : KCIO3 with some water may be first formed, and this
passes into the system ZnSO^ : Hg : KCIO3 or ZnS04 : KCl : H2O, or into a system
consisting of all these products. There are many reactions which are attributed
to Vhydrogene naissant — e.g. reductions with sulphurous acid, ferrous hydrates,
ammonium hydrosulphide, etc. — in which hydrogen is never evolved, and in
which the reduction may be performed in les systemes reducteurs. In the case of
zinc and ferric chloride it might be then argued that the effects attributed to
Vhydrogene naissant are really produced by the direct action of the metal ; in
symbols : Zn+2FeCl3=ZnCl2+2FeCl2 ; and not through the consecutive reactions;
Zn+2H20=Zn(OH)2+2H ; and H+reCl3=HCl+FeCl2.
The energy which is set free during a reaction under ordinary conditions appears
as heat, and this may make the gas at the seat of the reaction more reactive : or,
before the energy has degraded to heat, it may be available for doing chemical
work so that the difference between the so-called nascent and ordinary hydrogen
lies in the greater availability of the energy of the former. D. Tommasi argues
that the reducing properties of nascent hydrogen depend on the nature of the reaction
from which it is derived, and that the greater affinity of hydrogen in statu nascendi
arises from the momentary association of hydrogen with n calories of energy liberated
by the reaction. Accordingly the term " nascent hydrogen " is synonymous with
H+w cals. The difference in the reducing properties of hydrogen produced by
different reactions is determined by the amount of heat liberated in the reaction.
For example :
HgSO^Aq.-f Cd-21-5 Cals. H2S04Aq. + Zn= 377 Cals.
2HC]Aq. + Zn=34-2 „ H2S04Aq.+Mg= 112-0 „
2HBrAq.-|-Zn=34-2 „ 2Na (amalgam) + Aq.- 180-0 „
The greater the thermal value of the reaction, the greater is the activity of the nascent
hydrogen. In some cases, the reduction is not due to IL-\-n cals., but to the metal
M+w cals. ; for example, in the reduction of potassium chlorate by zinc and dilute
sulphuric acid, the zinc unites with the oxygen of the chlorate, forming zinc oxide :
KC1034-3Zn=KCl+3ZnO. If a chlorate, dissolved in a dilute solution of sulphuric
acid, be electrolyzed with platinum electrodes, perchloric acid is formed at the anode,
and with a zinc anode, potassium chloride, but no perchlorate appears at the
electrode.
When hydrogen is developed on the surface of a metal immersed in a suitable
solvent, W. Ostwald 13 has pointed out that the gas in the smallest — probably
submicroscopic — bubbles must be under a great pressure owing to the surface tension
of the water. He calculates this pressure to be near 15,000 atm. for gas bubbles of
approximately molecular dimensions, 10~7 cm. diameter. It has been suggested
that the gas under this enormous pressure is more chemically active than when it
is under ordinary atmospheric pressures.
The decomposition voltage of sulphuric acid with a platinized platinum electrode
has a definite and specific value which is different from the value obtained when
HYDROGEN 333
electrodes with other metals — e.g. smooth platinum, zinc, lead, etc. — are used and the
difference between the voltage required to liberate hydrogen at the surface of a metal
and at the surface of a platinized platinum electrode under similar conditions has
been styled the overvoltage. The point where the evolution of gas begins may be
taken as the point where bubbles of hydrogen begin to form, or the point where a
break appears in the voltage curve. M. le Blanc (1890) takes the overvoltage as
the excess back electromotive force generated at the given electrode over that at
a platinum electrode when the charging circuit is rapidly closed and as rapidly
broken making a new circuit through a commutator. The overvoltage depends
on the nature of the metal, current density, temperature, etc. W. A. Caspari ^^
found that with platinized platinum the overvoltage amounts to 0*005 volt ; with
smooth platinum, 0'08 volt ; meaning that the formation of gas bubbles could be
first observed on the electrode in iV-sulphuric acid, at these voltages. With iron,
sodium hydroxide was used. The overvoltages with some other metals are :
Au. Fe. Ag. Ni. Cu. Pd. Cd. Sn. Pb. Zn. Hg.
0-02 0-08 015 0-21 0*23 0-46 0*48 0-53 0-64 070 0*78 volt.
A. Coehn and K. Dannelberg, E. Miiller, and A. Thiel and E. Breuning, obtained
the same results with rather smaller voltages, but the order was the same as W. A.
Caspari's. These numbers may be taken to represent the potential at which the
hydrogen is liberated at the different electrodes. There is a similar overvoltage
at the oxygen anode — with platinized platinum, 0'39 volt ; polished platinum, 0*62 ;
palladium, 0*39 ; and gold, 0*59 volt. G. Carrara found the overvoltages quite
different in alcoholic and aqueous solutions. If there were no overvoltage, water
could be decomposed by a current of 1*22 to 1*23 volts, but the voltage actually
required is 1*23 plus the overvoltages at the two electrodes.
When hydrogen separates on a metal electrode, its chemical energy is augmented
by the overvoltage beyond what it would possess if it were in the gaseous state under
ordinary atmospheric pressure. The hydrogen from, say, a mercury or lead cathode
is evolved at a higher potential and therefore possesses greater energy than from a
platinum cathode. Lead or mercury thus possesses, in a high degree, the property
of rendering difficult the escape of hydrogen from the cathode, so that many electro-
lytic reductions are possible with these metals — e.g. of caffein or uric acid — which are
not produced by cathodes of other metals. While the reducing power and hydrogen
overvoltage generally run parallel, E. Miiller, J. Tafel and K. Neumann, C. F.
Bohringer, and A. Chilesotti have shown that it does not necessarily follow that all
reductions take place more readily at the cathode with the higher overvoltage.
For instance, nitrites are more readily reduced than nitrates at cathodes of zinc,
iron, lead, platinum, or gold, while nitrates are more readily reduced at cathodes
of spongy copper or silver, and at a mercury cathode in hot solutions. Caffein is
more readily reduced at a mercury than at a lead cathode, and the converse obtains
in the reduction of succinimide. Obviously, therefore, factors other than hydrogen
alone are involved, and these may mask the relation between the hydrogen over-
voltage and high reducing power. For example, one substance may be more
readily absorbed by one cathode than another, so that the effective concentrations
at the cathode are different ; the electrode may act as a catalytic agent on the
reaction, or apparently inert substances in the solution may modify the overvoltage.
J. Tafel has shown that traces of certain metals exercise what may be called a
poisonous influence on other metals in reducing the overvoltage or swper- voltage of
the hydrogen and so decreasing the effectiveness of the lead and mercury cathodes ;
thus, 0*004 mgrm. of platinum per 10 sq. cm. of lead cathode surface will prevent
many electrolytic reductions possible in its absence. Silver, tin, copper, mercury,
zinc, and iron are also enemies of electro-reduction. J. Tafel and B. Emmert found
that the toxic effect cannot be attributed to the formation of a skin of the metal
on the surface of the cathode.
Various hypotheses have been devised to explain what is taking place when
334 INORGANIC AND THEORETICAL CHEMISTRY
an overvoltage occurs. F. Haber suggested that the formation of films of gas on
the electrodes increases the resistance ; H. G. Moller considered the overvoltage
to represent the energy required to give a film of gas thick enough to generate
bubbles ; F. Foster suggested an oxide is formed at the anode, and a hydride or
soHd solution of a hydride at the cathode ; while W. Nernst and J. Tafel believe
that the effect is due to the slowness with which the electrode gets into equilibrium
with the surrounding conditions, for metals with a slight tendency to occlude gases
require energy to force the gas into the electrode and thus produce high voltages.
J. Tafel, E. Miiller, G. N. Lewis and R. E. Jackson and others have suggested that
the discharge of the H'-ions at the cathode, 2H*=H2 takes place in two stages,
H'=H and 2H=H2, so that monatomic hydrogen acts as an intermediate com-
pound, and while the reaction H''=H takes place quickly, the reaction 2H=H2
which causes the polarization of the electrode is slow. The potential of the hydrogen
electrode thus depends upon the concentration of monatomic hydrogen. C. W.
Bennett and J. G. Thompson, and W. D. Bancroft also favour an hypothesis based
on these reactions. D. Reichinstein ^^ attributes the poisoning of the electrodes
by the presence of certain impurities to the lowering of the overvoltage by increasing
the rate of conversion of the monatomic to ordinary hydrogen. Similarly also,
the effect of certain impurities —iron, cobalt, nickel, etc. — on the yield of caustic
soda by the amalgam process, observed by J. W. Walker and C. S. Paterson, is
attributed to the decrease in the overvoltage by the impurities acting catalytically in
increasing the speed of the conversion of monatomic into ordinary hydrogen. Similar
explanations have been applied to 0. Aschan's observation that impurities in the
sodium amalgam diminished the yield in the hydrogenation of benzoic acid ; to
E. Bamberger's observation of the great differences in the reducing power of different
samples of zinc dust ; and to G. Fernekes' observation that alcohol and many other
organic substances augmented the speed of the reaction between sodium amalgam
and water. The so-called nascent hydrogen is thus assumed to consist of electrically
neutral monatomic hydrogen, and the difference in the nascent hydrogen
derived from different sources is due to the difference in the effective concentration
of the monatomic hydrogen, which in turn is determined by the rate of conversion
of the monatomic into ordinary hydrogen.
Hydrogen is not a poisonous gas.i^ When small animals are placed in hydrogen,
they are " drowned," suffocated for want of oxygen. The injurious action of
hydrogen is therefore negative ; the gas acts merely by preventing access of oxygen
to the lungs. According to C. W. Scheele, F. Fontana, and H. Davy, the pure gas
excites disagreeable sensations and loss of muscular power, but when mixed with
air, it may be breathed a longer time. The violent symptoms described by G. Cardone
as attending the respiration by hydrogen must have arisen from impurities in the
gas. When hydrogen is inhaled, the voice becomes shrill — approaching falsetto.
The pitch of organ pipes and other wind instruments is raised if a blast of hydrogen
be used in place of air. According to F. Hatton, hydrogen has scarcely any appre-
ciable influence on bacteria, but M. Berghaus says their power of multiplying becomes
smaller in an atmosphere of hydrogen gas.
The detection and determination of hydrogen. — The determination of
hydrogen in the presence of methane and other hydrocarbons is based on the con-
traction which occurs after complete combustion with oxygen by explosion in a
eudiometer ; by fractional combustion with air in contact with palladium-asbestos
— methane does not oxidize at 100°, hydrogen does ; by absorption by spongy
palladium ; by the reduction of or absorption in palladium chloride — dry or in
solution — after the removal of olefine gases and carbon monoxide because these
gases also reduce palladious chloride. A nearly neutral one per cent, solution of
palladious chloride completely absorbs small quantities of hydrogen in a few hours ;
the absorption is retarded by using strongly acid solutions. According to F. C.
Phillips,i7 a one per cent, solution of palladious chloride is reduced in the cold by
hydrogen gas. 2000^^ P^-rt of hydrogen can be detected in a gas by this reaction.
HYDROGEN 335
The presence of hydrogen gas can be established by spectrum analysis. In the
absence of other hydrogen compounds, the gas can be passed over red-hot copper
oxide whereby water is formed. A. Gautier found that one part of hydrogen in
5000 of air is completely oxidized when passed over a layer of cupric oxide about
70 cm. long at the rate of 2 or 3 litres per hour ; with a tube 30 cm. long, about
70 per cent, of the hydrogen is oxidized. This reaction enables the amount of
hydrogen in dry air to be determined. C. Paal and C. Amberger used a colloidal
palladium solution for the absorption of hydrogen ; C. Paal and W. Hartmann
used a colloidal solution of palladium containing sodium picrate. A. LidofE
described a method for the volumetric determination of hydrogen by combustion
with magnesium powder ; T. Zerewitinoff used magnesium-methyl iodide. A.
Jacquelin proposed to absorb hydrogen from a mixture of methane, CH4, and
ethylene, C2H4, by sodium or potassium at about 300°. C. Zengelis found that
palladium or platinum absorbs hydrogen and the product colours blue a solution
made by dissolving a gram of molybdenum trioxide in dilute sodium hydroxide,
acidifying the solution with hydrochloric acid, and diluting the mixture to 200 c.c.
The atomic weight and valency of hydrogen. — Hydrogen was for a long time
the standard unit for the atomic weights ; but now oxygen =16 is the generally
accepted standard. Reports of the atomic weight of hydrogen, determined
through the ratio 0 : H lie between 16:0-99937 and 16:1-0087; and the best
representative value for the atomic weight of hydrogen is taken to be 1*008. The
molecular weight of the gas is then 2'016 — the molecule being therefore diatomic, H2.
According to W. Vaubel,!^ the molecular weight of hydrogen in the liquid state is
nearly 5-2. The valency of hydrogen is the unit for evaluating the valencies of the
other elements. R. de Forcrand i^ has favoured the assumption that hydrogen
is best regarded as a bivalent element ; this would entail doubling the valency of
all the other elements. The doubled scheme is said to offer some advantages in
dealing with the constitution of the subhalides and suboxides. There is a difficulty
already discussed, in assigning a place for hydrogen in the periodic table.
The quantities of some of the metals which are equivalent to a definite quantity
of oxygen have been already determined, and the results agree with the quantities
of the different metals found to be chemically equivalent to one gram of hydrogen.
One gram of a given metal dissolving in a suitable acid will always displace the same
amount of hydrogen whatever be the reacting liquid used — e.g. aluminium in sodium
hydroxide, in sulphuric acid, or in hydrochloric acid ; but the amounts furnished by
different metals are different. The weight of a metal required to displace one gram
of hydrogen is called the hydrogen equivalent of the metal. J. D. van der Plaats 20
determined the equivalent of zinc by measuring the volume of hydrogen evolved
by the dissolution of a given weight of the metal in sulphuric acid, and found that
13-8758 grms. of zinc gave 2-3767 litres of hydrogen, so that if a litre of hydrogen
weighs 0-89872 grm. under standard conditions, one gram of hydrogen is equivalent
to 64*89 grms. of zinc. J. W. Mallet 21 also determined the hydrogen equivalent
of aluminium by dissolving it in a solution of sodium hydroxide ; and E. Kohn-
Abrest, by dissolving it in hydrochloric acid. J. Torrey likewise determined the
hydrogen equivalent of iron. As a mean of twenty-one determinations, J. Thomsen
found the ratio H : Al=0-111902 ±0-000015 ; and he also found the ratio 0 : Al
=0-8878710-000018, from which it follows that 0 : H=7-9345± 0-0022.
The dissociation of the hydrogen molecule. — According to I. Langmuir
(1912),22 the electrical energy re qui red to maintain a tungsten wire at a given tempera-
ture in hydrogen increases at an abnormal rate with temperature, so that while the
theoretical value calculated from the heat losses by hot wires agrees with observa-
tions up to 1900° K., after that, there is a rapid increase until, at 3500° K., it is four
times the calculated value. No secondary electrical effects could be detected.
At small pressures, the heat losses are greater than at atmospheric pressures, pre-
sumably because of dissociation. This phenomenon is not observed with the other
gases tried. When a tungsten wire is heated to 1300°-2500° in hydrogen under a
336 INORGANIC AND THEORETICAL CHEMISTRY
pressure of O'OOl to 0*02 mm. , the gas slowly disappears. Hydrogen is not absorbed by
the wire to any great extent, but is deposited on the glass bulb, if the latter be cooled
by liquid air. When the liquid air is removed and the wire cooled, the hydrogen
is liberated. Langmuir explains the phenomenon on the assumption that hydrogen
is dissociated into atoms by the hot wire, and that some atoms diffuse into the tube
and are condensed in the same condition on the glass walls. The hypothesis that
the hydrogen is in the atomic condition is used to explain the greater chemical
activity of the gas, for, if phosphorus is present in the bulb, the two combine,
forming phosphine (phosphorus hydride), a reaction which has not been observed
with ordinary hydrogen under ordinary conditions ; the hydrogen also reacts with
oxygen at room temperatures ; tungstic oxide, WO3, and platinic oxide, Pt02, are
chemically reduced by the dissociated gas ; copper oxide is reduced almost imme-
diately to metallic copper ; ferric oxide is reduced to a lower oxide ; zinc oxide is
not reduced excepting over long periods of time when the colour of the dry zinc
oxide gradually turns grey, although silica, close beside it, remains quite white.
T. Ismardi (1915) has shown that the degree of dissociation a of hydrogen into
atoms is 0'557 at 3200° K. and 25 mm. pressure ; and 0*5 at the same temperature
and 50 mm. pressure. Similarly, at 3100°, the degree of dissociation is 0*255 at
100 mm. pressure, and 0*230 at 200 mm. pressure. The degree of dissociation a
was calculated from the expression : log {a^—l)-'^ =—Q/RT~log T+log^— 2'95,
where Q represents the heat of dissociation which is 64,000 cals., or 2H=H2
+64,000 cals. ; R is the gas constant ; and p the pressure. Assuming the heat of the
reaction at 2500° K. to be 64,000 cals., the free energy of the reaction 2H=H9 at
T° K. is given by G. N. Lewis and M. Randall as J=61,000-3'5T log T+0'00045r2
+20'2T. E. Briner estimates the heat of formation 2H=H2 at 2427° as 130 Cals.,
and the equilibrium constant as 0*10.
References.
1 H. B. Baker, Proc. Chem. Soc, 18. 40, 1902.
2 R. Hare, Tilloch's Phil Mag., 14. 238, 298, 1802 ; Trans. Amer. Phil. Soc, 6. 90, 1804 ;
Amer. Journ. Science, (1), 2. 281, 1820 ; T. Skidmore, ib., (1), 5. 327, 1822 ; B. Silliman, ib., (1),
1. 97, 1818 ; Amer. Min. Journ., 199, 1812 ; Trans. Amer. Phil. Soc, 11. 328, 1812 ; J. Newmann,
Journ. Roy. Inst, 1. 65, 1816 ; E. D. Clarke, ib., 2. 104, 1816 ; Ann. Chim. Phys., (2), 3. 39,
1812 ; The Gas Blowpipe, London, 1819 ; J. Priestley, Experiments and Observations on Different
Kinds of Air, London, 2. 124, 1775 ; G. G. Schmidt, Gilbert's Ann., 66. 84, 1820 ; W. A. Lampadius,
Schweigger's Journ., 19. 319, 1817 ; C. Ridolfi, ib., 20. 218, 1817 ; C. H. Pfaff, ih., 22. 385, 1818 ;
B. Hermann and C. Bischof, ib., 56. 123, 1829; A. Chodkiewieez, Scherer's Ann., 3. 248, 1799 ;
T. Cooper, ib., 5. 245, 1800 ; M. Hiibenthal, ib., 5. 244, 1799 ; P. Parrot, ib., 3. 239, 1799 ; 7. 280,
1801 ; J. 0. N. Rutter, Phil. Mag., (3), 1. 470, 1832 ; E. T. Hemming, ib., (3), 1. 32, 1832 ; J. F.
Daniell, ib., (3), 2. 57, 1833 ; C. Bischof, Journ. prakt. Chem., (1), 14. 129, 1819 ; J. Watt, Ann.
Phil, 11. 386, 1818 ; H. B. Leeson, ib., 14. 234, 1819 ; J. Holme, ib., 8. 471, 1816 ; J. T. Beale,
ib., 9. 252, 481, 1817 ; E. D. Clarke, ib., 9. 89, 163, 193, 326, 1817 ; 10. 373, 1817 ; G. Gray,
ib., 9. 479, 1817 ; T. S. Booth, ib., 10. 67, 1817 ; T. Osbrey, ib., 10. 366, 1817 ; R. W. Barchard,
ib., 10. 66, 1817 ; J. L. Gay Lussac, ib., 14. 320, 1830 ; Amer. Journ. Science, (1), 3. 87, 1821 ;
E. F. Smith, Chemistry in America, New York, 1914 ; The Life of Robert Hare, Philadelphia,
1917 ; A. Pleischl, ZeiL Phys. Mat, 1. 390, 1826 ; M. A. Gaudin, Compt Rend., 6. 681, 1838 ;
Journ. praht Chem., (1), 16. 54, 1839.
3 T. Drummond. Edin, Journ. Science, 5. 319, 1826 ; C. H. PfaflF, Pogg. Ann., 40. 547, 1837 ;
M. A. Gaudin, Compt Rend., 6. 861, 1835 ; A. Pleischl, Zeit Phys. Mat, 1. 390, 1826.
4 H. Moissan and J. Dewar, Bull Soc. Chim., (3), 17. 932, 1897 ; Compt Rend., 136. 641,
785, 1903.
5 D. P. Smith, Journ. Phys. Chem., 23. 186, 1939.
« P. MuUer, Zeit Chem., (2), 5. 507, 1869 ; F. Wohler, Liebig's Ann., 174. 60, 1874 ; C. R. A.
Wright and A. P. Luff, Journ. Chem. Soc, 33. 1, 504, 1878 ; A. Stromeyer, Pogg. Ann., 6. 471,
1826 ; G. Magnus, ib., 6. 509, 1826 ; F. Glaser, Zeit anorg. Chem., 36. 1, 1903 ; C. Brunner, Pogg.
Ann., 122. 153, 1864 ; R. Bunsen, Liebig's Ann., 146. 265, 1868 ; N. N. Beketoff. Compt Rend.,
48. 442, 1859 ; 79. 1413, 1874 ; B. Renault, ib., 76. 384, 1873 ; H. Pellet, ih., 77. 112, 1873 ;
78. 1132, 1874; 78. 1132, 1874; J. Lowenthal, Journ. prakt Chem., (1), 79. 480, 1860;
E. Schobig, ib., (2), 14. 289, 1876 ; N. Zelinsky and N. Glinka, Ber., 44. 2305, 1911 ; J. W. Russell,
Chem. News, 28. 277, 1874; Journ. Chem. Soc, 27. 3, 1874; A. R. Leeds, Ber., 9. 1456, 1876;
T. Poleck and K. Thiimmel, ib., 16. 2345, 1883 ; E. Reichhardt, Arch. Pharm., (3), 21. 585, 1883 ;
F. C. Phillips, Amer. Chem. Journ., 16. 255, 1894 ; E. D. CampbeU and E. B. Hart, ib., 18- 294,
HYDROGEN 337
1896 ; J. B. Senderens, Bull Soc. Chim., (3), 15. 991, 1896 ; J. Milbauer, Zeit. phya. Chem., 57.
649, 1907; 72. 380, 1911.
' J. A. Wanklyn and W. J. Cooper, Phil. Mag., (5), 30. 431, 1890 ; V. Meyer and M. von
Recklinghausen, Ber., 29. 2549, 2828, 1896 ; F. Jones, Journ. Chem. Soc, 33. 95, 1878 ; L. Ubbe-
lohde and L. Woromn, Petroleum, 7. 334, 1912 ; F. Lehmann, Arch. Pharm., 2lb\. 152, 1913 ;
A. C. Vournasos, Compt. Rend., 150. 464, 1910 ; Ber., 43. 2264, 2272, 1910 ; W. Foster, Chem.
News, 115. 73, 1917 ; H. B. Giles, ib., 15. 204, 1867 ; D. Borar, Journ. Chem. Soc, 99. 1414,
1911 ; A. Smith, General Chemistry for Colleges, London, 360, 1916 ; V. Meyer and H. Hirtz,
Ber., 29. 2828, 1896 ; J. Lowenthal, Journ. prakt. Chem., (1), 79. 480, 1860 ; W. IpatjefF, Journ.
Russian Phys. Chem. Soc, 41. 769, 1909 ; 43. 946, 1746, 1911 ; 44. 17J2, 1912 ; Ber., 42. 2078,
1909 ; 44. 1755, 3452, 1911 ; 45. 3226, 1912 ; S. Fokin, Journ. Russ. Phys. Chem. Soc., 42. 1074,
1910 ; P. Sabatier, Ber., 44. 1984, 3180, 1911 ; A. Skita and H. Ritter, ib., 43. 3393, 1910 ; ib.,
44. 668, 1911 ; F. W. Hinrichsen and R. Kempf, ib., 45. 2106, 1912 ; C. Paal, Ber., 45. 2221, 1912 ;
A. Brochet, Compt. Rend., 158. 1351, 1914 ; Bull. Soc. Chim., (5), 15. 554, 1914 ; N. Zelinsky and
B. Schtscherback, Journ. Russ. Phys. Chem. Soc, 44. 1880, 1913; N. Zelinsky. Ber., 44. 2779, 1911 •
H. H. Frank, ib., 37. 950, 1913 ; Ber., 45. 3595, 1912 ; H. Vosswinkel, Chem. Ztg., 37. 489, 1913 ;
C. Paal, Ber., 44. 1013, 1911 ; A. Karl, ib., 46. 3069, 1913 ; E. Windisch, ib., 46. 4010, 1913 ; A. C.
Chapman and H. D. Law, Analyst, 32. 250, 1907 ; R. Willstatter and others, Ber., 41. 2199,
1908 ; ib., 44. 3423, 1911 ; ib., 45. 1471, 1912 ; A, Skita and W. A. Meyer, ib., 45. 3312,
3579, 3589, 1912; H. Heymann, Zeit. phys. Chem., 81. 204, 1913; H. Foumier, Bull. Soc
Chim., (4), 7. 23, 1910 ; G. Varon, Ann, Chim. Phys., (9), 1. 144, 1914 ; A. Skita, Ber., 42. 1627,
1909 ; C. Paal, ib.y 41. 2282, 1908 ; German Pat., D.R.P., 236488, 1911; P. Breteau, Bull. Soc
Chim., (4), 7. 733, 764, 1911 ; H. Wieland, Ber., 45. 2615, 1912 ; N. V. Jurgens, German Pat.
D.R.P., 272340, 1914 ; P. Sabatier, J. B. Senderens, and A. Mailhe, Ann. Chim. Phys., (8),
4. 319, 1905 ; (8), 16. 70. 1909 ; D. Gauthier, ib., (8), 16. 289, 1909 ; A. Wohl and B. Mylo,
Ber., 45. 322, 1912 ; H. Wieland, Ber., 45. 484, 1912 ; 0. Starke, ib., 46. 2335, 1913 ;
J. Milbauer, Zeit. phys. Chem., 57. 380, 1911 ; N. Zelinsky and N. Glinka, Ber., 44. 2305,
1911.
8 F. Kuhlmann, Liebig's Ann., 29. 272, 1839 ; P. Sabatier and J. B. Senderens, Compt. Rend.,
\2A. 1358, 1897 ; 128. 1173, 1899 ; 130, 1761, 1900 ; 131. 40, 1900 ; 132. 1254, 1901 ; 133.
321, 1901 ; 134. 514, 1902 ; 135. 225, 1902.
» T. de Saussure, Mem. Soc. Phys. Hist. Nat. Geneve, 8. 163, 1839 ; Journ. prakt. Chem., (1),
14. 152, 1838 ; C. F. Schonbein, ib., (1), 89. 344, 1863 ; A. J. Nabokich and F. A. Lebedeff, Zeit.
Land. Vero. Oest., 8. 789, 1905 ; N. Bronislaff, Anz. Akad. Cracow, 911, 1906 ; J. Nikitinsky,
Centrb. Bakt. Parasit., 16. ii, 681, 1906 ; B. Niklewsky, ib., 20. ii, 469, 1908.
" J. H. Gladstone and A. Tribe, Journ. Chem. Soc, 33. 306, 1878.
11 L. Cohn, Arbeitsmethoden fur organisch-chemische Laboratorien, Hamburg, 1907 ; M. Berthe-
lot. Bull. Soc Chim., (2), 7. 53, 1867.
12 D. Tommasi, Ber., 11. 345, 1878 ; 12. 1701, 1879; Chem. News, 40. 245, 1879; 41. 1,
176, 1880 ; Traiti theorigue et pratique de V electrochimie, Paris, 105, 1889 ; Journ. Phys,
Chem., 1. 555, 1897 ; R. Franchot, ib., 1. 75, 1897 ; Bull. Soc Chim., (2), 38. 148, 1882 ;
(3), 17. 961, 1897 ; Monit. Scient., (3), 8. 829, 1878 ; T. L. Phipson, Chem. News, 40. 184, 1879 ;
5. Kern, ib., 31. 112, 1875 ; J. Thomsen, Ber., 12. 2030, 1879 ; G. Bodlander, Ueber langsame
Verbrennung, Stuttgart, 427, 1889.
1^ W. Ostwald, Lehrbuch der allgemeinen Chemie, Leipzig, 2. ii, 685, 1902.
1* W. A. Caspari, Zeit. phys. Chem., 30. 89, 1899 ; A. Coehn and K. Dannenberg, ib., 38.
609, 1901 ; F. Haber and R. Russ, ib., 47. 257, 1904 ; J. Tafel and K. Neumann, ib., 50. 713,
1905 ; E. Brunner, ib., 56. 331, 1906 ; M. le Blanc, ib., 5. 469, 1890; G. Carrara, ib., 69. 75, 1909;
J. Tafel, ib., 34. 187, 1900 ; 50. 641, 713, 1905 ; J. Tafel and B. Emmert, ib., 52. 349, 1905 ;
A. Thiel and E. Breuning, Zeit. anorg. Chem., 83. 329, 1913 ; E. MiiUer, ib., 26. 11, 1901 ; Zeit.
Elektrochem., 9. 955, 1905 ; 11. 509, 681, 1907 ; 14. 429, 1908 ; D. A. Maclnnes and L. Adler,
Journ. Amer. Chem. Soc, 41. 194, 1919 ; W. Thomsen, Chem. News, 99. 157, 1909 ; A. Besson
and L. Foumier, Compt. Rend., 150. 1752, 1910 ; C. N. Otin and B. Waser, German Pat.
D.R.P., 235955, 1911 ; G. Kolsky, ib., 257559, 1913 ; C. F. Bohringer, Zeit. Elektrochem., 12.
745, 1906 ; A. Chilesotti, ib., 12. 146, 173, 197, 1906 ; F. Haber, ib., 8. 539, 1902 ; F. Haber and
R. Russ, Zeit. phys. Cliem., 47. 257, 1904 ; H. G. MoUer, ib., 65. 226, 1908 ; Ann. Physik., (4),
25. 725, 1908 ; F. Foster, Zeit. Elektrochem., 16. 353, 1910 ; Zeit. phys. Chem., 69. 236, 1909 ;
G. N. Lewis and R. F. Jackson, ib., 56. 207, 1906 ; Proc Amer. Acad., 41. 399, 1906 ; C. W.
Bennett and J. G. Thompson, Trans. Amer. Electrochem. Soc, 29. 269, 1916 ; W. D. Bancroft,
ib., 29. 301, 1916; E. Newbery, Journ. Chem. Soc, 105. 2419, 1914; 107. 1051, 1066, 1107, 1359,
1916; 109. 470, 1917; Mem. Manchester Lit. Phil. Soc, 60. 11, 1916; D. A. Mclnnes and A. W.
Contieri, Proc Nat. Acad. Science, 5. 321, 1919; Journ. Amer. Chem. Soc, 41. 194, 2013, 1919;
E. A. Harding and D. P. Smith, ih., 40. 1530, 1918; 41. 1892, 1897, 1919; E. Newbery, ib., 41.
1887, 1895, 1919.
15 D. Reichinstein, Zeit. Elektrochem., 16. 927, 1910 ; J. W. Walker and C. S. Paterson, Trans.
Amer. Electrochem. Soc, 3. 185, 1903 ; W. D. Bancroft, ib., 29. 301, 1916 ; 0. Aschan, Ber., 24.
1865, 1891 ; E. Bamberger, ib., 27. 1548, 1894 ; G. Fernekes, Journ. Phys. Chem., 7. 611, 1903 ;
F. M. Frederiksen, ib., 19. 696, 1916 ; G. N. Lewis and R. F. Jackson,, Proc. Amer. Acad., 41»
399, 1906 ; Zeit. phys. Cliem., 56. 207, 1906.
i« C. W. Scheele, CrelVs Ann., 2. 229, 291, 1785 ; H. Davy, Elements of Chemical Philosophy,
VOL, I. Z
338 INORGANIC AND THEORETICAL CHEMISTRY
London, 1812 ; F. Fontana, Recherches physique sur la nature de Vair nitreux et de Vair dephlogis-
tique, Paris, 1776 ; G. Cardone, Quart. Journ. Science^ 20. 393, 1825 ; F. Hatton, Jourii. Chem.
Soc.y 39. 247, 1881 ; M. Berghaus, Arch. Hygiene, 62. 172, 1907.
1' F. C. PhUlips, Amer. Chem. Journ., 16. 406, 1895 ; E. E. Reid, ih., 47. 416, 1912 ; E. D.
Campbell and E. B. Hart, ib., 18. 294, 1896 ; W. Hempel, Ber., 12. 636, 1879 ; A. JacqueUn,
Ann. Chim. Phys., (2), 74. 203, 1840 ; A. Gautier, ib., (7), 22. 5, 1901 ; C. Zengelis, Zeit. anal.
Chem., 49. 729, 1910 ; O. Brunck, Chem. Zig., 34. 1313, 1331, 1910 ; C. Paal and C. Amberger,
Ber., 43. 343, 1910 ; C. Paal and J. Genim, ib., 41. 808, 1908 ; C. Paal and W. Hartmann, ib.,
42. 3930, 1909 ; A. LidoflF, Journ. Russian Phvs. Chem. Soc, 39. 195, 208, 1907 ; T. Zerewitinoff,
Ber., 41. 2233. 1908; 43. 3590, 1910; 45. 2384, 1912; E. Erdmann and F. Bedford, ib., 42.
1324, 1909 ; G. Abelmann, ib., 47. 2935, 1914 ; B. Oddo, ib., 44. 2048, 1911 ; Oazz. Chim. ItaL,
41. 709, 1911 ; H. Hibbert, Journ. Chem. Soc, 101, 328, 1912 ; V. Nesmjeloff, Zeit. anal. Chem ,
48. 232, 1909 ; G. von Knorre, Chem. Ztg., 33. 717, 1909 ; J. A. A. Auzies, Bull. Soc. Chim., (4),
9. 814, 1919 ; P. Lebeau and A. Damiens, ib., (4), 13. 366, 1913 ; Compt. Rend., 156. 144, 325,
489, 1913 ; W. Hempel, Zeit. angew. Chem., 25. 1841, 1912 ; L. Vannino and A. Schinner, ib., 26.
55, 1913 ; J. P. Wibaut, Chem. Weekbl, 11. 489, 1914.
18 W. Vaubel, Journ. prakt. Chem., (2), 57. 337, 1898 ; (2), 59. 246, 1899.
i» R. de Forcrand, Compt. Rend., 140. 764, 1905.
2» J. W. Mallet, Amer. Chem. Journ., 12. 205, 1890 ; H. N. Morse and E. H. Keiser, ib., 6.
347, 1884 ; J. Torrey, ib., 10. 74, 1888 ; J. D. van der Plaats, Compt. Rend., 100. 52, 1885 ; H. C.
Reynolds and W. Ramsay, Journ. Chem. Soc, 51. 854, 1887.
" J. W. MaUet, Phil. Tran^., 111. 1003, 1880 : E. Kohn-Abrest, BvU. Soc. Chim., (3), 33. 121,
1905 ; J. Thomsen, Zeit. phys. Chem., 13. 398, 1894 ; Zeit. anorg. Chem., 12. 1, 1896 ; J. Torrey,
Amer. Chem. Journ., 10. 74, 1888.
22 I. Langmuir, Trans. Amer. Electrochem. Soc, 20. 225, 1911 ; Journ. Amer. Chem. Soc, 34.
860, 1912 ; 35. 931, 1913 ; I. Langmuir and G. M. J. Mackay, ib., 36. 1708, 1914 j Phil. Mag.,
(6), 27. 188, 1914 ; L Langmuir, Journ. Amer. Chem. Soc, 37. 417, 1915 ; 38. 1145, 1916 ; G. N.
Lewis and M. Randall, ib., 36. 1969, 1914 ; A. E. Freeman, ib., 35. 927, 1913; T. Ismardi, Zeit.
Elektrochem., 21. 405, 1915 ; I. Langmuir and G. M. J. Mackay, ib., 20. 498, 1914; E. Briner,
Journ. Chim. Phys., 12. 109, 1914.
§ 9. The Difhision of Gases
Owing to the fact that air is fourteen times as heavy as an equal volume of
hydrogen under the same physical conditions, if a jar of hydrogen be placed mouth
upwards under a jar of air, mouth downwards, most of the hydi;ogen will flow
upwards into the upper cylinder, and air will flow downwards into the lower cylinder.
The action is analogous with what would occur if the lower cylinder contained a
light oil and the upper cylinder water. The oil and water would change places. The
two liquids can be left an indefinite time without mixing, for the two liquids — oil
and water — are immiscible. Several of the earlier chemists used to argue that gases
like those under consideration would behave like the liquids, but J. Priestley i noticed
that this is not the case ; and in 1805, J. Dalton's experiments demonstrated that
two gases — hydrogen and air — when brought into contact, do not arrange themselves
according to their specific gravities, but spontaneously difiuse, mutually and equally
through each other, and spread throughout the two vessels so that a homogeneous
mixture of air and hydrogen is obtained. This phenomenon is so marked, that it
was once considered that the molecules of a gas mutually repelled one another
owing to the presence of a " self-repulsive force." Gases are miscible in all propor-
tions ; had the two liquids been mutually soluble in one another in all proportions
— say alcohol and water instead of oil and water — they also would diffuse one into
the other so as to form a homogeneous solution of alcohol and water.
The process of diffusion in the case of liquids appears to be very much slower
than with gases. The molecules of gases seem to lead a more or less independent
existence. This is illustrated^ by the rapidity with which the molecules of, say,
ammonia can travel from one end of a room to the other and affect the sense of
smell. In liquids, however, the molecules are much less mobile. This can easily
be proved by dropping a small grain of aniline dye into a tumbler of clear still water.
The water will be uniformly coloured in a few weeks. The molecules of solid sub-
stances have practically lost their mobiHty but not all. Carbon laid in contact
HYDROGEN
339
with pure, hot, solid iron will diffuse into a mass of the metal ; gold or platinum in
contact with lead will diffuse into the lead.^ J. Violle found that carbon will diffuse
into the solid body of hot porcelain ; and metallic silver will diffuse into hot glass,
staining it yellow. According to C, E. van Orstrand and F. P. Dewey (1915), the
coefficient of diffusion, of gold into sohd lead at 197° is 0-0076 ; at 150°, 0-004:3 ; and
at 100°, 0-0002. The penetration of gold into lead at 197° can be detected at a depth
of 2-5 cm. after 54 days ; the penetration of the lead into gold at a depth of 0-2 mm.
— but the methods of detecting lead in gold are not so sensitive as for gold in lead.
The transfer of gases in bulk from one vessel to another is an effect of gravitation,
whereas diffusion is not an effect of gravitation. C. L. Berthollet (1809) ^ con-
firmed J. Dalton's observation that hydrogen is rather more penetrative and
diffusive than any other gas tried. In 1829, T. Graham experimented on the
interdiffusion or miscibility of gases with one another, and the subject was pursued
further by J. Loschmidt (1870) and A. von Obermayer (1880). J. Loschmidt
placed two tubes containing different gases one over the other, and, after opening
a channel between them, determined the amounts of the different gases which had
changed places after the lapse of a definite interval of time. The results were
expressed in terms of what is called the coefficient o£ diffusion, h, which
represents the quantity of gas travelling per second through a surface of one
sq. cm., and along a distance one centimetre in length. Then, at 0°, the co-
efficient of diffusion, Tcq, of hydrogen into air is ^0=^"^^ J ^^^^ steam, 0-69 ;
into oxygen, 0-67 ; and into
CoHe C2H4 SO2 NgO COg
0-46 0-48 0-48 0*43 0-63
CH4
0-65
irbon dioxide into
N2O C2H4 CO O2 Air H2
0-092 0101 0-140 0-141 0-142 0-553
CH4
0159
^0
It is doubtful if the assumption that the coefficient of diffusion h is altogether
independent of the relative proportions of the two gases, is quite valid.* It is
assumed that the gases exert no chemical or physical attractive forces on one
another. J. Loschmidt found that the quotient obtained by dividing the coefficient
of interdiffusion by the square root of the product of the molecular weights of the two
interdiffusing gases is approximately constant — 1*8. This is illustrated in Table IV.
Table IV.- — The Interdiffusion of Gases.
Diffusion
coefficient k.
k
f^M^M^
Carbon dioxide into air .....
00512
1-825
Carbon dioxide into hydrogen
0-2001
1-877
Carbon dioxide into oxygen .
0-0507
1-904
Carbon dioxide into carbon monoxide .
0-0506
1-776
Carbon dioxide into methane
0-0571
1-514
Carbon dioxide into nitrous oxide .
0-0354
1-557
Oxygen into hydrogen
0-2598
2-079
Sulphur dioxide into hydrogen
0-1728
1-955
Carbon monoxide into oxygen
0-0649
1-942
Carbon monoxide into hydrogen .
0-2312
1-730
J. Loschmidt also found that the coefficient of diffusion A; at a pressure p, and at
absolute temperature T, is related with the diffusion coefficient Tcq at 273° K.,
and standard pressure, such that
, , / T Y760 , , ,, , .,n760
where a denotes the coefficient of thermal expansion of the gas, and 6 the tem-
perature. J. Stefan and J. Loschmidt found that the index n is 2, but A. von
340
INOKGANIC AND THEOKETICAL CHEMISTRY
Obermayer found n=2 (nearly) applies only for binary mixtures of carbon dioxide
with nitrous oxide, air, and oxygen ; for other mixtures n=l'5 (nearly).
Long before J. Loschmidt, Thomas Graham ^ showed that the speed at
which the molecules of a gas can diffuse or travel through thin porous membranes
or septa — ^porous earthenware', plaster of Paris, etc. — into air or a vacuum
is related to the specific gravity of the gas. For example, hydrogen diffuses
nearly four times as fast as oxygen ; the relative densities of oxygen and
hydrogen are nearly as 1 : 16 ; and the relative rates of diffusion of the two gases
are nearly as -y/ie : 1 ; i.e. as 4:1. Hence follows Graham's law of diffusion :
the relative speeds of diffusion of gases are inversely proportional to the square
roots of their relative densities. Graham measured the speed of diffusion of gases
through thin porous plates and found the numbers indicated in the last column
of the subjoined table.
Table V. — Speeds
OF Diffusion of some
Gases and Graham's Law.
Gas.
Relative density
H = l.
Speed of diffusion
calculated from
1
Observed speed
of diffusion
(Hydrogen = 1).
V Relative density
Hydrogen, Hg
1
1
1
Methane, CH^
8
0-354
0-351
Ethylene, C2H4
14
0-267
0-266
Carbon monoxide, CO
14
0-267
0-278
Nitrogen, Ng
14
0-267
0-265
Oxygen, Oj .
16
0-250
0-248
Hydrogen sulphide, H2S
17
0-243
0-248
Nitrous oxide, NgO
22
0-213
0-214
Carbon dioxide, CO 2
22
0-213
0-212
Sulphur dioxide, SO 2
32
0-177
0-177
The third column represents the theoretical numbers calculated on the assumption
that the speeds of diffusion are inversely as the relative densities, Graham's
numbers have been recalculated to H=l instead of air=l, and whole number
approximations for the relative densities have been introduced. The observed
numbers for the speed of diffusion agree very closely with those obtained by
calculation. There is an interesting application of the law :
To find the relative density of a gas by comparing its speed of diffusion with
that of another gas of known density. — Let Di and Dg represent the relative
densities of two gases one of which, Di, is known, the other, D^, being unknown.
Suppose that the relative speeds of diffusion of the two gases Vi and V2 are
known. Then, it follows from Graham's law, Fi V'i)i= FgVA 5 or 7i2Z)i== 72^-^2-
Given any three of these numbers, the fourth can be calculated by arithmetic.
Examples. — (1) The speeds of diffusion of carbon dioxide and of ozone were found by
Soret (1868) to be as 0-29 (Fi) is to 0*271 (V^). The relative density of carbon dioxide is
22 (I>i) when H=l. What is the relative density of ozone (D^) ? From the preceding
relation, it follows that V^2=0-29x V22 ■^0-271=0-29 x4-69-^0-271 ; or J[)2=(5-02)2=25
nearly.
(2) A. Ladenberg (1898) found that ozonized air required 367-4 seconds to diffuse under
conditions where pure oxygen required 430 seconds : what is the specific gravity of the
ozonized air, assuming that the specific gravity of oxygen is unity ? Ansr. 1-3698.
The slow mutual diffusion of gases makes them dissolve each other irresistibly
so that there is no limit to the extensive expansion of the smallest bubble of gas
which may be formed or liberated in air. It is this ceaseless interchange of gases
which plays a part in preventing cities being speedily enveloped and stifled by their
own noxious exhalations from the natural functions of their teeming population,
and from the deadly vapours of their fires and furnaces. The ready diffusion of gas
through the walls of buildings plays a part in ventilation. Most building materials
I
HYDKOGEN
341
Diffusion Experiments.
are porous, and permit the passage of gases through them in both directions. The
diffusion does not take place so readily when the walls are saturated with moisture
—e.g. in new buildings, etc. The diffusion is illustrated by the following
demonstration :
One leg of a U-tube is enlarged, as shown in Fig. 16, and the widened end fitted with a
plug of plaster of Pans. The mercury of the other leg of the U-tube is connected with a battery
and electric bell, and a wire fused in the other leg of the tube, so that when the liquid rises
electric contact is made, and the bell will ring. If a beaker of hydrogen be placed over
the plaster of Paris plug, hydrogen will rapidly diffuse into the enlargement, and create
a pressure which depresses the mercury in the one leg, and raises it in the other so as to
ring the bell. A device based
on this principle has been sug-
gested as an alarm indicator
for the escape of coal gas in
rooms, or fire-damp in coal
mines. These gases, like hy-
drogen, diffuse through the
walls of porous pots faster than
the air can escape and produce
an internal pressure. Con-
versely, if the enlarged bulb be
bent downwards and a jar of
carbon dioxide be lifted up-
wards so that the porous sep-
tum is bathed in this gas, air
escapes from the interior faster
than the heavier carbon dioxide
can enter. This reduces the
pressure in the interior, and
causes a movement of the mercury in the opposite direction. Instead of using a plaster
of Paris septum, porous pots fitted on to glass tubes can be employed, and the conditions
can be so arranged that the pressure of the gas sprays a jet of liquid like a miniatm-e
fountain.
V In difEusing through porous walls, or through unglazed earthenware vessels,
the gases actually travel through the pores of the material. The diffusion
of gases through metals and indiarubber appears to follow a different procedure,
for the gas probably dissolves in these substances, diffuses while in solution through
the solid, and is liberated on the
opposite side. The diffusion of
gases through rubber is well illus-
trated by filling toy balloons of
thin rubber with hydrogen gas ;
the balloons soon collapse. If
the balloons be inflated with air,
and then immersed in hydrogen
gas, they swell out and burst.
For similar reasons, a soap bubble
blown with air and then made
to float on carbon dioxide gas, will gradually sink in the gas owing to the diffusion
of the heavy gas through the walls of the bubble.
The separation of a mixture of gases by diffusion.^ — If a slow current
of electrolytic gas, that is, the mixture of hydrogen and oxygen obtained
by the electrolysis of water, be allowed to pass through the stem of a church-
warden clay pipe, or porous earthenware tube, and the gas issuing from the pipe
be collected in a gas trough, it will no longer explode when brought in contact
with a flame. On the contrary, it will rekindle a glowing chip of wood, showing
that oxygen is present. In passing through the tube, hydrogen escapes by
diffusion through its porous walls much more rapidly than the heavier oxygen.
If the porous tube be surrounded by a glass tube, and fitted with a gas
delivery tube, as shown in Fig. 17, the gas which collects in the receiver A will
Fig.
17.— H. St. C. Deville's Experiment on
Atmolysis,
342
INORGANIC AND THEORETICAL CHEMISTRY
contain an excess of oxygen, and that in receiver B an excess of hydrogen. This
phenomenon — the separation of one gas from another by diffusion — has been called,
by T. Graham, atmolysis — ar/xo?, vapour ; \vw, I loosen. If a current of
steam be passed through a porous tube at a high temperature, and if the porous tube
be surrounded by another tube of glazed impervious porcelain, the water vapour will
be dissociated by the heat into hydrogen and oxygen ; since the hydrogen diffuses
much faster than the oxygen, hydrogen will pass from the inner tube into the
annular space between the two tubes. The hydrogen may be collected in a suitable
gas trough. Similarly, the residual oxygen can be collected from the gaseous
steam passing along the inner tube. Of course, the hydrogen is contaminated with
more or less oxygen, and the oxygen with more or less hydrogen. In 1906, A. Jouve
and G. Gautier ^ proposed to separate carbon monoxide and hydrogen industrially
by filtration through a porous partition.
Are the molecules of a gas all alike ? — Experiments on atmolysis
enable a conditional answer to be returned. If a gas like hydrogen or oxygen be
filtered through a septum of porous earthenware, no difference can be detected in
the properties of the gas on each side of the partition. If some of the molecules
of, say, hydrogen have a greater density than the others, it should be possible to
separate two kinds of hydrogen of different density. This cannot be done, and hence,
it is inferred that within the limits of experimental observation, the molecules of
hydrogen are all alike, for if there be a difference, it is too small to be detected.
Similar remarks apply to oxygen and many other gases. W. Ramsay and J. N.
Collie (1897) ^ showed that helium is probably a single substance because repeated
diffusion did not alter its density ; Lord Rayleigh and W. Ramsay (1895) showed
that atmospheric nitrogen is probably a physical mixture of nitrogen with a heavier
gas, because repeated diffusion altered its density ; and F. W. Aston (1913)
separated neon into two gases with densities 19'9 and 22*1, and hence inferred
that neon is probably a mixture of two gases of different densities.
The effusion of gases. — T. Graham (1832) » found that the law of diffusion
holds good for the passage of a gas through a very fine aperture in a metal plate.
T. Graham called the phenomenon the effusion of gases to distinguish it from diffu-
sion. The theory was indicated by D. Bernoulli lo in 1738. The speed of effusion
varies inversely as the square root of the density. If a gas of density Dj flows
out of the tube in the time ti, then, according to the law of effusion : Dj : Z)2=<i- : ^2^-
In words, the times required for the efflux of equal volumes of different gases,
under like conditions of pressure, etc., are directly proportional to the square
roots of the densities of the gases. Thus T. Graham found (air unity) :
Table VI.
Time of efflux.
®**- ""* density.' ""
Glass tube.
0-277
0-75«
0-987
0-984
1-053
1-199
1-218
Perforated brass
plate.
Hydrogen
Methane
Carbon monoxide
Ethylene
Nitrogen
Oxygen
Nitrous oxide
Carbon dioxide
0-263
0-745
0-984
0-985
0-986
1-051
1-237
0-276
0-753
0-987
0-986
1053
1-203
The method for determining the specific gravity of coal gas based on this principle
is considered one of the best. Thus, the time of effusion observed for gas was
2' 25-1" or 145-1 seconds, and for air, 3'40-8" or 2208 seconds. Hence, the specific
HYDROGEN
343
1
gravity of the gas is (145-l/220-8)2=0-432. R. Bunsen (1857) ^ utiUzed this fact
to determine the specific gravity of a gas when but a small quantity is available.
In N. H. Schilling's effusion apparatus (1879) the gas is introduced into a glass tube,
luted to a brass cover, vid the cocks 6, c, Fig. 18. This tube is placed in a cylinder
filled with water and the temperature is indicated by the thermometer T. The
perforated platinum plate is fixed at h. The time taken for the tube to sink
from the level P to the level 0 on the cylinder is noted.
The experiment is repeated with another gas of known b
density. A. Ladenberg (1898) ^2 used this apparatus to
determine the specific gravity of ozone. A. Debierne
(1910) used Bunsen's effusion method for determining the
density of the emanation from radium ; and F. Emich g tTfU^T
(1903) modified the apparatus so that it could be used at
high temperatures between 1400° and 2000°.
Examples. — (1) N. H. Schilling (1879) foimd that a given
volume of air effused in 285 seconds while the same volume of
coal gas effused in 209 seconds, hence the relative density of coal
gas with respect to air unity is (209)2/(285)2=0-538.
(2) A. Ladenberg (1898) found that a mixture containing
86' 16 per cent, of ozone required 267 '5 seconds under conditions
where pure oxygen required 430 seconds. Hence, determine the
specific gravity of ozone. From a previous example, it follows
that the specific gravity of the ozonized oxygen is 1*3698,
oxygen=l. The specific gravity of ozone is 1*456, if oxygen be
unity ; and 46'6, if oxygen O2 be 32. Note that this does not
establish the molecular weight by Avogadro's hypothesis, because,
in determining the proportion of ozone in the mixtiire, by estimating the amount of
iodine liberated by a given volume of the gas, it was assumed that the reaction pro-
ceeds as indicated in a preceding equation, which in turn assumes that the formula of
ozone is O3.
FiQ. 18. — Schilling's
Effusion Apparatus.
References.
1 J. Priestley, Experiments and Observations on Different Kinds of Air^ Birmingham, 3. 390,
1781 ; Trans. Amer. Phil. Soc.y 5. 15, 1802 ; J. Dalton, Mem. Manchester Lit. Phil. Soc, (2), 1.
244, 1805.
2 J. C. Roberts-Austen, Phil Trans., 187. 383, 1896 ; Proc. Boy. Soc, 67. 101, 1900 ;
F. Guthrie, Phil. Mag., (5), 16. 321, 1883 ; G. E. van Orstrand and F. P. Dewey, U.S. Ge.ol. Sur.
Prof. Paper, 65. 83, 1915 ; J. VioUe, Compt. Bend., 94. 28, 1882; A. Colson, ib., 93. 1074, 1881 ;
94. 26, 1882.
3 C. L. Berthollet, Mem. Arcueil, 2. 463, 1809; T. Graham, Quart. Journ. Science, 2. 74,
1829 ; J. Loschmidt, Sitzber. Akad. Wien, 61. 367, 1870 ; 62. 468, 1870 ; A. von Obermayer,
ib., 81. 1120, 1880 ; 85. 147, 748, 1882 ; 87. 188, 1883 ; 96. 546, 1887 ; J. Stefan, ib., 63. 63,
1871 ; 65. 323, 1872 ; K. Waitz, Wied. Ann., 17. 201, 251, 1882 ; V. Hausmaniger, Sitzber. Akad.
Wien, 86. 1074, 1882.
* 0. E. Meyer, The Kinetic Theory of Gases, London, 1899.
5 T. Graham, Phil Mag., (3), 2. 175, 269, 351, 1833 ; Phil. Trans., 136. 573, 1846 ; 139.
349, 1849.
6 T. Graham, Phil. Trans., 153. 385, 1863.
7 A. Jouve and G. Gautier, French Pat. No., 372045, 1906.
8 W. Ramsay and J. N. CoUie, Compt. Bend., 123. 214, 1896 ; Nature, 54. 546, 1896 ; Lord
Rayleigh and W. Ramsay, Phil. Trans., 186. 187, 1895 ; F. W. Aston, Eng., 96. 423, 1914 ; Nature,
104 334 393 1919.
» T. Grah'am, Phil. Mag., (3), 2. 175, 1833 ; Phil. Trans., 136. 573, 1846 ; 153. 385, 1863 ;
Trans. Boy. Soc. Edin., 12. 222, 1834 ; Proc. Boy. Soc, 12. 612, 1863 ; G. Baudrimont, Journ.
Pharm., (3), 29. 266, 1856 ; F. Exner, Pogg. Ann., 155. 321, 443, 1875 ; Sitzber. Akad. Wien,
73. 263, 1873.
^° D. Bernoulli, Hydrodynamica, Argentorati, 1738.
^^ R. Bunsen, Gasometrische Methoden, Braunschweig, 1857 ; A. de Negri, Ber., 3. 913, 1870 ;
M. Plettner, Dingier' s Journ., 229. 537, 1878 ; N. H. Schilling, ib., 155. 194, 1860 ; T. C. Mendenhall,
Amer. Chem., 6. 91, 1875 ; A. Wagner, Zeit. anal. Chem., 16. 76, 1877 ; J Krell, Journ. Gas-
beleucht., 42. 212, 1899 ; R. ThreKaU, Proc. Boy. Soc, 77. 542, 1906.
12 A. Ladenberg, Ber., 31. 2508, 1898 ; F. Emich, Monatsh., 24, 747, 1903 ; A. Debierne,
Compi. Bend., 150. 1740, 1910.
CHAPTER VIII
OXYOEN
§ 1. History of the Discovery oi Oxygen
A substance is discovered only when it has been shown to differ from all other known
substances.' — G. W. von Leibniz.
The history of the discovery of oxygen is intimately connected with Lavoisier's
discovery of the cause of combustion. C. W. Scheele's laboratory note-books,
preserved in the Koyal Academy of Science at Stockholm, are said to prove
that he discovered oxygen some time before 1773. C. W. Scheele exposed
alkaline liver of sulphur, moist iron filings, phosphorus, etc., to the action of air
confined in a jar over water ; and he also burnt phosphorus, hydrogen, and a candle
in air confined under similar circumstances. By measuring the volume of water
which entered the vessel at the end of the process, C. W. Scheele found what fraction
of the original volume had disappeared. He concluded :
Thus much I see from the experiments mentioned, that the air consists of two fluids
differing from each other, the one of .which does not manifest in the least the property of
attracting phlogiston, while the other, which composes between a third and a fourth part of
the whole mass of air, is peculiarly disposed to such attraction.
C. W. Scheele failed to find the lost air in the liver of sulphur, etc. He then
attempted to explain the facts by assuming that heat is a compound of fire-air
(oxygen) and phlogiston, and concluded that by a double affinity, the fire-air com-
bined with the phlogiston of the liver of sulphur, and escaped through the pores of
the glass by which it had been confined. Where weight disappears, analysis is
impossible, and C. W. Scheele left the composition of air to be explained by those
who believed that elastic fluids cannot penetrate glass. In some further experi-
ments, he sought an air which would behave towards liver of sulphur, phosphorus,
etc., like that part of common air which had the pecuHar property of attracting
phlogiston from inflammable bodies. He first obtained what he called Feuerluft —
fire-air — by heating a mixture of nitre and sulphuric acid, and found that a candle
burned in the gas briskly and brightly. He obtained fire-air in several other ways —
namely, by heating red oxide of mercury ; black oxide of manganese ; black oxide of
manganese and sulphuric acid ; nitre ; etc. — and noticed that it was absorbed by
those substances which caused a portion of common air to disappear. He said :
These experiments show, therefore, that this fire-air is just that air by means of which
fire bums in common air ; only it is there mixed with a kind of air which seems to possess
no attraction at all for the inflammable substance, and this it is which places some
hindrance in the way of the otherwise rapid and violent inflammation.
C. W. Scheele's mind was probably a little misled by the phlogiston hypothesis,
for he did not see that during combustion the burning body absorbs something from
the air. The vitiated air — verdorbene Luft — which remained did not support com-
bustion. C. W. Scheele also found the specific gravity of the vitiated air to be rather
less, and that of fire-air rather greater, than that of common air.
C. W. Scheele did not publish an account of his work until 1777 ;^ meanwhile,
Joseph Priestley independently prepared the same gas which he called dephlogisticated
344
OXYGEN 345
air. He encountered the gas while examining the effect of heat upon a great
variety of substances confined in a cylinder inverted over mercury, and he
decomposed them by focussing the sun's rays upon them by means of " a burning
lens of 12 inches diameter, and 20 inches focal distance." Joseph Priestley 2
announced his discovery of oxygen in these words :
On the 1st of Augiist, 1774, I endeavoured to extract air from mercurius calcinatus per
fte — that is, mercuric oxide, or red oxide of mercury — and I presently found that by means
of this lens, air was expelled from it verj'^ readily. Having got about three or four times as
much as the bulk of my materials, I admitted water to it and foimd that it was not imbibed
by it. But what surprised me more than I can well express, was that a candle burned in
this air with a remarkably brilHant flame.
and furthermore he said ^ that he ascertained the nature of this air very gradually
during the following March, 1775. He then tried the effect of adding one measure of
nitric oxide to two measures of the new air, and found that the new air behaved like
common air. He added :
I did not take notice of what I might have observed if I had not been so fully possessed
by the notion of there being no air better than common air, that the redness was really deeper,
and the diminution something greater than common air would have admitted. I now con-
cluded that all the constituent parts of air were equally and in their proper proportion imbibed
in the preparation of this substance, and also in the process of making red lead.
Until the first of March, J. Priestley said that he had no suspicion of the new air
being fit for respiration ; he then tried the mouse experiment of John Mayow,
and found the new air was longer respirable than common air. His ideas,
however were not so clear as those of J. Mayow, for instead of regarding air, with
J. Mayow, as a constituent part of nitric acid, J. Priestley thought air to be a com-
pound of nitric acid and earth ; and added that in December, 1777,
No doubt remained in my mind that atmospheric air, or the thing that we breathe,
consists of the nitric acid and earth, with so much phlogiston as is necessary to its elasticity,
and likewise so much more as is necessary to bring it from a state of perfect purity to the mean
condition in which we find it.
It is therefore fairly evident that neither of the so-called discoverers of oxygen —
Priestley and Scheele — had clear or accurate views of the nature of this gas, because
they interpreted their results in the light of hypotheses which we now know to be
wrong. This, however, does not detract from the merit of the discovery per se.
Between August, 1774, and March, 1775, when J. Priestley had no doubt that the
new air had all the properties of common air, A. L. Lavoisier, having heard an
account of the new gas from J. Priestley himself, ascertained its relation to various
bases and to the atmosphere ; and thus, said W. V. Harcourt,^ a property of this
gas which under Priestley's observations led to nothing, in the hands of A. L.
Lavoisier gave rise to one of the most important investigations in the annals of
chemistry.
It must be added that J. Priestley's observation ^ in 1772 that a candle burnt
with an enlarged flame in the gas obtained by heating nitre was perhaps as much a
discovery of oxygen as the experiment of August 1, 1774. J. Priestley then said :
All the kinds of factitious air on which I have yet made the experiment are highly
noxious, except that which is extracted from saltpetre or alum ; but in this, even a candle
burned just as in common air. In one quantity which I got from saltpetre, a candle not only
burned, but the flame was increased, and something was heard hissing similar to the decrepi-
tation of nitre in an open fire ; this experiment was made when the air was fresh made, and
while it contained some particles of nitre which it would probably have deposited afterwards.
Joseph Priestley seems to have thought that the gas was fixed air, and he did not
recognize his mistake until three years later.
For a century before this time, philosophers had suspected some land of relation-
ship between nitre and air, although there was no agreement as to the form of that
relationship. Some, like N. Henshaw (1661),^ maintained that the effects were due to
346 * INORGANIC AND THEORETICAL CHEMISTRY
the distribution of actual particles of nitre in the air, for " the air is everywhere full
of a volatile kind of nitre." R. Boyle also, in 1664, made a similar suggestion ;
T. Hobbes considered that nitre consisted of "many orbs of salt filled with air " ; while
R. Hooke (1664) believed that the portion of air which rendered air a solvent of com-
bustible bodies, is " like, if not the very same, with that which is fixed in saltpetre " ;
and J. Mayow followed this up by demonstrating that the gas in the atmosphere
which, in combining with bodies, produces the phenomenon of combustion, is iden-
tical with one of the components of nitre because (i) that salt is produced in the
atmosphere ; and (ii) the effects produced by both in enabling substances to burn
are the same — for instance, he showed that the acid component of nitre, of nitric
acid, and of the product obtained by heating antimony in the focus of a burning
glass, all furnish the same substance.
There has been some discussion as to whether the alchemists were acquainted
with oxygen, for there are indications in old books that this gas was known in the
early centuries. Zosimus, in the fourth century, said :
Take the soul of copper which is borne upon the water of mercury, and disengage an
aeriform body.
According to F. Hoefer,^ the water of mercury must here have referred to liquid
mercury, and the soul of copper to red mercuric oxide which floats on mercury which
has been heated in air. If so, the aeriform body given off when the soul of copper
is heated must have been oxygen. This is an ingenious and probable interpretation
of the passage, although it is easy to read into old writings a meaning which the
original wets not intended to express. Again, in 1742, J. H. Cohausen ^ referred to
the white dove mentioned in the writings of Philalethes. It was said that atmospheric
air is not respirable if the white dove be removed. In 1660, Robert Boyle ^ related,
in a gossiping style, that Cornelius Drebell invented a kind of submarine, and that
he had a chemical liquid which he accounted the chief secret of his navigation ;
for when the air was fouled by respiration, he could, by unstopping a vessel full of
the liquid, speedily restore such vital parts as would make the air fit for use again.
Although clothed in mystical language, as was customary with the writers on
alchemy of this period, it does seem as if the preparation of oxygen and its property
of supporting life may have been suspected by some of the alchemists ; but, their
statements are so often mixed with what we know to be false, and their means of
handling gases were so crude, that it requires some licence to be able to say that they
discovered this gas. It has also been claimed that oxygen was discovered by
E. Swedenborg,io half a century before Priestley, but the claims are based upon
obscure and unintelligible passages which carry their own refutation.
However, P. Eck de Sultzbach ^i in 1489, did know that red oxide of mercury
gave off a spirit when heated ; about 1557, J. Cardan, 12 and about 1674, Ole Borch
— the familiar term for Olaus Borrichius — obtained a fiatus by the action of heat
on saltpetre. This must have been either oxygen or nitrous oxide gas, because
J. Cardan said that the gas nourished flame, and rekindled a glowing splint. A
few years later, R. Boyle ^3 said :
On September 4, 1678, I exposed one ounce of minium in an open glass to the sunbeams
concentrated by a burning glass, and found that it had lost three-fourths of a grain of its
weight, though much of the minium had not been touched by the solar rays. . . . On
May 30, the same experiment was repeated in a light glass phial sealed hermetically, but
such plenty of air was produced that the glass phial broke into a hundred pieces.
Stephen Hales prepared the gas in a similar manner about 1727, and collected it
over water ; and in 1774, P. Bayen obtained the same gas by heating precipitated
mercuric oxide in a retort. In the light of subsequent events, it appears as if these
observers had really isolated oxygen gas, but one and all failed to establish its
identity as a distinct individual. In some cases, the observations do not seem to
have been clearly understood, probably because of the dominating influence of the
idea that air was the only ponderable gaseous element. Had any one of these
OXYGEN 347
observers not only isolated the spirit or flatus, but also shown that it had specific
properties, sui generis^ which distinguished it from other known substances, he
would have been credited with the discovery of oxygen.
The Chinese appear to have known something about oxygen in the eighth century,
long before Joseph Priestley's and C. W. Scheele's experiments. Whatever be the
facts of the case, the work of the Chinese played no part in the European discovery
of this gas. In a paper On the chemical knowledge of the Chinese in the eighth
century y^'^ J. Klaproth, quoting from the Chinese, said :
There are many substances which rob the atmosphere of part of its yin, the chief of these
are the metals, sulphur and carbon. . . . The yin of the air is always pure, but by the aid
of fire, yin can be extracted from nitre, or from a black mineral (black oxide of manganese)
found in the marshes. It also enters into the composition of water, where the union is so
close that decomposition is extremely difficult. . . . Gold never amalgamates with the yin
of the air.
It will therefore be obvious that many erring steps have stumbled on the threshold
of the discovery of oxygen. Although J. Priestley and C. W. Scheele are usually
credited with the discovery of this gas, it is extremely difficult to decide whose name
deserves the highest place in the discovery, for R. Hooke and J. Mayow inferred its
existence in nitre and in air ; P. Eck de Sultzbach and R. Boyle disengaged the gas
from red lead or minium ; S. Hales collected the gas in a jar over water ; and
J. Cardan, C. W. Scheele, and J. Priestley observed that it supported combustion.
Before answering the question : Who discovered oxygen ? it is necessary to answer :
What is meant by the term discovery ?
References.
1 C. W. Scheele, Chemische Abhandlung von der Luft und dem Feuer, Upsala, 1777 ; A. E.
Nordenskiold, Nachgelassene Briefe und Aujzeichnungen, Stockholm, 1892 ; Alembic Club Beprints,
8, 1894.
2 J. Priestley, Experiments atid Observations on Different Kinds of Airy London, 2. 29-103,
1775 ; Alembic Club Reprints, 7, 1894.
^ J. Priestley, Experiments and Observations on Different Kinds of Air, London, 2. 113, 1790.
* W. V. Ilarcourt, Phil. Mag., (3), 28, 478, 1846 ; M. Speter, Lavoisier und seine Vorldufer,
Stuttgart, 1910 ; S. M. Jorgensen, Die Entdeckung des Sauerstoffs, Stuttgart, 1909.
5 J. Priestley, Phil. Trans., 62. 245, 1772.
^ N. Henshaw, Aero-chalinos, or a Register of Air, London, 1664.
' F. Hoefer, Histoire de la chimie, Paris, 2. 271, 1843.
" J. H. Cohausen, Hermippus redivivas, Frankfurt, 1742 ; E. Philalethes, Ripley Revtvd,
London, 1678 ; H. C. Bolton, Amer. Chem., 4. 170, 1873.
^ R. Boyle, New Experiments Physico-mechanical touching the Spring of Air, London, 1660.
^^ E. Swedenborg, Prodromus principiorum rerum naturaliiim, Amsterdam, 1721.
11 P. Eck de Sultzbach, Theatrum chemicum, Argentorati, 4. 1007, 1622.
12 J. Cardan, De rerum varietate, Basil, 668, 1557 ; 0. Borrichius, Hermeiis, ^gypiiorum et
cJiemicorum sapientia, Hafnise, 1674.
13 R. Boyle, Philosophical Works, London, 2. 633, 1725.
1* J. Klaproth, Mem. Imp. Acad. St. Petersburg, 2. 476, 1810 ; W. Duckworth, Chem. News,
53. 250, 1886.
§ 2. The Action of Heat on Mercuric Oxide
In the thirteenth century, the Latin Geber knew that if mercury be heated
in air, it forms the red oxide ; and P. Eck de Sultzbach, as previously stated,
appears to have known that if the red oxide be heated, it gives oS a spirit or gas.
Here then is a paradoxical result : Heating mercury in oxygen furnishes mercuric
oxide ; heating mercuric oxide furnishes oxygen and mercury. Several investiga-
tions 1 have been made to determine the conditions of temperature and pressure
which determine the course of this reaction — evidently one of the balanced type,
symbolized 2Hg0^2Hg-f-02. Unlike the reaction between steam and iron in
348
INORGANIC AND THEORETICAL CHEMISTRY
closed vessels, Fig. 4, Cap. Ill, the reaction is here complicated by the vaporization
of the mercury. If pQ denotes the partial pressure of the oxygen at any assigned
temperature, and p„^ that of the mercury, Guldberg and Waage's law gives the rela-
tion poPfn^=coiista.nt. The product poPm^ is least when the partial pressure of the
mercury vapour is twice that of the oxygen, i.e. when the two elements are in the same
proportion as they exist in the solid phase. When the system is in equilibrium,
mercur}^ and oxygen are uniting to form mercuric oxide, and mercuric oxide is decom-
posing to form mercury vapour and oxygen. When the speeds of the two reactions
are perfectly balanced, the system is in equihbrium, and the gases exert a definite
pressure called the equilibrium or dissociation pressure on the walls of the vessel
as measured by the manometer. This pressure thus measures the concentration of
the gases, and one-third of the pressure is due to oxygen, and two-thirds to the
mercury. When the temperature of mercuric oxide in vacuo is raised the speed of
decomposition of the oxide is augmented more than the rate of re-combination of the
two gases, and consequently, the equilibrium pressure or the concentration of the
gases increases as the temperature is raised, until finally, all the mercuric oxide is
decomposed, and the mixed gases then obey the ordinary gas laws. J. B. Taylor
and G. A. Hulett represent the relation between the absolute temperature, T, and
the dissociation pressure, P, by the expression log P= — 5273'5T~^-f-l'75 log T
—0-001033r-f 5-9461. The following values of the total pressure P ^,t different
temperatures are selected from G. B. Taylor and G. A. Hulett's determinations :
Temperature
360°
390°
420°
450°
480° C.
Equilibrium pressure .
90
180
387
810
1581 mm
1500
6
.o
J
er
'or
I
cu
ma
of
ric
hi
u
3/7
l-
\
1
Q.
-y
— -
<n
■o:..:x.
/.
;;^^;-*v:
/
o^'
y
/
^
/>> j
rrr
•.-.-0-.I
The curve showing the equilibrium pressure at different temperatures is indicated
in Fig. 1. When the relation between the total pressure of the oxygen plus
mercury and temperature can be represented by a point on this curve, the system
is in equilibrium ; if by a point to the right of the curve, mercuric oxide will be
decomposing and this will continue until the pressure
attains its equilibrium value ; and if by a point to the
left of the curve, the gases will be recombining and mer-
curic oxide will be formed until the pressure is reduced to
its equilibrium value. Assuming the partial pressure of
oxygen in air is one-fifth of 760 mm., i.e. 152 mm., and
that mercuric oxide is heated in air, the oxide will com-
mence to decompose when heated above 426°, and the
temperature must be kept below this limit if it is desired
to form mercuric oxide by heating mercury in air.
Were it not for the cost, the preparation of oxygen
by heating mercuric oxide would be very convenient.
Ten grams of mercuric oxide give not quite half a litre of
oxygen. The mercury is a by-product of the reaction.
Mercuric oxide decomposes more rapidly if it be pre-
viously mixed with powdered platinum or with certain
metal oxides — e.g. ferric oxide, manganese dioxide, cad-
mium oxide, etc. These substances are not chemically
changed during the decomposition of the mercuric oxide, and they can be used
indefinitely over and over again. Hence, the oxides, which accelerate the decompo-
sition of mercuric oxide are called catalytic agents. Stannic oxide or alumina do
not exert any appreciable catalytic effect.
Many other oxides are available in place of mercuric oxide — e.g. gold and silver
oxides decompose at temperatures even lower than mercuric oxide, while manganese
peroxide (pyrolusite) decomposes at a higher temperature — over 400°. The prepara-
tion of oxygen by heating manganese dioxide was formerly one of the cheapest
methods of preparing the gas on a commercial scale. The process was also used by
C. W. Scheele in 1777 ; and by J. Priestley in 1779. Unlike silver, mercury, and gold
300°
500
400"
Temperature.
Fig. 1. — Equilibrium Pres-
sure of Mercuric Oxide
(Partial pressure of oxygen
= \ Total pressure).
OXYGEN 349
oxides, the manganese oidde does not break down into the corresponding metal and
oxygen, but rather into a complex oxide similar in composition to the mineral
hausmannite — Mn304. The reaction is symbolized : 3Mn02=Mn304+02.
Fairly pure manganese dioxide will furnish 88 to 89 litres of oxygen per kilogram ;
but pyrolusite, the native form of manganese dioxide, may give as low a yield as 34
litres of oxygen per kilogram, and the gas is contaminated with carbon dioxide and
nitrogen or nitrogen oxides derived from impurities in the mineral. 2 Gold, silver,
and mercuric oxides are not often used as sources of oxygen on account of the expense.
It will be observed that the alternate oxidation of mercury by roasting in air and
the decomposition of the resulting oxide suggests a means of extracting oxygen from
the air. Similarly, palladium sponge is directly oxidized when heated to redness
in air, and the resulting oxide, Pd20, reforms the metal with the evolution of oxygen
at a little higher temperature. When rhodium sponge is similarly treated, the oxide,
RhO, is formed ; and with iridium sponge, the oxide Ir30. These two oxides give
oxygen and the metal ^ at about 1200°.
References.
1 J. Myers, Ber., 6. 11, 1873 ; H. Debray, Compt. Rend., 77. 123, 1873 ; H. Pelabon, ib.,
128. 825. 1899 ; Mem. Soc. Bordeaux, (5), 5. 68, 1901 ; T. Camelley and J. Walker, Journ. Chem.
Soc, 53. 80, 1888 ; W. H. Echols, Chem. News, 44. 189, 1881 ; G. B. Taylor and G. A. Hulett,
Journ. Pkys. Chem., 17. 565, 1913.
2 M. Carlevaris, Bull. Soc. Chim., (2), 4. 255, 1865 ; H. St. C. Deville and H. Debray, Campt.
Rend., 50. 868, 1860 ; J. B. J. D. Boussingault, ib., 50. 890, 1860 ; W. H. Echols, Chem. News, 44.
189, 1881.
8 T. Wilm, Bull. Soc. Chim., (2), 38. 611, 1882; H. St. C. DeviUe and H. Debray, Compt.
Rend., 87. 441, 1878.
§ 3. The Action o£ Heat on Potassium Chlorate
Potassium chlorate is a white crystalUne solid which melts to a clear liquid when
heated to about 340°. According to A. Killiet and J. M. Crafts,i potassium chlorate
begins to give off oxygen below its melting point and the decomposition goes on
several weeks before it becomes imperceptibly small ; when the limit is reached,
a rise of temperature starts a fresh decomposition, tending towards a new limit.
At about 10° above its melting point, the melted chlorate appears to boil, because
bubbles of oxygen gas are copiously evolved. The potassium chlorate is decomposing.
When the bubbling ceases, the molten mass becomes very viscid or even solidifies.
The potassium chlorate has decomposed into potassium perchlorate, potassium
chloride, and oxygen. If the temperature be raised still further — over 600° — the
mass again melts to a clear liquid and the potassium perchlorate decomposes, giving
off more oxygen. The final products of decomposition are potassium chloride and
oxygen, and the reaction is accordingly symbolized : 2KC103=2KCl+302.
Hence, potassium chlorate can be used in place of mercuric oxide for the preparation
of oxygen gas. Ten grams of potassium chlorate will give nearly 2| Htres of oxygen.
Purified and fused potassium chlorate gives oxygen free from chlorine and ozone,
O3. The gas is of a high degree of purity, and this process was used by E. W. Morley
in his work on the atomic weight of oxygen.
If potassium chlorate be suddenly heated to a temperature above that at which
decomposition occurs, the salt may detonate in an open vessel under ordinary pressure.
Some disastrous explosions have been produced by potassium chlorate. M. Berthe-
lot's experiment (1899) 2 illustrates the explosive nature of this salt.
One end of a glass rod is drawn out into a thread, and the narrow end is dipped several
times in molten potassium chlorate so that each layer of salt solidifies before the rod is dipped
again. When a bead has been formed at the end of the rod, dip the rod into a test-tube
heated red hot at one end so that the salt is about a centimetre from the bottom of the tube.
350 INORGANIC AND THEORETICAL CHEMISTRY
Take care not to touch the sides of the tube. As the chlorate melts, it slowly drops to the
bottom of the test-tube ; each drop of chlorate as it falls explodes with a sharp
detonation.
C. L. Berthollet 3 first made oxygen by the chlorate process in 1785 ; and in
1832, J. W. Dobereiner noticed that the potassium chlorate decomposes completely
at a much lower temperature if it be mixed with manganese dioxide ; the reaction
also progresses more smoothly and is more under control. Other agents can be
employed — e.g. the oxides of copper, iron(ic), lead, cobalt, vanadium, uranium,
and tungsten. With ferric oxide, the evolution of oxygen begins between 110° and
120° ; with manganese dioxide, between 200° and 205° ; and with platinum black,
between 260° and 270°. The stimulating action of alumina and chromic oxide is less
marked ; and baryta, lime, magnesia, and zinc oxide have no influence. Powdered
glass, sand, and china clay promote the decomposition of the chlorate to a small
extent. Although oxygen can be easily obtained by heating manganese dioxide,
Mn02, to about 600° in a fireclay, porcelain, or iron retort (3Mn02->Mn304H-02),
and potassium chlorate, when heated alone, does not give off oxygen below about
240°, yet a mixture of the two gives off oxygen at about 205° — a temperature below
the melting point of the chlorate. After the action, manganese dioxide still remains,
but the potassium chlorate has decomposed into potassium chloride and oxygen.
Manganese dioxide can be recovered from the residue by lixiviating the mass with
water. The water dissolves the potassium chloride, and leaves the manganese
dioxide as a residue. The temperature of the reaction with ferric oxide and cupric
oxide is nearly the same as with manganese dioxide ; with platinum black, it is
270°, and 285° with lead dioxide.'* Cobaltic and nickelic oxides behave like
manganese dioxide and accelerate the decomposition of potassium chlorate. The
nitrogen is probably derived from nitrogeneous impurities in the dioxide, and is not
a result of the oxidization of atmospheric nitrogen because a similar result is obtained
if the dioxide be heated in a current of inert gas — say carbon dioxide.
C. F. Schonbein ^ reported that the gas obtained from lead dioxide, mercuric
oxide, silver oxide, and chlorates contains traces of ozone, but A. R. Leeds showed
that the ozone reactions obtained by C. F. Schonbein were produced by traces of
chlorine. According to W. Spring and E. Prost, the amount of chlorine formed is
insignificant if the chlorate be decomposed in a platinum vessel, but quite marked
quantities are obtained when porcelain vessels are used. In the presence of carbon
dioxide or phosphoric oxide, it is possible that chloric anhydride, CI2O5, is disengaged
from the chlorate, and that this unstable gas is instantly decomposed into chlorine
and oxygen. According to 0. Brunck, ozone, not chlorine, is formed in the reaction
in amounts which increase with increasing proportions of manganese dioxide mixed
with the chlorate. Thus, with equal parts of chlorate and the dioxide the oxygen
contained 0*3 per cent, of ozone, whereas with 25 times as much of the dioxide, the
oxygen contained 1*55 per cent, of ozone. If alkaUes be present no ozone is formed,
and with sodiima carbonate, sodium peroxide is found in the residue. Nickelic and
cobaltic oxides, like manganese dioxide, give oxygen contaminated with ozone ;
while mercuric oxide, lead dioxide, and silver oxide, give ozonized oxygen when
heated with potassium chlorate. 0. Brunck also claims that if these mixtures are
heated in a stream of carbon dioxide no ozone is formed. It is quite certain that the
gas from the mixture of potassium chlorate and manganese dioxide has usually the
smell of ozone or chlorine. Traces of these two substances behave so much alike in
the usual test, that one can be easily mistaken for the other. Both chlorine and
ozone, for example, give a blue coloration with a solution of starch and potassium
iodide. E. H. Cook ^ has reported 0*03 per cent, of chlorine in the oxygen derived
from the chlorate. H. McLeod has shown that when potassium chlorate and
manganese dioxide are heated, some chlorine is produced, and claims that there is
no evidence of the simultaneous formation of ozone. For instance, the residue
left after the reaction is over gives an alkaline solution when extracted with water.
0. Brunck obtained no alkaline residue. The gas also gives a precipitate of silver
OXYGEN 351
chloride when it is passed through a solution of silver nitrate, suitable precautions
being taken to prevent fine particles of potassium chloride being carried along with
the stream of gas. 0. Brunck explains the discrepancy between his own results
and H. McLeod's by showing that the latter's apparatus contained traces of organic
matter which would decompose the ozone, and that H. McLeod used unnotig high
temperatures. Above 400°, 0. Brunck also obtained chlorine, and he suggests
that the manganese dioxide forms with the chlorate, permanganic anhydride,
Mn207, and that this gives the ozone form of oxygen, O3, in accord with the scheme :
MngOy -> 2Mn02 + O3 ; or graphically :
I
02=Mn:-0. ^i _ 02=Mn , O^^
Og^Mnj-O^ i Og^Mn O
and if chromic oxide is used a chromic anhydride, Cr206, is similarly formed and
decomposed. Most of the ozone is decomposed at the temperature of the reaction,
203=302.
Potassium chlorate, mixed with manganese dioxide, is commonly used in the
laboratory for the preparation of oxygen ; and with the idea of lessening the violence
of the action still more, the addition of 10 to 30 per cent, of common salt has been
recommended ; 7 but this is usually considered unnecessary. The chlorine can be
removed by passing the gas through magnesia, whiting, or soda ash distributed over
some inert material like glass wool or asbestos. If a highly pure gas is needed, fused
potassium chlorate can be used alone. The cost of oxygen by the chlorate process
is nearly double that by the pyrolusite process. By the former process, 1000 cub. ft.
of oxygen are said to cost £8 to £10, and by the latter £4 to £5, against 35. 6d. by
the liquid air process.
References.
1 A. RilUet and J. M. Crafts, B.A., Rep., 493, 1882.
2 M. Limousin, Journ. Pluirm. Chim., (5), 2. 178, 1880 ; M. P. E. Berthelot, Compt. Bend.,
129. 926, 1899 ; A. Dupre, Journ. Soc. Chem. Ind., 21. 217, 1902 ; D. Dollner, Chem. Ind., 22.
443, 1899 ; G. Lunge, Zeit. angew. Chem., 12. 537, 1899 ; C. A. Lobry de Bruyn, ih., 12. 933,
1899 ; R. Gartenmeister, Chem. Ztg., 31. 174, 1907 ; M. Couleru, ih., 31. 217, 1907 ; H. Landolt,
ih.y 31. 285, 1907 ; L. G. Marquart, ih., 31. 286, 1907.
3 C. L. Berthollet, Mem. Acad., 276, 1785 ; J. W. Dobereiner, Liehig's Ann., 1, 236, 1832 ;
E. Mitscherlich, Sitzher. Akad. Berlin, 62, 1841 ; G. J. Fowler and J. Grant, Journ. Chem. Soc,
57. 273, 1890.
* E„ Wiederhold, Pogg. Ann., 116. 171, 1862 ; 118. 186, 1863 ; German Pat. D.R.P., 299505,
1915.
5 C. F. Schonbein, Journ. prakt. Chem., (1), 65. 96, 1855 ; A. R. Leeds, Chem. News, 42.
304, 1880 ; T. C. Kingzett, ib., 25. 242, 1872 ; H. H. Croft, ib., 25. 87, 1872 ; C. F. Rammelsberg,
Pogg. Ann., 134. 534, 1868 ; G. Bellucci, Ber., 8. 905, 1875.
« E. H. Cook, Journ. Chem. Soc, 65. 802, 1898 ; J. C. G. de Marignac, Liebig's Ann., 44. 13,
1842 ; J. S. Stas, Chem. News, 73. 15, 1896 ; 0. L. Erdmann and R. F. Marchand, Journ.
prakt. Chem., (1), 31. 274, 1844 ; A. Wachter, ib., (1), 30. 321, 1843 ; H. Schulze,i&., (2), 21. 407,
1880 ; A. Vogel, Repert. Pharm., (3), 3. 145, 1849 ; J. C. Poggendorf, Pogg. Ann., 77. 17, 1849 ;
M. E. Chevreul, Compt. Rend., 29. 296, 1849 ; F. Bellamy, Monit. Scient., (4), 1. 1145, 1887 ;
W. Spring and E. Prost, Bull. Soc Chim., (3), 1. 340, 1889 ; F. Sestini, L'Orosi, 18. 5, 1895 ;
H. McLeod, Jo«m. Chem. Soc, 55. 184, 1889; 65. 202, 1894; 69. 1015, 1896; 0. Brunck, Ber.,
28. 1790, 1893 ; Zeit. anorg. Chem., 10. 222, 1895.
'' L. von Babo, Liebig's Ann. Suppl, 2, 265, 1862 ; *H. Landolt, Chem. Ztg., 3. 276, 1888.
§ 4. The Occurrence and Preparation of Oxygen
The occurrence of oxygen. — Oxygen is widely distributed on the earth in very
large quantities. It is an essential constituent of air and water. About one-fourth
— 23-2 per cent. — of the atmosphere by weight, and about one-fifth by volume consists
352 INORGANIC AND THEORETICAL CHEMISTRY
of free oxygen. J. H. Jeans' estimate i of the number of molecules of oxygen per c.c.
at a height h kilometres is
/<=0 /t = 20 A=80 ;t = 160 /i = &00
21X10" 7XlOi« 25X10" 3 X 10« 0
Water contains about 88"8 per cent, of combined oxygen. Oxygen also forms a
material part of rocks since a great many minerals contain a considerable proportion
of oxygen. F. W. Clarke estimates that 45 to 53 per cent. — nearly one-half — the
total weight. of rocks, and eight-ninths of the water which make up the half-mile
crust of the earth, is combined oxygen. Oxygen therefore is by far the most abun-
dant element, being nearly equal in amount to all the others put together. Natural
waters hold a small amount of oxygen in solution.
Oxygen is an essential constituent of animal and vegetable tissue and fluids. It
is absorbed from the atmosphere by animals and plants during respiration ; and
given off by plants when they assimilate carbon dioxide from the air in sunlight.
This fact was estabhshed by J. Priestley on AugustlTth, 1776, by showing that plants
could live in fixed air in which animals perish, and that plants can restore to fixed
air the properties of common air when in sunlight, but not in darkness. The expla-
nation of J. Priestley's observation was possible only after the function oi oxygen
in respiration was recognized. R. Perceval, and J. Senebier showed that the amount
of oxygen given off depends on the proportion of carbon dioxide in the atmosphere,
and J. Ingenhousz proved that sunlight is necessary for the reaction. The oxygen
contains no ozone. Some plants — e.g. hactarium photometricuin, can produce
oxygen without the agency of light ; and chlorophyllous animals give off oxygen
in sunUght.2
H. Draper ^ showed, in 1877, that oxygen is present in the sun ; J. Trowbridge
(1896), however, beheved that the spectral lines thought to be oxygen were due to
iron ; but later, from the analogy of the solar spectrum with the spark spectrum ot
water vapour, J. Trowbridge (1902) assumed that the sun contains dissociated water
vapour and consequently also free oxygen.
The preparation of oxygen. — There are many methods available for preparing
oxygen — the particular process to be employed must be determined by cost and
convenience.* If but a few litres of gas, not specially purified, are required, cost is
not very serious, and convenience is perhaps the most important factor ; if pure
oxygen be required, a complicated apparatus may be needed, and neither cost nor ,
labour must be spared. An elaborate apparatus may be needed to remove traces of
impurities — ^say, traces of ozone and chlorine from the oxygen. Pure potassium
chlorate alone will give a gas of a high degree of purity. If large quantities of gas are
needed, say for industrial purposes, the cost factor is of prime importance. Gener-
ally speaking, the success of industrial operations depends upon the ability of the
chemist to manufacture his products cheaply. In former times, oxygen was made
commercially by heating pyrolusite, and also by the chlorate method ; but the cost
was too great for these processes to compete successfully with the cheaper methods
by barium peroxide, or the fractional distillation of liquid air. The different
methods of preparing oxygen can be conveniently classed :
1. Processes dependent on the decomposition of oxides or oxy-compounds by heat. —
The methods of preparing oxygen by heating mercuric oxide, and by heating potassium
chlorate are typical. Several dioxides also yield oxygen when heated^ — e.g. manga-
nese dioxide, lead dioxide, barium dioxide, etc. The nitrates of potassium or sodium
give oxygen contaminated with nitrogen or nitrogen oxide.^ Oxygen was probably
obtained by this method by J. Priestley in 1771 — some years before the date gene-
rally accepted for his discovery of this gas. The first British patent for the manu-
facture of oxygen was by S. White ^ in 1849, and this was effected by heating nitre.
When the vapour of sulphuric acid is passed over fragments of brick or earthen-
ware heated to bright redness, oxygen, sulphur dioxide, SO2, and water are formed :
2H2S04=2H20+2S02-4-02. By washing the products in water or in a solution
OXYGEN 353
of sodium hydroxide, the sulphur dioxide is arrested. In this way, H. St. C. Deville
and H. Debray ^ obtained 140 litres of oxygen from a kilogram of concentrated
sulphuric acid — ^between 6 and 7 per cent, of the acid escaped decomposition.
W. S. Squire (1875) patented this process of making a mixture of oxygen and sulphur
dioxide for the manufacture of sulphur trioxide — the water was removed by a desic-
cating agent. H. St. C. Deville and H. Debray also obtained 6*8 litres of oxygen
by calcining zinc sulphate au rouge hlaiw. The sulphate decomposes in an analogous
manner to the acid.
When potassium or sodium permanganate is heated, it also furnishes fairly pure
oxygen.8 It is well to cover the salt with a heavy layer of glass wool to retain any
dust from the decomposition of the permanganate. The product of this reaction,
after scrubbing by passage through a layer of solid potassium hydroxide, and drying
by passage over phosphorus pentoxide, has been used in some refined determinations
of the relative density of the gas. The decomposition proceeds at a lower tempera-
ture in the presence of steam forming an alkali manganite (G. Kousseau), or a mixture
of manganic oxide and alkali hydroxide (C. M. T. du Motay and C. R. M. de Mare-
chal), and a regular stream of oxygen.^ The hypochlorites or hypobromites furnish
oxygen when they are heated. Cupric metaborate, CUB2O4, gives ofE oxygen i^ at
about 1000° forming the sesquiborate : 12CuB2O4=6Cu2O.2B2O3+3O2+10B2O3.
2. Processes dependent on the decomposition of oxides and oxy -compounds hy
chemical means. — There is probably no real distinction between many of the thermal
and chemical processes. Heat may be required to start the reaction in either class ;
and the thermal processes all involve chemical reactions. Many dioxides yield
oxygen when treated with water or dilute acid, in some cases at atmospheric
temperatures. Sodium peroxide, for example, is slowly decomposed by water into
sodium hydroxide and oxygen : 2Na202-f-2H20=4NaOH+02. If a catalytic
agent, say manganese dioxide, be present, the reaction is not inconveniently slow for
small quantities of gas. The trade name for a mixture of fused sodiimi peroxide
with a small quantity of manganese dioxide is oxone ; oxylith 11 is a compressed mix-
ture of sodium peroxide with about 62 per cent, of dry chloride of lime. It slowly
decomposes in contact with water, giving ofi oxygen : CaOCl24-H20+Na202
=Ca(OH)2+2NaCl+02. Numerous other mixtures have been patented for this
purpose ; as well as vessels — generators or auto-generators — for producing the gas.
When sodium peroxide is gently warmed with a salt containing water of crystalliza-
tion— e.g. Glauber's salt, or sodium carbonate — a steady stream of oxygenis given off.12
Many other methods are available for the preparation of oxygen. Heating sulphuric
acid with manganese dioxide ; 13 with chromic acid or potassium dichromate ; 1*
with potassium permanganate ; etc., and this more particularly when a powerful
oxidizing agent, rather than gaseous oxygen, is required. The reaction with manga-
nese dioxide is symbolized : 2Mn02+2H2S04=2H20-h2MnS04+02, and with
potassium dichromate : 2K2Cr207+8H2S04=2Cr2(S04)3-[-2K2S04+8H20+302.
Other salts rich in oxygen may be used. Hydrogen peroxide is sometimes convenient
for preparing small quantities of oxygen, although the cost is rather high. For
instance, a 10 per cent, solution of hydrogen peroxide mixed with a substance which
provokes its catalytic decomposition — e.g. chloride of lime ; 15 manganese dioxide ;
lead dioxide ; i^ potassium ferricyanide in alkaline solution 17 (2K3FeCy6-|-2KOH
+H202=2H20+2K4FeCy64-02) ; or potassium dichromate.i^ \ concentrated
solution of potassium permanganate gives ofi oxygen at ordinary temperatures
when acidified with sulphuric acid.
Heat a mixture of, say, 20 grams of potassium permanganate with 80-100 c.c. of dilute
sulphuric acid (one volume of the concentrated acid, with four volumes of water) in a flask
with a delivery tube and safety funnel. Oxygen begins to come off when the temperature
is about 50°, and continues in a steady stream. Ten grams of the permanganate with
between 40-50 c.c. of the dilute sulphuric acid give just over a litre of gas.
A. Baumann (1890) i^ charges a Kipp's apparatus with lumps of pyrolusite and a
mixture of commercial hydrogen peroxide with 15 per cent, by volume of sulphuric
VOL. I. 2 a
354 INORGANIC AND THEORETICAL CHEMISTRY
acid. A steady and continuous stream of oxygen is said to be evolved. The
hydrogen peroxide can be replaced in any of the preceding mixtures by a peroxide
of the alkalies or alkaline earths. L. Wolter (1908) recommends a fused mass of
equal parts of sodium peroxide, and potassium nitrate with one-eighth of its weight
of magnesia. Hydrochloric acid is dropped on the coarsely pulverized mass.
L.Santi recommends warming a solution of ammonium chloride with barium peroxide :
2Ba02+4NH4Cl->2BaCl2+4NH3+2H20-|-02, for a steady stream of oxygen.
G. Neumann 20 similarly charges a Kipp's apparatus with cubes made from barium
dioxide, pyrolusite, and gypsum in the proportions 2:1:1, together with hydro-
chloric acid (specific gravity 1*12).
There are numerous other reactions in which oxygen is evolved — e.g. a mixture
of steam and chlorine 21 passed through a red-hot tube gives a mixture of hydrogen
chloride, HCl, and oxygen : 2Cl2+2H20=4HCl+02. According to D. Muller,
a temperature of 120° suffices for the reaction. Chlorine water decomposes into
oxygen and hydrogen chloride when exposed to sunlight. The finely-divided metals
of the platinum family decompose chlorine water even in darkness, forming
oxygen and hydrogen chloride ; according to C. F. Schonbein,22 bromine and
iodine water do not decompose in this way in darkness.
According to J. L. Gay Lussac, dry chlorine will displace the oxygen from certain
oxides. This is the case, for instance, with the oxides of the alkaline earths, lead,
or cadmium. If silver hydroxide be placed in a jar of chlorine gas, the oxygen and
chlorine will change places.23 So also if commercial " chloride of lime " which
contains much calcium hypochlorite, CaOCl2, be heated to redness : 2CaOCl2
=2CaCl2+02. According to H. St. C. Deville and H. Debray, a kilogram of the
chloride of lime will furnish 40 to 50 litres of oxygen contaminated with a little
chlorine which can be removed by washing the gas with soda lye. E. Mitscherlich 2*
has also shown that a fairly steady stream of oxygen can be obtained by heating
a mixture of a concentrated solution of bleaching powder, and a small quantity of a
nickel or cobalt salt, at about 85°. Other hypochlorites, as well as hypobromites,
can be employed. The cobalt or nickel salt forms a higher oxide which acts catalyti-
cally. The cobalt or nickel salt can be replaced by other metal oxides — e.g. oxides
of manganese, iron, or copper. A current of chlorine or bromine passed through a
boiling solution of caustic alkali or milk of lime, containing a salt of the metal, gives
a 90 per cent, yield of oxygen. A hypochlorite or hypobromite is first formed,
e.g.: 2NaOH+Br2=NaOBr+NaBr-f HgO ; and the hypobromite, NaOBr, then
decomposes : 2NaOBr=2NaBr+02. According to C. F. Schonbein, an aqueous
solution of iodic acid can be boiled without decomposition, but if platinum black
be present, it decomposes into iodine and oxygen.
By heating a mixture of calcium sulphate and silica, N. A. Helouis 25 obtained a
mixture of sulphur dioxide and oxygen, and a residue of calcium silicate. By
heating a mixture of sodium nitrate with twice its weight of zinc oxide, sodium
zincate, ZD(0Na)2, and a mixture of nitrogen with 71*4 per cent, of oxygen is formed :
2ZnO-f4NaN03=2Zn(ONa)2+2N2+502.
4. Processes in which oxygen is obtained from the atmosphere. — Many ingenious
processes — both chemical and physical— have been devised for the continuous
manufacture of oxygen, and inventors have been particularly sanguine — at first.
These processes are often of much theoretical interest ; but however interesting a
process may be, and however enthusiastic the inventor, an installation will have a
short life commercially if it cannot bring money into the pockets of the investors.
{a) Chemical processes. — These depend on the oxidation of a substance by air so
as to form a compound which gives up the oxygen and re-forms the original substance.
This can be again oxidized, and so the cycle can be continued indefinitely. Thus,
mercury can be oxidized by heating it in air at about 400°, and the resulting mercuric
oxide broken up into oxygen and mercury at about 600°. The oxidation is too
slow for the process to be industrially useful. J. T. A. Mallet (1865) 26 patented a
process for extracting oxygen from air by the alternate oxidation of cuprous chloride,
OXYGEN
355
1 800
4CuCl-f 02=2Cu20Cl2, by exposure to moist air ; and, subsequent deoxidation
of the cupric oxychloride by heating to dull redness. C. M. Tessie du Motay and
C. E. Marechal (1866) 27 heated a mixture of pyrolusite and caustic alkali at a high
temperature while exposed to air, oxygen from the air takes part in a reaction
forming an alkali manganate — say Na2Mn04 ; the manganate then gives up its
oxygen when heated in a current of steam, and the residue again forms manganite
when heated in air. This process has not proved a great success although many
works were erected and numerous modifications have been patented. G. Kassner
(1889) 28 heated a mixture of chalk and lead oxide in contact with air whereby calcium
plumbate, Ca2Pb04, is formed. When this is treated with potassium or sodium
carbonate, caustic alkali remains in solution and calcium carbonate and lead peroxide
are precipitated : Ca2Pb04+2Na2C03+2H20=4NaOH+2CaC03-f PbOg. When
the precipitate is dried, and heated to about 500°, oxygen is evolved, and the residue is
ready for the regeneration of calcium plumbate. The caustic lye obtained as a by-
product is claimed to make the process economical ; but this is doubtful since the
process has had no commercial success in spite of numerous modifications.
Oxygen was formerly made on a manufacturing scale by the barium peroxide
process of L. Q. and A. Brin (1880). This depends upon a very interesting reaction
discovered by J. B. J. D. Boussingault 29 in 1851. When barium oxide — BaO is
heated in air to about 500°, it is rapidly oxidized to barium dioxide : 2BaO+02
=2Ba02. If the barium dioxide be heated to a still higher temperature, 800°, the
oxygen is given off and barium oxide remains as a
residue : 2Ba02=2BaO+02. The phenomena attend-
ing the decomposition of barium peroxide in closed
vessels are quite analogous with the decomposition of
mercuric oxide except that non-volatile barium oxide
appears in place of volatile mercury. The curve. Fig.
2, can therefore be readily interpreted. Some other
oxides behave in a similar manner. For example,
as found by H. Debray and A. Joannis, and by G. H. ^^^q- j^q~ iooq-
Bailey and W. B. Hopkins, cupric oxide, CuO, at a red Temperature **"
heat forms cuprous oxide, CU2O ; and re-forms cupric ^^^ 2.— EquiUbrium Pressure
oxide when heated in air at a lower temperature. Curve of Barium Peroxide.
J. B. J. D. Boussingault tried to apply the barium
peroxide reaction industrially, but it was found that after the baryta had been
oxidized about a dozen times, it lost its power. M. Gondolo claimed to have used
the barium oxide over a hundred times without deterioration, by mixing the barium
oxide with a little lime or magnesia, and potassium manganate. The cause of the
trouble was recognized by MM. Brin Freres in 1879. The barium oxide can be
reoxidized and used over and over again, provided the air be freed from carbon
dioxide, organic matter, dust, and any substance which forms a compound
with barium oxide which is not decomposed under the given conditions. The
regulation of the temperature offered practical difficulties which were overcome
by keeping the temperature constant in the vicinity of 700°. Barium oxide is then
transformed into the dioxide if the pressure be about 2 kilograms per sq. cm. — nor-
mally the atmospheric pressure is 1 '033 kilograms per sq. cm. The peroxide is decom-
posed into the oxide and oxygen at the same temperature under a reduced pressure —
about 0*05 kilogram per sq. cm. The gas pumped off under these conditions
contained about 90-96 per cent, of oxygen, and 4-10 per cent, of nitrogen.
After an industrial life of nearly 20 years, Brin's process failed to compete
successfully against the newer and cheaper method of preparation by the fractional
distillation of liquid air. The relative costs of the two processes are said to be :
3s. 6cZ. per 1000 cub. ft. for the liquid air process against from 75. to 125. for Brin's
process. The same amount of oxygen costs I65. Sd. when manufactured by elec-
trolysis with electrical energy at \d. per unit. The oxygen obtained from liquid air
is about 98 per cent, purity ; that from the barium process about 94 per cent.
—
n
Fo
3a 0^
V
/
rmati
on
J
J
m
/
Wrm,
/
'ik
m~m-_
'd
356 INORGANIC AND THEORETICAL CHEMISTRY
purity. In 1907, the British Oxygen Co. made about 30,000 cub. ft. of oxygen per
day by the liquid air process, and in 1917, about 1,000,000 cub. ft. per day.
[b) Mechanical and physical processes. — Oxygen is more soluble than nitrogen
in water ; and J. T. A. Mallet 30 patented a process for extracting oxygen from the
atmosphere by pumping o£E the air dissolved under pressure by water. The product
was again dissolved under pressure in water ; and again pumped from the liquid.
By repeating the operation eight times, a gas containing 91 per cent, of oxygen was
obtained. The use of liquids other than water — e.g. glycerol — have been patented
as solvents ; but these processes have not been commercially successful. Although
the ratio of oxygen to nitrogen dissolved by the liquid may be greater than with
water, the actual amount dissolved may be less, and, according to G. Claude.^i celui
rend illusoire leur usage. Similar remarks apply to the greater adsorption of
oxygen from air by wood charcoal than is the case with the absorption of
nitrogen.32 Repeated absorption followed by the expulsion of the adsorbed gases
by heating under a reduced pressure furnishes a product rich in oxygen.
P. Margis 33 obtained a 95 per cent, oxygen by the repeated diffusion of air through
indiarubber membranes — oxygen passes through the rubber much faster than
nitrogen. The fractional distillation of liquid air furnishes most of the oxygen gas
for commerce.
5. Processes dependent upon the electrolysis of water. —Nearly all the oxygen on
the market is now obtained by the liquid air process. Comparatively little oxygen
is obtained by the electrolytic process, since this is profitable only when the hydrogen
can be readily sold at good prices. 34 As indicated in connection with the electro-
lytic preparation of hydrogen, one ampere of electricity decomposes 0*335 grm. of
water per hour, and liberates 0*0373 grm. or 0*414 litre of hydrogen, and 0*298 grm.
or 0*207 litre of oxygen. By converting the number of calories involved during the
formation of water, and converting into equivalent electrical units, it follows from
Kelvin's rule that 1*5 volts are needed for the decomposition of water. In practice
2*5 volts is the minimum employed ; and with this voltage 12 kilowatt hours are
needed to furnish a cubic metre of oxygen and two cubic metres of hydrogen ; in
practice 12 to 14 kilowatt hours are consumed in producing these amounts of hydro-
gen and oxygen. This works out at about 16.s'. Sd. per 1000 cub. ft., when the
electrical energy costs ^d. per unit. A. d'Arsonval installed a plant for electrolytic
preparation of oxygen about 1885. He used as electrolyte a 30 per cent, solution
of sodium hydroxide with sheet-iron cylinders as electrodes. He used a current
density of 2 amps, per sq. decimetre, and enclosed the anode in a woollen bag to
serve as diaphragm. The hydrogen was not used. With 60 amps, about 100 or
150 litres of oxygen were obtained per diem. D. LatchinofE used an apparatus
in which the gases were under pressure. Other forms are indicated in connec-
tion with the electrolytic preparation of hydrogen. There have been several
explosions from the use of electrolytic oxygen, owing to the hydrogen getting
mixed therewith. 35
The oxygen obtained by any of these processes is pumped into steel cylinders
under a pressure of 100-150 atmospheres, and sold as compressed oxygen. The gas
may be obtained from the cylinders at any desired rate by regulating the valve.
References.
1 J. Priestley, Experiments and Observations on Different Kinds of Air, Birminglmm, 1700 ;
R. Perceval, Trans. Irish Acad., 4. 85, 1790 ; J. Scnebier, Ann. Chim. Phys., (1), 4. 261, 1790 ;
(1), 11. 89, 1791 ; J. Ingenhousz, Experiments on vegetables, discovering their great power of pvri-
fying common air in sunshine, etc., London, 1779 ; G. Bellucci, Ber., 8. 905, 1875 ; J. H. Jeans,
The Dynamical Theory of Gases, Cambridge, 356, 1916.
2 T. Engelmann, Bot. Zeit., 46. 66, 677, 693, 709, 1888.
3 H. Draper, Amer. Journ. Science, (3), 14. 89, 1877 ; J. C. Draper, ib., (3), 16. 256, 1878 ; C. A.
Young, ib., (3), 4. 356, 1872 ; J. Trowbridge, Phil. Mag., (5), 41. 450, 1896 ; ib., (6), 4. 156, 1902.
4 J. Philipps, Der Sauerstoff, Berlin, 1871.
5 J. Lang, Pogg. Ann., 118. 282, 1863.
OXYGEN 357
« S. White, Brit. Pat. No., 12536, 1849 ; L. Mond., ib., 2566, 1862 ; L. T. Thome, Journ. Soc.
Chem. Ind., 8. 82, 1889 ; 9. 246, 1890.
' H. St. C. Deville and H. Debray, Compt. Bend., 51. 822, 1860 ; C. Winkler, Dinglefs Journ.,
223. 408, 1877 ; W. S. Squire, Brit. Pat. No., 3278, 1875 ; H. A. Achereau, ib., 668, 1867.
« R. Bottger, Journ. prakt. Chem., (1), 104. 316, 1867.
» G. Rousseau, Compt. Mend., 103. 261, 1886 ; C. M. T. du Motay and C. R. M. de Marechal,
Dingler's Journ., 196. 230, 1870 ; J. H. Parkinson, Chem. Ztg., 6. 802, 1892.
1" W. Guertler, Zeit. anorg. Chem., 38. 456, 1904; 40. 253, 1904.
" G. F. Jaubert, Compt. Rend., 134. 778, 1902.
12 H. J. Turner, Amer. Chem. Journ., 37. 106, 1907^ G. F. Jaubert, Rev. Oen. Chim., 7. 365,
1904.
13 A. Vogel, Journ. prakt. Chem., (1), 1. 446, 1834 ; C. Winkler, ib., (1), 98. 340, 1866.
1* W. H. Balmain, Journ. Pharm., 2. 499, 1842.
15 J. Volhard, Liebig's Ann , 253. 246, 1889 ; G. F. Jaubert, Compt. Rend., 134. 778, 1902 ;
Bull Soc. Chim., (3), 27. 566, 1902.
i« R. Bottger, Journ. prakt. Chem., (1), 107. 48, 1869 ; M. Tonneau, Repert. Pharm., 45. 304,
1893.
17 G. Kassner, Chem. Ztg., 13. 1302, 1338, 1889 ; Zeit. angew. Chem., 5. 448, 1890 ; A. Gawo-
lowsky, Pharm. Ztg., 35. 702, 1890 ; H. le Chatelier, Compt. Rend., 117, 109, 1893.
18 J. Robbins, Pogg. Ann., 122. 256, 1864.
19 A. Baumann, Zeit. angew. Chem., 4. 79, 1890 ; L. Santi, BoU. Chim. Farm., 43. 673, 1904 ;
L. Wolter, Chem. Ztg., 32. 1067, 1908.
20 G. Neumann, Ber., 20. 1584, 1887.
21 0. Binks, Brit. Pat. No., 1563, 1860 ; J. L. Gay Lussac and L. J. Thenard, Recherches
physico-chimiques, Paris, 2. 143, 1811 ; D. Muller, Compt. Rend., 40. 906, 1856 ; L. Weber, Pogg.
Ann., 112. 619, 1861.
22 C. F. Schonbein, Journ. prakt. Chem., (1), 98. 76, 1866.
23 J. Schie], Uebig's Ann., 132. 322, 1864.
2* E. Mitscherlich, Lehrbuch der Chemie, Berlin, 2. 143, 1847 ; E. T. Kirkpatrick, Brit. Pat.
No., 1300, 1870 ; T. Fieitmann, Liebig's Ann., 134. 64, 1865 : F. Stolba, Journ. prakt. Chem.,
(1), 97. 309, 1866 ; G. Deniges, Journ. Pharm. Chim., (5), 19. 303, 1889 ; G. F. Jaubert, German
Pat. D.R.P., 157171, 1902.
25 N. A. Helouis, Ber., 15. 1221, 1882 ; J. H. Pepper, Dingier' s Journ., 167. 39, 1865.
26 J. T. A. Mallet, Brit. Pat. No., 2934, 1865 ; 3171, 1866.
27 C. R. M. de Marechal and C. M. T. de Motay, Brit. Pat. No., 85, 1866 ; Dingler's Journ.,
196. 230, 1870 ; M. Dutremblay and M. Lugan, Journ. Pharm. Chim., (6), 6. 392, 1897 ; F. A.
Bowman, Patentblatt, 12. 1035, 1891 ; J. H. Parkinson, ib., 13. 597, 1892 ; G. Webb and G. H.
Rayner, ib., 14. 720, 1893 ; F. Fanta, ib., 15. 318, 1894.
28 G. Kassner, Dingler's Journ., 274. 135, 183, 226, 270, 1889 ; 278. 468, 1890 ; Chem. Ztg., 17.
1242, 1893 ; 22. 225, 1898 ; G. L. Schaefer, ib., 24. 564, 1900.
29 J. B. J. D. Boussingault, Ann. Chim. Phys., (3), 35. 5, 1851 ; (5), 19. 464, 1880 ; L. Q. and
A. Brin, Brit. Pat. No., 1416, 1880 ; Mem. Soc. Ingen. Civ., 104. 450, 1881 ; M. Gondolo, Compt.
Rend., 66. 488, 1868 ; F. C. G. Muller, PJiarm. Ztg., 34. 665, 1889 ; B. Gerdes, Zeit. comp. fluss.
Guse, 2. 5, 1898 ; H. Debray and A. Joannis, Compt. Rend., 99. 585, 1884 ; 100. 999, 1885 ;
G. H. Bailey and W. B. Hopkins, Chem. News, 61. 116, 1890.
3« J. T. A. Mallet, Dingier' s Journ., 199. 112, 1871 ; Brit. Pat. No., 2137, 1869 ; C. W. Harrison,
ib., 435, 1873 ; A. Stamm, ib., 8285, 1884 ; N. A. Helouis, ib., 2080, 1891 ; Ber., 15. 1221.
1885.
31 G. Claude, Compt. Rend., 131. 447, 1900.
32 M. Montmagnon and M. de Laire, Bull. Soc. Chim., (2), 11, 261, 1869 ; J. Dewar, German
Pat. D.R.P., 169514, 1905.
33 P. Margis, Deut. Ind. Ztg., 23. 314, 1882 ; N. A. Helouis, Brit. Pat. No., 2080, 1881 ;
T. Graham, Compt. Rend., 63. 471, 1866.
3* A. d'Arsonval, Elektrotech. Zeit., 197, 1891 ; D. Latchinoff, Brit. Pat. No., 15935, 1888 ;
M. Lefebvre, ib., 1045, 1859.
35 E. Bosshard and A. Hauptli, Zeit. angew. Chem., 18. 1531, 1905 ; J. C. A. S. Thomas and
F. H. van Leent, ib., 15. 1236, 1902 ; A. Fraenkel, Mitt. tech. Gewerbe Wien, 160, 1907.
§ 5. Catalysis
Materializing abstractions is a vice of thought. — H. S. R. Elliott.
The action of manganese dioxide on the decomposing potassium chlorate is very-
curious. It acts as a stimulant. We do not know precisely how the manganese
dioxide does its work, although we can form a rough idea of what is taking place.
Many other oxides act similarly, but not quite so vigorously — e.g. ferric, copper,
358 INORGANIC AND THEORETICAL CHEMISTRY
cobalt, or nickel oxide, vanadium pentoxide, V2O5. tungstic oxide, WO3, uranic
oxide, U3O8, may be used in place of manganese oxide. It is quite a common thing
to find that the speed of reactions is accelerated or retarded by the presence of a
foreign substance whose composition at the end of the reaction is the same ^s it was
at the beginning. For example, in the combustion of hydrogen, platinized asbestos
or moisture may act as catalytic agents ; similarly metallic oxides stimulate the
decomposition of mercuric oxide, potassium chlorate, etc. Indeed, there are probably
few chemical reactions, if any, which are not affected by the presence of a catalytic
agent. These agents are conveniently grouped together as catalytic agents, and the
general phenomenon is called catalysis ; if the catalytic agent retards the speed of the
reaction, the phenomenon is conveniently called negative catalysis. It must be
clearly understood that catalysis is simply a term for grouping those reactions
whose speed is modified, or for those reactions which can be started by the presence
of a small amomit of a substance which is found to possess, at the end of the
reaction, the same chemical composition as it had at the beginning. The
catalytic agent may be chemically affected by interaction with the products of the
reaction, etc. W. Ostwald ^ ingeniously compares the action of a catalytic agent
with the action of oil on a machine, or of a whip on a sluggish horse. W. Ostwald,
and his followers, beUeve that the reaction must be actually in progress before the
catalytic agent can act ; although if W. Ostwald's analogy be pursued it runs against
his hypothesis, for we know that friction may be so great as to stop the running of a
machine, when a little lubricating oil would have prevented the stoppage. Ostwald's
limitation is quite arbitrary, and, so far as we can see, does not agree with all
the facts.
Some have tried to evade the difficulty either by refusing to recognize it, or by
reserving the term catalytic reactions for those reactions whose speed is merely
accelerated by the catalyst ; and using the term trigger reactions for those reactions
which do not start unless their potential energy is released by contact with
another substance. Of course, the introduction of a new term does not remove the
difficulty. Again, different products may be obtained by a reaction with and
without a catalytic agent ; and further, different catalytic agents, with the same
reacting materials, may furnish different end-products. So far as the evidence goes :
Catalytic agents can not only start, accelerate, or retard the speed of chemical
reactions, but they can also in some cases direct or determine the course of a
reaction.
Francis Bacon long ago cautioned us against allowing words to govern thought
instead of thought governing words. The word catalysis itself explains nothing.
To think otherwise would lay us open to Mephistopheles' gibe :
A pompous word will stand you instead
For that which will not go into the head.
This means that too much trust must not be placed in words. It is just when ideas
fail that a word comes in most opportunely. There is no difficulty in covering an
obscure idea by a word so that the word appears to explain the idea. In passing
back from the word to the idea, it becomes easy to believe that the " subjective
abstraction has an objective existence," or that because there is a word, some-
thing real must lie behind the word. These remarks about the term catalysis
might be applied, mutatis mutandis, to many of the terms in common use
in chemistry — passive resistance, chemical affinity, the ions of the ionic theory,
adsorption, colloids, etc.
References.
1 W. Ostwald, Die Schule der Chemie, Braunschweig, 1. 88, 1903.
OXYGEN 359.
§ 6. Consecutive Reactions
Theories are abstractions which, while they place in rehef that which is important for
certain fixed cases, neglect almost necessarily, or even disguise, what is important in other
cases. A theory always puts in place of a fact something different, something more simple,
which is qualified to represent it in some certain aspect, but for the very reason that it is
different does not represent it in other aspects. — E. Mach (1892).
The representation of a chemical reaction by means of an equation emphasizes
the character of the initial and of the end products of the reaction, but it conveys
no idea of the mechanism of the reaction — how the different materials interact to
give the final products. With the growth of knowledge, reactions represented by
the older chemists by simple equations resolve themselves into reactions of greater
and greater complexity. The regular type of chemical equation shows but the
beginning and end of the reaction. Such equations are sometimes considered to
represent " the essential and determining features " of the reaction, because they
indicate what might be called the main products of the reaction, and they are accord-
ingly used in the arithmetic of chemistry. However, chemists are continually
striving to obtain a completer view of the real mechanism of a reaction. The
truth, not simplicity, is the ultimate object of their quest. There can be no doubt
that quite a number of intermediate stages temporarily subsist before the drama
of the reaction closes with the final act — the formation of the end products. There
is plenty of evidence leading us to infer the existence of a kaleidoscopic sequence of
changing scenes during the progress of what are usually considered simple reactions.
Some suppose that water has no more right to representation in the chemical equation
than the glass of the vessel in which the reaction occurs. As we progress in our
studies, we shall find that water profoundly modifies the properties of most substances
with which it is in contact. This, said C. L. Berthollet (1803), is a striking illustration
of the effect of words on the ideas we form, and even on the results of observation.
We begin by considering a solvent as the liquid employed in making solutions,
and that these in turn are mere mixtures of solvent and the dissolved substance ;
consequently, attention is rarely directed to the action of the solvent under other
conditions because in them it retains the name of solvent. It must, however, never
be forgotten that all the substances present in a reacting system exercise an action,
and if there are circumstances in which the solvent may be neglected, there are others
in which it contribates efficaciously to the result.
The fate of the molecules of manganese dioxide. — As a result of quite a number
of experimental investigations on the decomposition of potassium chlorate ^ and a
study of the available circumstantial evidence, we are able to get, in imagination, a
peep behind the curtain which hides the course of the reaction. Firstly, it is not
quite correct to say that the manganese dioxide is not changed in any way during
the reaction because a microscopic examination of the manganese dioxide, before
and after the reaction, shows that it has undergone a physical, if not a chemical,
change — crystalline manganese dioxide has apparently become amorphous. The
manganese dioxide does appear to take part in the reaction in spite of the fact
that it has the same chemical composition at the end as it had at the beginning.
Secondly, the manganese dioxide is probably oxidized by the decomposing chlorate
to form one of the unstable higher oxides of manganese, but exactly what oxide
we do not know. This uncertainty is expressed by writing the unknown oxide
MnOw+2j where the numerical value of n is not known with certainty. This
stage of the reaction can then be represented by the equation :
i7iKC103+Mn02=>KCl+MnOn+2 (1)
Thirdly, the unstable oxide produced by the oxidizing action of the potassium
chlorate probably breaks down almost as soon as it is formed, regenerating the man-
ganese dioxide, and liberating free oxygen :
2MnOn+2=2Mn02+w02 (2)
360 INORGANIC AND THEORETICAL CHEMISTRY
The manganese dioxide so formed is again oxidized, and the oxide again decomposed
regenerating manganese dioxide anew. This cycle of changes continues until the
potassium chlorate is all decomposed. The opening and closing scenes are repre-
sented :
2KCiO8[+MnO2]=2KCl[+Mn0j]-f302
Equations (1) and (2), expressed in the most general form, indicate that we are
dealing with a reaction in which
A->M aJid M->B
where A and B respectively denote the initial and final products of the reaction, and
M the intermediate products. In the reaction just considered, M is represented by
MnO„+2. Under the prevailing conditions, A does not form B directly. Consecu-
tive reactions are those in which intermediate products are produced which do not
necessarily appear as final products in the reaction. Consecutive reactions occur
in stages ; one stage must be in progress before another can start. The speed of
formation of B from A obviously depends on the speed of the intermediate reactions.
If the reaction A->M be very rapid, and M->B be very slow, the intermediate product
M will accumulate in the system, and could be recognized and probably isolated.
Several examples are known. On the other hand, if A->M be very slow, and M->B
be very fast, it would be hopeless to look for intermediate products, and the evidence
in support of the assertion that the reaction involves a sequence of consecutive or
intermediate reactions must be circumstantial, not direct proof. It will be obvious
that the same reasoning must apply in a longer series of intermediate reactions, say,
A->M ; M->N ; N->B. Similarly, one or more of the intermediate reactions
might be a concurrent reaction or an opposing reaction.
The favourable influence of some inert powders — powdered glass, sand, and china
clay — shows that the effect may be in part due to an action similar to the effect of
finely-divided particles in promoting the evolution of gases from liquids. There is
evidence to show that all catalytic agents do not act in the same way ; thus the acid
oxides of vanadium, uranium, and tungsten, and chromic oxide, phosphoric acid, and
phosphorus pentoxide favour the evolution of chlorine ; 2 oxide of silver, and the
dioxides of lead and barium favour the formation of perchlorate.
References.
^ C. F. Schonbein, Journ. prakt. Chem., (1), 65. 96, 1855 ; H. McLeod, Journ. Chem. Soc,
55. 184, 1889 ; 65. 202, 1894 ; 69. 1015, 1896 ; W. H. Sodeau, ib., 77. 137, 717, 1900 ; 79. 247, 939,
1901 ; 81. 1066, 1902 ; E. J. Mills and G. Donald, ib., 41. 18, 1882 ; E. J. Mills and J. SteVenson,
ib., 41. 23, 1882 ; G. J. Fowler and J. Grant, ib., 57. 273, 1890 ; H. Warren, Chem. News, 58.
247, 1889; V. H. Veley, ib., 58. 260, 1889; W. R. Hodgkinson and F. K. Lowndes, ib., 58.
187, 223, 309, 1889 ; 59. 53, 1889; A. Brunck, Ber., 26. 1790, 1893 ; Zeit. anorg. Chem., 10. 222,
1895; F. Bellamy, Monit. Scient, (4), 1. 1145, 1887; J. Scobai, Zeit. j)hys. Chem., 44. 319,
1903 ; E. Baudrimont, Compt. Bend.., 73. 254, 1871 ; Journ. Pharm. Chim., (4), 40. 161, 1871 ;
E. Jungfleisch, Bull. Soc. Chim., (2), 15. 6, 1871 ; W. Spring and E. Prost, ib., (3), 1. 340, 1889;
G. Krebs, Zeit. Chem., 13. 243, 1870 ; E. J. JVIills, Phil. Mag., (5), 23. 375, 1887 ; M. Berthclot,
Ann. Chim. Phys., (5), 10. 377, 1877 ; Compt. Rend., 85. 1219, 1877 ; E. Wiederhold, Pogg.
Ann., 116. 171, 1862 ; 118. 186, 1863 ; A. Wagner, Zeit. anal. Chem., 21. 508, 1882.
2 W. Spring and E. Prost, Bull. Soc. Chim., (3), 1. 340, 1889 ; J. G. Fowler and J. Grant,
Journ. Chem. Soc, 57. 273, 1890.
§ 7. Concurrent or Side Reactions
Compounda are not marked by nature with chemical formulae but by properties, and it
is by these we have to distinguish them.- — J. D. Henrichs.
Attention must be again directed to the curious way potassium chlorate decom-
poses when heated. When potassium chlorate, KCIO3, is heated, not only is
OXYGEN 361
potassium chloride, KCl, formed, but, as G. S. Serullas and N. A. E. Millon have
shown, just after the development of oxygen has begun, the residue contains a
considerable amount of potassium perchlorate, KCIO4. According to J. C. G. de
Marignac, when the chlorate has lost from 4 to 5 per cent, of oxygen, there is between
64 and 65 per cent, of perchlorate, and between 12 and 13 per cent, of chlorate in
the residue ; and when the chlorate has lost between 8 and 9 per cent, of oxygen,
the residue contains between 65 and 66 per cent, of perchlorate and no chlorate.
Hence, part of the salt is transformed into perchlorate and part into oxygen and
chloride. Contrary to the opinion of N. A. E. Millon, J. C. G. de Marignac could
find no sign of the formation of potassium chlorite, KCIO2, at any stage of the
process. 1 It will be observed that an indefinitely long array of possible equations
could be deduced for the thermal decomposition of potassium chlorate, for, as
J. Bottomley (1878) has shown, if potassium perchlorate and chloride, as well as
oxygen, be the end-products of the reaction, the equation can be written :
2mKC103=2nKC104+2(m-ri)KCl+(3m-4w)02
where m and n are any integers subject to the condition that 3m be not less than 4%.
For instance, if m be 11, and n be successively assigned values 1, 2, 3, ... 8, the
reaction can be symbolized by one of the following eight equations :
22KClO3=2KClO4+20KCl+29O2
22KC103=:4KC104+18KCl+2502
22KC103=16KC104 f 6 'kCI+oV
and similarly when other values are assigned to m. It must be emphasized that
most chemical equations represent unproved and simplified hypotheses as to the
course of reactions. A chemical equation should, if possible, summarize ascertained
facts, and symbols should not be treated as if equation-building were merely an
algebraic operation.
The following illustrates one of the many possible ways ^ of building equations to satisfy
the rules of chemistry only when the initial and final products are known. In the preparation
of oxygen by the action of sulphuric acid, H2SO4, upon potassium dichromate, suppose
the by-products are chromic sulphate, Cr2(S04)3, potassium sulphate, K2SO4, and water,
and that it is desired to find their relative proportions: Write a;K2Cr307+2/H2S04
'^zQT^{^0^)^+uK^^O^+vll^O+wOi. It follows that «=w (K) ; x=z (Cr) ; lx^v+2w (O) ;
2y=2v (H) ; y='Sz-\-u (SO4). There are here five algebraic equations and six unknowns ;
hence it is possible to solve these equations only in terms of any one of the unknowns, a:,
y, z, u, V, w- — say u. Accordingly, x=u ; y=4:U ; z=u ; v=4:U ; ic=^u. Obviously
each of the unknowns must be a positive whole number, and w=2 is the smallest number
which will satisfy this condition. In that case, x=2 ; y=S ; z=2 ; ti = 2 ; v=S ; w=3 ;
or 2K2Cr207+8H2S04->2Cr2(S04)3+2K2S044-8H20 + 362.
In the numerous systems of balancing equations, there is a temptation to suppose
that the result of the algebraic operation represents a real process. Chemical
equations cannot be demonstrated by the manipulation of chemical symbols. J. von
Liebig (1846) stigmatized this operation " a senseless form of jugglery." Practice in
the art of balancing equations according to algebraic rules may have limited uses,
but it is utterly bad if it conveys the impression that reactions must take place as
the equation demands. Facility in the art may thus display ignorance, not
learning. Accordingly, equation-building is not emphasized so much as formerly
in modern works on chemistry.
The fate 0! the molecules of potassium perchlorate. — When potassium chlorate
is heated, part of the chlorate decomposes into potassium cliloride and oxygen :
2KC103->2KCl+302, and part oxidizes another part of the chlorate into potassium
perchlorate, KCIO4 ; in symbols, KC103+3KC103->KCl-f 3KCIO4. These two
reactions proceed side by side — concurrently, yet independently. Measurements
of the relative proportions of potassium perchlorate and oxygen formed at different
362 INOKGANIC AND THEOEETICAL CHEMISTRY
temperatures show that the potassium perchlorate reaction proceeds nearly twice
as fast as the other reaction. The lower the temperature, the greater the relative
speed of the perchlorate reaction. Hence, as the potassium perchlorate accumulates
in the system, the molten mass becomes more and more viscid, and if the temperature
be below the melting point of potassium perchlorate (610°), the mass solidifies
when enough potassium perchlorate has accumulated in the system, even though
the temperature be higher than the melting point of potassium chlorate (340°).
When the temperature is raised high enough, the potassium perchlorate decomposes
into potassium chloride and free oxygen. Here again the opening and closing
scenes are represented by the equation : KC104=KCl+202 ; but the whole
reaction can be perhaps better represented by the scheme :
Between 340-610° Above 630°
Airnin >7KCl+3KC104->4KCI+602
bKUU3<52KCl+302
or generally remembering our ignorance of the molecular weight of the molecules
of solid potassixmi chlorate, etc.
Between 340-610° Over 610°
fm-4-..)KC10 <^i^KCl+|mKC104->mKCl+|m02
The final products of both reactions are potassium chloride and oxygen, and this
is the sole justification for representing the reaction by the equation 2KCIO3
->2KCl+302. The fact that no appreciable quantity of potassium permanganate is
formed when a mixture of potassium chlorate and manganese dioxide is decomposed,
shows that the catalytic agent particularly favours one of the two reactions.
This view of the mechanism of the decomposition of potassium chlorate by
heat shows how the relative proportions of potassium chloride and perchlorate, and
oxygen depend on the temperature, and almost an infinite number of equations
are possible. This must be borne in mind when reading many text-books, for the
reaction is often represented by a set of complex equations — e.g. P. F. Frankland
and J. Dingwall 3 represented the reaction by 8KC10o=:5KC104-l-3KCl-|-202 at a
moderate heat, followed by 2KC103=KC104-l-KCl-{-02 at a higher temperature ;
and finally, at a still higher temperature : KC104=KCl+202. It can be shown
that all so far proposed are special cases of the simple equations described in the
text.
The cyclic reactions between the manganese dioxide and potassium chlorate
proceed rapidly at a temperature much lower than that at which the perchlorate
reaction has acquired an appreciable velocity. In fine, the catalytic agent accele-
rates at least one of the two concurrent reactions. It must not be supposed that
the above outline gives a complete representation of this remarkable reaction.
The products of the reaction may interact with themselves or with the catalytic
reagent. In some cases part of the oxygen comes off as ozone, and the products of
the reaction may contain a little chlorine. Traces of potassium permanganate
have been detected among the residual products. The chlorine and potassium
permanganate are probably formed by a reaction between the potassium chloride
and the manganese dioxide.
References.
1 G. S. S^rullas, Ann. Chim. Phys., (2), 45. 204, 270, 1830 ; N. A. E. Millon, ih., (3), 7. 298,
1843 ; J. C. G. de Marignac, Liehig's Ann., 44. 13, 1842.
2 J. Bottomley, Proc. Manchester Lit. Phil. Soc, 17. 94, 1899 ; Chem. News, 56. 277, 1887 ;
37. 110, 1878 ; W. J. Karslacke, ib., 96. 41, 1907 ; 0. C. Johnson, ib., 42. 51, 1880 ; W. Ackroyd,
ib., 82. 154, 1900 ; H. C. Madan, ib., 51. 265, 1885 ; J. C. Waddell, ib., 101. 253, 1910 ;
A. L. Taturn, Western Chem. Met., 5. 135, 1908 ; E. E. Junderich, School Science, 4. 93, 1904.
3 P. F. Frankland and J. Dingwall, Journ. Chem. Soc, 51. 274, 1887 ; F. L. Teed, ib., 51.
283, 1887 ; L. Maumen6, Chem. News, 53. 146, 1886.
OXYGEN 363
§ 8. The Physical Properties of Oxygen
Oxygen is at ordinary temperatures a colourless gas without smell. H. V.
Regnault (1847)i found that the weight of a litre of oxygen at 0° and 760 mm. at
Paris weighs 1 42980 grms., and this value corrected by J. M. Crafts (1888) for the
difference in volume between the full and the vacuous globe gave 1 '43011 grms.,
and corrected by J. Thomsen for 45° latitude at sea level, 1-42929. P. von Jolly's
value at Paris is 1-42892 grms., when corrected by Lord Rayleigh, 1-42971. E. W.
Morley found for the weight of a normal litre of oxygen 1-4290010-000034 grm.
A. Leduc also found 1-42939 grms. in 1891, and 1*4293 grms. in 1896, at Paris.
Lord Rayleigh (1893) gave 1-42952 grms. at n.p.t. ; and J. Thomsen found 1-42904
grms. at n.p.t. when reduced to 45° latitude and sea level. E. W. Morley (1896)
gave 1-42900 grms., J. Thomsen 1*42906 grms. at 45° and sea level. A. Jaquerod
and A. Pintza (1904) gave 1-4292 grms. ; A. Jaquerod and F. L. Perrot (1905),
1-42893 grms. ; P. A. Guye's calculation of R. W. Gray's determination (1905)
gave 1-42896; A. Jaquerod and M. Tourpaian (1911), 1-4290; and A. F. 0.
Germann (1915), 1-42906. The best representative value for the weight of one
litre of oxygen under these conditions is taken to be 1*42905 grms. A. F.
de Fourcroy, L. N. Vauquelin, and B. R. Seguin found the relative density of
oxygen to be 1-087 ; R. Kirwan, 1103 ; H. Davy, 1*088 ; W. Allen and W. A.
Pepys, 1-088 ; J. B. Biot and F. J. Arago, 11036 ; T. Thomson, 1-1056 ; T. de
Saussure, 1-1056; P. L. Dulongand J. J. Berzelius, 1*1026 ; H. Buff, 1-106; J.B.A.
Dumas and J. B. J. D. Boussingault, 1*1057 ; J. von Wrede, 1*1052. The gas in.
these early determinations was often saturated with moisture, and the measure-
ments afflicted with numerous errors so that they are only of historical value.
Later more accurate determinations are by H. V. Regnault, who found the relative
density of oxygen, air unity, to be 1*10564 ; P. von Jolly, 1-10505 ; J. P. Cooke,
1-10534; A. Leduc, 110506 (1891) and 1*10523 (1896); J. Giesen with the
microbalance, 1*1051 ; and Lord Rayleigh, 1*10530. The reported numbers for
the relative density of oxygen are between 15*861 and 15*96 when hydrogen
is unity. The best representative values are taken to be 1'10523 when air is
unity, and 15*87 when hydrogen is unity. Lord Rayleigh (1908-11) found no
appreciable difference in the density of oxygen prepared from different sources —
the electrolysis of water heating potassium chlorate or potassium permanganate —
outside the limits of experimental error. The vapour density of oxygen at —182°
is normal.
According to J. Dewar, the specific gravity of liquid oxygen at 760 mm. and
-182*5° is 1*1181 ; at -195*5°, 1*1700 ; and at -210*5°, 1*2386. J. Drugman and
W. Ramsay 2 give 1*1321 at —183-6° and 759 mm., and 1*1310 at —183*3° at 762 mm.
The specific volume of the liquid is then 08838 ; and the molecular volume 28*28.
According to G. le Bas, the atomic volume of oxygen at the critical temperature
is 26*5 ; the atomic volume of terminal oxygen in organic compounds is 7*2, and of
ethereal oxygen, 10*8. J. Dewar estimates the molecular volume at absolute zero
to be 21*21. E. C. C. Baly and F. G. Donnan give the interpolation formula
1*248874— 0-00481(r— 28) for the density of the liquid oxygen at a temperature
between 69*28° K. and 88*94° K. The specific gravity of solid oxygen at —252*5°
is 1*4256. J. Dewar represents these results by the expression 1*5154— 0*004420r,
where T denotes the absolute temperature ; otherwise expressed, liquid oxygen is
about 1*13 times as heavy as an equal bulk of liquid water at 4°. J. K. H. IngHs and
J. E. Coates have measured the specific gravity of mixtures of liquid oxygen and
nitrogen.
The weight of an oxygen atom is 16 x 1*56 x 10-^4 grm. ; the mean diameter of
the molecule of oxygen 3*62 XlO-^ cm. ; the mean free path of the molecule 6-3
X 10-6 cm. ; the number of molecules per c.c. is 2*75 x lO^^ ; the collision frequency
1*64 X 1029 per sec. per c.c. of gas ; and the molecular velocity is 42,500 cm. per sec.
364
INORGANIC AND THEORETICAL CHEMISTRY
The value of J. D. van der Waals' a=0-(X)273, and his 6=0-00142.3 L. L. Grunmach
calculates the molecular weight of liquid oxygen to be 41 51. The viscosity of
oxygen at 0° is 0-001873 ; A. von Obermayer gives 0-000189 ; and K. Schmitt,
0-0001926. K. L. Yen claims 7^=0-000204235 at 23° and 760 mm. with an accuracy
of 0-15 per cent. The viscosity increases with a rise of temperature such that at 6°
A. von Obermayer found that the viscosity is 0-0001928(1 +0-000283^) ; the observed
value at 20° is 0-0002060; at 99-74°, 0-0002485; and at 185-8°, 0-0002885, or
0-0001878(1 +0-003665^)0-787; W. Sutherland gave for the viscosity -q at 0°,
7j=rjo{{273+T)l(T-\-C)}{TI273)^, where C=127 ; F. Kleint gives C=136 ; H.
Markowsky, 138 ; and Lord Rayleigh, 128-2. According to 0. Volker,* the viscosity
decreases on a falling temperature, being 0*0001693 at —39-48° ; 0-0001474 at
-76-12° ; 00001128 at -1298° ; and 0-0001050 at -152-5°. The surface tension 5
of liquid oxygen is 13*074 ±0-066 dynes per cm., and the specific cohesion is 23-038.
The molecular weight by R. Eotvos' formula is 41-51, so that the molecules of the
liquid are rather more complex than corresponds with the formula O2. E. C. C.
Baly and F. G. Donnan, however, find the molecular surface energy a{Mv)^
=1-917(153-77— T), which gives a constant 1*917, not far from that required for
a normal liquid with molecules O2. According to P. L. Dulong,^ the velocity of
sound in oxygen gas at 0° is 317-17 metres per second at 0° ; and S. R. Cook finds
the velocity is 328*55 metres per second at 21° ; 282*4 at —28*4° ; 264-26 at —66-5° ;
210-12 at -137*5° ; and 173*92 at -1830°.
E. H. Amagat 7 has investigated the isothermal pv-curves for Boyle's law at 0°,
15*65°, 99*50°, and 199*50° for pressures from one to 1000 atm., and for 0° and 15*6°
up to 3000 atm. The values for 0° are :
Pressure
. I
500
1000
1500
2000
2500
2900 atm
Volume
. 1 -00000
0-00231
0-00174
0-00153
0-00141
000133
0-00128
pv
. 10000
11570
1-7360
2-2890
2-8160
3-3238
3-7120
150
50
/I ^
/
There is a minimum in the curve for ^v =0-91 35 near 150 atm. and 0° ;. for
^v=0*9920 near 150 atm. and 15*65° ; and the minimum was not observed at higher
temperatures. According to A. Leduc,
the compressibility of oxygen ^ at 11*2°,
~{d{pv)/dp)jpv, is 0*00076. A. Jaquerod
and 0. Scheuer give the compressibihty of
oxygen between 400 and 800 mm. pressure
as 0*00097 ; Lord Rayleigh gives 0*00094 ;
and D. Berthelot, 0*00085. In 1886, C.
Bohr measured the value of pv for pres-
sures less than normal, and concluded that
at a pressure of 0*70 mm., as indicated in
Fig. 3, there is a break in the curve showing
the relation between the pressure and the
product pv. C. Bohr's observation was
confirmed by E. C. C. Baly, W. Ramsay,
A. Campetti, and A. Batelli ; audit was thought to agree with the anomalous result
in the radiometer repulsion of oxygen at a pressure of 0*76 mm. which was found by
W. Crookes to be six to twelve times as great as that of nitrogen or of carbon monoxide
or dioxide ; and of H. Ebert's experiments on the dark space in vacuum tubes. It
might be thought that experimental work with all this backing could be accepted
with some degree of confidence. W. Sutherland ^ interprets the result as an effect
of the spontaneous change of oxygen into ozone, at a certain degree of rarefaction,
and with increasing rarefaction the transformation of oxygen into ozone continues
so as to keep the number of molecules of ozone per unit volume constant, when all
the oxygen molecules are used up, ozone alone remains, and it follows Boyle's law.
No chemical test enabled R. Threlfall and F. Martin to detect ozone in suitably
expanded oxygen. Neither Lord Rayleigh nor M. Thiesen could estabHsh Bohr's
0 5 10 15
Pressure in mm
Fig. 3 — C. Bohr's Critical Value in the
pv-curve of Oxygen at 0-7 mm. pressure.
OXYGEN 365
anomaly. Lord Rayleigh found that there is no deviation from Boyle's law
exceeding one part in 4000 for pressures between O'Ol and 150 mm. of mercury.
M. Thiesen attributes Bohr's anomalous result to some unrecognized experimental
error. Liquid oxygen is very compressible in comparison with many other liquids —
vide water ; A. Eucken found for the compressibility, j3, between 10 and 20 atm.,
j8==l-95 X 10~3 kgrm. per sq. cm. ; and calculated from the relation Cp—Cv=Tva^l^,
at 20-4° K., ^3=2-06 X 10-3 kgrm. per sq. cm.
The coefficient of thermal expansion of oxygen gas is given by P. von Jolly lo as
a^ =0-00367430 ; H. K. Onnes' value for the liquid oxygen is a==00157 at
-252-6°, or 20-4° K. The thermal conductivity of oxygen n at 0° is 0-00005694,
and between 7° and 8°, 0-0000563; S. Weber gives 5-768x10-5. The specific
heat at constant pressure,i2 Cp, is 0*2175 ; or from 20° to 440° 0^=0*224 ; and
from 20° to 630°, 0-230. The specific heat at constant volume is 0-1544 ; or, the
molecular heat at constant volume, at the absolute temperature T, is represented
by W. Nernst and H. von Wartenberg by 4-68+0-00026T cals. per gram-molecule.
For constant pressures the molecular specific heat may be taken as (7p=6-50
4-0-OOlOT, P. A. Miiller, and 0. Lummer and E. Pringsheim's measurement of
the ratio of the two specific heats of oxygen is 1*398 from 5° to 14° ; and 1-402 from
16° to 20°. G. N. Lewis and M. Kandall give for the best representative value of
the molecular heat of oxygen gas, (7j,=6-504-0-0010T, which gives at 0°, 6*77 ;
H. V. Regnault found 6-85; and M. Pier 6-89. The equation also gives 8-77 at
2000^, while M. Pier found 6-70. The higher values of L. Holborn and L. Austin, and
of A. Eucken were obtained indirectly with oxygen admixed with nitrogen. The
entropy of oxgyen gas at 25°, calculated by G. N. Lewis and G. E. Gibson, is 48-23
per gram- molecule, when the increase of entropy from absolute zero to the first
transition point T is <j)=jCpd log r=2-20 ; 17 -5/23-5 =0*74 is the entropy of the
transformation to the second form ; the increase of entropy when the temperature
of the solid rises from the first to the second transition temperature is 2-58 ; and
167 '4/42 "5 =3 -94 is the entropy of the second transition temperature ; and the
increase of entropy in rising from the second transition temperature to the melting
point is 2*62 ; the increase of entropy in passing from the solid to the liquid state
is 105-5/54-1=1-95; in passing from the melting to the boiling point, 6-52 ; in
passing from the liquid to the gaseous state, 1599/90*3=17-72 ; and in passing from
the boiling point to 298° K., 7-96.
For a long time oxygen proved incoersible to all attempts to liquefy it by
compression and cooling. J. Natterer, for instance, obtained no liquid at a pressure
of 1354 atm. On the 16th December, 1877, L. Cailletet, and a Httle later E. Pictet,i3
obtained the liquid ; and soon afterwards, S. von Wroblewsky, K. Olszewsky,
J. Dewar, and many others prepared the liquid in quantity and investigated its
properties. The critical temperature of oxygen is between —113° and —119° — say
—118° ; the critical pressure lies between 44-1 and 50-0 atm. — say 50 atm. ; and the
critical volume is 0-00426. If therefore the temperature is near but below or at —119°,
a pressure of 50 atm. will liquefy the gas ; and if the temperature exceeds —119°
no pressure, however great, can liquefy the gas. Liquid oxygen has a pale blue
colour. L. Grunmach found the boiling point of oxygen to be —182-65° at 762-22 mm.
pressure ; its vapour pressure i^ at —182-4° is 800 mm. ; at —193°, 200 mm. ; and
at —211*2°, 7*5 mm. K. Scheel gives for the boiling point of oxygen at a pressure
p mm. of mercury, — 183-0° +0-01258(^9 -760) -0-000007 (;? -760)2. Xhe vapour
pressure of liquid oxygen at T° absolute is given by log p=— 399/1 -\-l'75 log T
— 0-0051 r+6-9484. According to W. P. Juliusberger, the vapour pressure
between —212° and —119° is given by logio ;)=3-54595— 313-7^-1+1-40655
logio T mm. The latent heat of vaporization of liquid oxygen is 580 cals. per
gram, or 1856 cals. per gram-molecule. The last-named constant does not
vary linearly with temperature, since at —183° the latent heat of vaporization is
52*09 cals. ; at —201*5°, 59*10 cals. H. Alt's value at 760 mm. is 50*97 cals.
per gram; A. Eucken's value is 1599 cals. per gram-molecule. J. Dewar (1896)
366 INORGANIC AND THEORETICAL CHEMISTRY
cooled liquid oxygen by a spray of liquid nitrogen and obtained a hard pale blue
solid with a melting point —227° at 0*9 mm. pressure. According to A. Eucken,
the melting point is 54-1° K., or —218*9° ; and the latent heat of &sion of oxygen
is 105'5 cals. per gram-molecule.
There are signs of two transition points respectively at — 249*5° and —230*5°,
in the heating curve of solid oxygen. These are supposed to correspond with at
least three allotropic forms of the solid element. A. Eucken (1916) estimates that
the heat of the y to j8 transformation is 17*5 cals. and of the j3 to a 167'4 cals.
With the previous notation, therefore,
-249-5* -230-5' -218 9° -182-5°
a-Oxygensoiid~^-Oxygengoiid— y-OxygeUsoiid— Oxygeniiquid— Oxygeugas
It is very unusual to find the heat of fusion less than the heat of transition such
as is the case with solid oxygen. A. Eucken found the molecular heats of the
different forms of solid oxvgen to be : a-oxygen or oxygen III to be 0^3=2 '5,
C,=2*48 at 17-0° K., and 0^=4*42, a,=4-28 at 21*8° K. ; ^S-oxygen, or oxygen II,
Cp=5*92, C„=5-62 at 26*95° K., and 0^=^10*52 and C„=9*12 at 39*5° K. ; y-oxygen,
or oxygen I, Cp=ll*0 at 44*7° K., and 10*76 at 51*3° K. The value of C^ is here
calculated from the value of Cp from W. Nernst and E. A. Lindemann's relation
Cp~C^—ATCp, where ^ is a constant 3*2x10" 4. A. Eucken further gave for
liquid oxygen, 0^=12*81 at 57*4° K., and 12*62, at 73° K., Debye's function=113.
W. Wahl 15 found that oxygen becomes viscid when cooled near to the point of
solidification, and crystals grow in the viscid mass very slowly. If the cooling be
rapid, a vitreous glass is formed. The crystals which grow in the cooling mass are
dark between crossed nicols, but their crystalline form has not been established. If
the oxygen be cooled by boiling hydrogen, a fine-grained mass of double refracting
crystals belonging to the hexagonal system is formed. The transition point of a-
to j3- oxygen is not far below the melting point. Most of the physical constants of
solid oxygen, determined at the boiling point of hydrogen, refer to the second
crystalline form. J. Dewar found that when the charcoal vacuum was turned off,
the vapour pressure remained near 0*46 mm. for some time and then rose rapidly to
1*11 to 1*12 mm. and then remained constant during melting. W. Wahl suggests
that 0*46 mm. is the vapour pressure of the a- and j8-crystals at the triple -
point ; and 1*12 mm. the vapour pressure of the a-crystals, liquid and vapour.
The existence of two crystalline modifications of oxygen is interesting in view of
the polymorphism of sulphur, and the analogies between the compounds of sulphur
and oxygen. Ozone also is polymerized oxygen. The phenomena observed by
I. Langmuir i^ to be associated with the formation of atomic hydrogen are largely
duplicated when oxygen is substituted for hydrogen. The resulting atomic oxygen
reacts with tungsten even at liquid air temperatures.
The mean index of refraction of oxygen gas is less than that of any other gas ;
for white light /x=l "000270 ; for the C-ray, 1*000255 ; for the 6^-ray, 1*000294 ; for
the E-my, 1*000315 ; and for the green mercury line, /Lt=l*0002727. E. Mascart
(1877),i7 L. Lorenz (1880), H. C. Rentschler (1908), and C. and M. Cuthbertson (1909)
have determined the dispersion of oxygen. The latter find for wave-lengths, from 4861
to 6563, Cauchy's formula /x=l-fO-00026509A-i-f (l+7*33A-2 lO-H) holds good;
but the general results with Cauchy's formula are not very good. Better results are
obtained with a formula of the type used by W. Sellmayer, with the dispersion formula
^—l=cl{nQ^-n^), where Hq is the frequency of the free vibration which has received
a theoretical interpretation from the electron theory of P. Drude. The constants
are /x— 1=3*397 X 1027/(12804 X 1027— w). The index of refraction of liquid ox}^gen
for the D-ray is, according to G. D. Liveing and J. Dewar, 1*2236. According to
J. Dewar, the atomic refraction of the liquid for the D-Mne is 3*182, and this is not far
from J. H. Gladstone's value 3*0316 for gaseous oxygen. J. H. Gladstone gives
for the atomic refraction of singly linked oxygen in combination 2*8 ; and of doubly
linked oxygen 3*4 ; and J. W. Briihl 1*506 for hydroxylic oxygen ; 1655 for
OXYGEN 367
ethereal oxygen ; and 2-328 for carbonylic oxygen. J. F. Homfray calculates
the atomic refraction of quadrivalent oxygen to be 2*73. J. W. Briihl also gives
for the atomic dispersion Ry—Ra=0'01d for hydroxylic oxygen ; 0*012 for ethereal
oxygen ; and 0'086 for carbonylic oxygen. According to L. H. Siertsema,
E. Verdat's constant for oxygen at a pressure of 10 k.grams. per sq. cm. and 7°,
is 00002722A-1+000001915A-3 for wave-lengths between 0'423/x and 0-684/x.
For pressures between 38 and 100 atm., the constant changes proportionally with
the density of the gas.
In 1864, J. Pliicker and W. Hittorf i^ established the fact that one and the
same element can under different conditions produce more than one spectrum.
Several different spectra of oxygen have been recorded. A. Schuster distinguishes
four main spark spectra of oxygen : (i) The elementary line spectrum which appears
at the highest temperature to which oxygen can be subjected such as occurs when
the Leyden jar and air-break are introduced into the electric circuit. This spectrum
consists of a number of lines particularly in the more refrangible part of the spectrum,
(ii) The compound line spectrum appears at a lower temperature than the elementary
line spectrum, and predominates when the Leyden jar and air-break are removed
and the exhaustion of the tube is not very high ; if the exhaustion is high a special
spectrum from the negative glow is superposed on the four-line spectrum. There
are : one red line, two green lines, and one blue line as illustrated in Fig. 4. Ac-
cording to A. Schuster, these four lines have the wave-lengths 6156*86 (red) ;
5435-55 (green) ; 5329 41 (green) ; and 4367-62 (blue). With an increase of
pressure, the more refrangible lines widen first while the blue line remains sharp.
Red. Green Blue.
61.56-8 5435-5 5329 4 4367-6
llll|llll|llll|llll|llll|llll|llll|llll|llll|llll|llll|llim^
20 30 40 50 60 70 80 90 ■ 100 110 120 130 140 150 160
Fig. 4. — The Compound Line Spectrum of Oxygen in a Geissler's Tube.
(iii) The continuous spectrum appears at the lowest temperature at which oxygen
becomes luminous. According to E. Becquerel, an excess of oxygen in the oxy-
hydrogen flame produces a yellow colour probably due to the continuous spectrum
of oxygen, (iv) The negative-glow spectrum was first observed by A. Wiillner in
1872, and is always seen in the glow surrounding the negative electrode in oxygen.
It consists of five bands — three red, two green. The least refrangible red band is
so faint that it may escape observation, and the two red bands are so close that
with a small dispersion they appear as one line. With high optical powers the
two green bands can be resolved into a series of lines. A. Schuster further described
the appearance of a vacuum tube filled with pure oxygen as it is sparked while
being gradually exhausted :
At first the spark has a yellow colour, and the spectrum is perfectly continuous. Almost
immediately, however, four lines are seen in the capillary part above the continuous
spectrum. One of these lines is in the red, two are in the green, and one is in the blue.
The discharge still passes as a narrow spark throughout the length of the tube. In the
wide part the spectrum remains continuous, and it extends more towards the red than in
the capillary part. It seems as if the four lines had taken away part of the energy of the
continuous spectrum. As the pressure diminishes, these lines increase considerably in
strength, the spark spreads out in the wide part of the tube, and the intensity of the
continuous spectrum is, therefore, considerably diminished, while it still forms a prominent
part in the spectrum of the capillary part. When the pressure is small the continuous
spectrum decreases in intensity. At the same time the negative glow, with its own charac-
teristic spectrum, gradually extends through the negative half of the tube into the capillary
part. The continuous spectrum has now entirely disappeared ; the bands of the negative
pole and the four lines stand out on a perfectly black background. It is imder these
conditions that the change from the compound line spectrum to the elementary Ime
spectrum is best studied. The mere insertion of the Leyden jar, I find, makes hardly
368 INORGANIC AND THEORETICAL CHEMISTRY
any difference ; the jar does not seem to be charged at all. If, in addition to the jar, we
insert a movable air-break, which can be opened or closed at will, while we look through
the spectroscope, we shall be able to see alternately two perfectly distinct spectra. If the
air-break is closed, the four lines of the compound spectrum only are seen ; if the air-break
is opened, these four lines will disappear entirely, and the elementary line spectnmi will
come out.
The spectrum of oxygen is peculiar in that it does not show up clearly
in the presence of other gases ; for example, air in a vacuum tube gives the same
spectrum as nitrogen. All carbonaceous impurities should be excluded from
oxygen whose spark spectrum is under investigation, because the carbon is readily
oxidized to carbon monoxide at the high temperature and the spectrum of the
impurity may be the more brilliant, and entirely eclipse that of the oxygen. Several
descriptions of the carbon monoxide spectrum have been published ^^ which have
been attributed to oxygen. C. Runge and F. Paschen 20 reduced the line spectrum
of oxygen to two series of triplets and two series of doublets. All foui' series can
be represented by formulae of the type X''^=a'\-hn~'^—cn~^, where a, b, and c are
constants. J. J. Balmer's formula — vide hydrogen — is a special case of this more
general expression. B. Reismann found that when a Geissler's tube of oxygen is
excited by a direct current discharge the series, elementary line, and banded spectra
appear at the cathode, and only the series spectrum at the anode. J. Stark and
co-workers, and H. Wilsar have measured the spectrum of oxygen canal rays. The
Stark effect with the spectrum of oxygen has been examined by U. Yoshida.
The absorption spectrum of oxygen has attracted some attention because
certain dark lines of the solar spectrum — the so-called telluric lines — may be in
part derived from the absorptive power of atmospheric oxygen. The absorption
spectrum of oxygen is very feeble and it must be examined in a long tube, with
the highly compressed or liquefied gas — since the two are practically identical.
J. Janssen 21 used a tube 20 to 60 metres long, with the gas at 27 atm. pressure.
There are two strong absorption bands in the red corresponding with the A and B
Fraunhofer solar lines with wave-lengths from 6340 to 6225 ; a strong band in the
yellow from 5820 to 5730 ; a feeble band in the green at 5350 ; and a feeble one
in the blue at 4810. 0. C. Lester found two series of absorption bands in the
oxygen of the solar spectrum. J. Tyndall could scarcely detect any absorption of
the invisible heat radiations by oxygen at atmospheric pressure. W. Burmeister
found gaseous oxygen has no absorption bands in the infra-red.
The oxygen gas produced by heating potassium permanganate is positively
electrified.22 Liquid oxygen is virtually a non-conductor of electricity. For the
phenomena associated with the ionization of oxygen, see hydrogen. The mean
values of J. Zeleny's, A. P. Chattock's, and J. Franck's determinations of the
velocity of the positive and negative oxygen ions when the electric discharge is in
a field of 1 volt per cm. are respectively r32 and 1'83 cms. per second. The mean
value of J. S. Townsend's and E. Salleo's determinations of the diffusion coefficient
of the positive and negative ions per sq. cm. per sec. are respectively 0'0275 and
0'040. J. S. Townsend ^3 gives dn/dt=—33S0n^, where n denotes the concentration
of the ions, and dn/dt the velocity of combination of the ions to form ordinary mole-
cules. A. Erickson and P. Philips have studied the effect of temperature on this
reaction. The potential of the oxygen electrode is discussed later. The discharge
potential has been discussed in connection with hydrogen. W. C. Rontgen found the
minimum potential for a 4- point to be respectively 2402 and 1975 volts for pressures
205 and 110 mm., and J. Precht, 2800 volts for a -f point, and 2350 volts for a —
point with a pressure of 760 mm. A. L. Hughes and A. A. Dixon 24 found the ionizing
potential is dependent on the least energy necessary to ionize the molecules of a
gas by the impact of electrons, and amounts to 9 "2 volts for oxygen ; J. Franck
and G. Hertz found 9 volts ; F. M. Bishop, 9 volts ; and the value calculated by
K. T. Compton's formula F=0-194(^— 1)"* is 8*4 volts, where V denotes the ionizing
potential, and K the specific inductive capacity.
OXYGEN ^ 369
The dielectric constant of liquid oxygen 25 lies between 1'4:65 and 1-491 at —182°.
H. Rohmann gives the dielectric constant of oxygen gas at 0° and one atm. pressure as
r000547 ; according to A. Occhialini, oxygen gas at 13'5° and compressed to a density
of 35 has a dielectric constant 1 01831, and at a density 110, 1 -05843. The results
agree with Mossotti and Clausius' formula but not with (Z— l)/D=constant. In
1847, M. Faraday's experiments indicated that oxygen gas is diamagnetic, but it
was subsequently shown to be paramagnetic having a magnetic susceptibility of
+0-12x10-6 at 20° and 1 atm. ; and +6*2x10-6 at 16° and 40 atm. pressure
(volume units). E. Becquerel noticed the magnetic quahties of oxygen — more
particularly of wood charcoal saturated with adsorbed oxygen. The magnetic
qualities are greatly enhanced when oxygen is condensed in the liquid state. Liquid
oxygen then shares with iron, nickel, and cobalt the property of being magnetic.
If the magnetic moment of iron be unity, that of liquid oxygen is 0*001. When a
little liquid oxygen is placed in a cup between the poles of an electromagnet, the
liquid leaps up to the poles, and remains attached until all has evaporated. 26 If
the current is cut o£E while the oxygen is attached to the poles of the electromagnet,
the oxygen detaches itself. A thin test tube of liquid oxygen, suspended by a long
thread, will attach itself to an electromagnet, and fall away as the current is cut
off. The magnetic susceptibility of liquid oxygen 304x10-6 volume units or
241x10-6 mass units at —182° ; 280x10-6 at —208°. For the solid, at —253°
the magnetic susceptibility is 375x10-6 (mass units), and at —259°, 436x10-6.
At its freezing point therefore the magnetic susceptibility of liquid oxygen is nearly
1*3 times as great as that of solid oxygen.
The solubility of oxygen. — The solubility of oxygen in water is small ;
100 vols, of water absorb nearly 5 vols, of oxygen at 0°. The absorption coefficient
was measured by R. W. Bunsen 27 in 1855. L. W. Winkler (1891) and
C. Bohr and J. Bock (1891) measured the coefficient of absorption of water between
0° and 100° ; L. W. Winkler's results are rather lower than C. Bohr and J. Bock's.
The latter are here indicated along with oj, the weight of gas in grams taken up
by 100 grams of solvent at the indicated temperature and a total pressure — gas
plus the vapour of the solvent — of 760 mm. :
0° 4° 8° 12° 16° 20° 24°
B . . 0-04961 004496 0*04098 0'03732 0-03425 0-03171 0-02954
CO . . 0-00704 0-00637 0-00578 0-00526 0-00481 0-00443 0-00411
L. W. Winkler represented his results for the absorption coefficient j8 at 6°
between 0° to 30° by the empirical formula j8=0-04890— 0-0013413^+0-04283^2
-0-0629534^3; j. j. Fox's formula (1909) is j8=0-04924 -0 0013440^
+O-O42875202_o63924^3 ; c. Bohr and J. Bock's values between 25° and 100° are :
i8
a)
Unit volume of water increases 0-00115 unit by the absorption of one volume of
oxygen. Water is easily supersaturated with oxygen, and the excess is given off
with difficulty.28 W. E. Adenby and H. G. Becker have studied the rate of
solution of oxygen in water.
According to G. Geffcken,29 the solubility A of oxygen in acids solutions at
15° and 25°, when A for oxygen in water is 0*0363 (15°) and 0-0308 (25°) is :
Sulphuric acid.
15° 25°
0-0338 0-0288
0-0335 0-0251
According to A. Christoff, the solubility A of oxygen in 95-6 per cent, sulphuric
acid is 0*03166 between 17° and 20° ; in 0-06162 per cent, acid, 13*30 ; in
0-03582 per cent, acid, 15*61 ; and in pure water, 0*03046. The value of A for
fotassium hydroxide in |iV-solutions is 0*0291 (15°) and 0*0252 (25°), and for normal
VOL. 1, 2 b
25°
30°
40»
50°
60°
80°
100°
0-02904
0-02676
0-02326
0-02070
0-01893
0-01726
0-01679
0-00403
0-00368
0-00310
0-00263
0-00221
0-00135
0-00000
Hydrochloric acid.
Nitric acid.
15° 25°
15° 25°
lN-&cid .
. 0-0344 0-0296
0-0348 0-0302
2A^-acid
. 0-0299 0-0267
0-0315 0-0284
370 INORGANIC AND THEORETICAL CHEMISTRY
solutions, 0-0234 (15°) and 00206 (25°) ; with J^'-solutions of sodium hydroxide,
A=0-0288 (15°) and 0-0250 (25°); with normal solutions of sodium hydroxide,
A=0-0231 (15°) and 0-0204 (25°). Similarly, with JiV-solutions of potassium
sulphate, A=0-0294 (15°) and 0*0253 (25°), and with normal solutions, A=0-0237
(15°) and 0-0207 (25°). With JiV-solutions of sodium chloride, A=0-0308 (15°) and
0-0262 (25°) ; and with normal solutions, A=0-0260 (15°) and 0*0223 (25°). The
solubility of oxygen in salt solutions diminishes regularly with an increase in the
concentration of the salt. C. G. MacArthur measured the solubility of oxygen in
solutions of lithium, sodium, potassium, rubidium, caesium, ammonium, magnesium,
calcium, and barium chlorides ; sodium and potassium bromide ; potassium iodide,
nitrate, and sulphate ; and in sodium sulphate. The density determinations also
furnished data for calculating the degree of hydration of the salts. C. G. MacArthur
says that allowing for the hydration of the ions, the results show that particular
ions increase the solubiUty to a definite extent which is specific for a particular ion.
Sea water was found by F. Clowes and J. W. H. Biggs to dissolve a little more than
70 per cent, of the amount of oxygen dissolved by distilled water ; and the amount
dissolved diminishes regularly with the salinity of the water. According to J. S.
Maclaurin, the absorption coefficient of oxygen in aqueous solutions of potassium
cyanide is
KCy in 100 grms. solution
Absorption coefficient
The solubility of oxygen in organic solvents. — Oxygen is a little less than five
times as soluble in ethyl alcohol as in water. According to E. E. 0. Libarsch, the
solubility of oxygen in unit volume of water is 0-0289, and in aqueous alcohol at
20° and 760 mm.
Per cent, alcohol . . 9-09 16-67 28'57 33-33 50*00 66-67 80-0
A 0-0278 0-0278 00249 00267 00350 00495 0-056
The solubility thus decreases with increasing concentration of the alcohol ; reaches
a minimum with nearly 28 per cent, solutions, and again increases with increasing
concentration. The variation of the absorption coefficient with temperature for
99-7 per cent, solutions of ethyl alcohol is j8=0-2337 (0°), 0-2301 (5°), 0-2266 (10°),
0-2232 (15°), 0-2201 (20°), and 0-2177 (24°) ; he represents the absorption coefficient
P for temperatures, 6, between 0° and 23-4°, by ^=0-23370-0-0374688^,
-}-0*053288^2 . ^th methyl alcohol and acetone, the values of A are :
0° 5° 10° 20° 30° 40° !S0°
1
10
20
30
50 grms
0-029
0-018
0-013
0-008
0-003
A (methyl alcohol) .
0-31864
0-30506
0-29005
0-25374
0-21569
0-21569
0
A (acetone)
0-2997
0-2835
0-2667
0-2313
0-1935
0-1533
0
and M. G. Levi represents the values for methyl alcohol by A=0*31864— 0-002572<?
—0-042866^2 . and for acetone, A=0-2997— 0-00318^— 0-0412^2, With petrokmn,
the absorption coefficient at 10° is 0*229, and at 20°, 0*202. C. G. MacArthur
measured the solubility of oxygen in solutions of sugar, and he found signs of an
oxidation of the sugar. C. Bohr has measured the absorption of oxygen by blood.
The solubility of oxygen in solids. — Solids adsorb, or occlude, oxygen in an
analogous manner to hydrogen. ^o G. Neumann found that if the following metals
be heated to 450° in oxygen.
Silver. Gold. Platinum.
Volumes of oxygen occluded .... 4-1-54 32-8-48-5 630-770
per volume of metal. Palladium is oxidized 3i to Pd20, or a mixture of Pd20 and
PdO. According to C. Engler and L. Wohler, platinum, like palladium, forms a
true compound or a mixture of compounds : PtO and Pt02. According to
E. Goldstein,32 the platinum electrode of a Geissler's tube rapidly absorbs oxygen
when red hot. R. Lucas attributes the absorption of oxygen by platinum to the
presence of iridium ; pure platinum, says he, absorbs no oxygen. The case ot
10°
O'^
-10"
-50"
-100°
-150°
61-7
66-7
70-0
105-0
165-0
245-0 vol^.
OXYGEN 371
silver 33 is peculiar. Molten silver dissolves about ten times its volume of oxygen,
and gives it up again on cooling. In cooling, a solid skin forms on the exterior
surface ; as the interior cools, the gas bursts through the solid crust, driving out a
fepurt of the still fluid metal — the phenomenon is called the spitting or rockage of
silver. Molten platinum behaves in a somewhat similar way. A little oxygen
still remains dissolved in the cold metal, and this can be recovered by heating it to
redness in vacuo.
According to S. Kern,3* steel dissolves from 0*025 to 0'06 vol. of oxygen. H. V.
Regnault ^5 said that mercury dissolves a little oxygen, but E. H. Amagat found no
evidence of this between 0° and 100°, even at 420 atm. pressure. According to
F. Leblanc, molten litharge dissolves oxygen which is rejected when the oxide solidifies.
Charcoal^^ dissolves 9'25 its volume of oxygen at 12° and 724 mm. pressure, as
shown by T. de Saussure in 1814 and according to L. Joulin, 26 vols, at 0° and
2*36 atm. pressure, and 230 vols, at —185°. According to J. L. Baerwald, with
one volume of charcoal at 760 mm. pressure, the volume of oxygen absorbed at
different temperatures is
Oxygen absorbed
Neither platinum nor palladium show this remarkable increase in its own absorptive
power at low temperatures. According to G. Craig, the coke from lignite which
has been heated to redness and cooled with the exclusion of air, absorbs oxygen
with the formation of water.
The coefficient of difhision of oxygen into carbon dioxide is 0136, and of hydrogen
into the same gas, 0*538. T. Graham 37 has shown that oxygen travels through
indiarubber, 2 J times as rapidly as nitrogen. Hydrogen travels 5 J times as fast
as nitrogen. According to M. Berthelot, cold glass is impermeable to oxygen, but
at 650° he found a glass bulb lost 8 per cent, of the gas in 2 hours, and Jena glass
at 800° scarcely lost any gas in IJ hours. Oxygen does not diffuse through cold
silver, but it does if the metal be heated, say at 800° ; and L. Troost showed that
at this temperature about 1700 c.c. of gas will travel through a sq. metre of the
metal 1 mm. thick per hour, and 3300 c.c. if the plate is J mm. thick. The perme-
ability of hot silver to oxygen is connected with its power of occluding this gas. For
the diffusibility of oxygen through rubber, see hydrogen.
Liquid oxygen at —190° dissolves about 380 times its volume of nitrogen and
the boiling point is then changed to —188*8°. The suggested cause of the
discrepancies in the boiling point determined by different investigators is probably
due to the contamination of the liquid with nitrogen. Liquid oxygen indeed
rapidly absorbs nitrogen from the air^S—Fig. 25, Cap. XL J. K. H. Inglis and
J. E. Coates studied the densities and partial pressures of solutions of oxygen
and nitrogen, and J. K. H. Inglis the isothermal distillation of nitrogen and
oxygen. Liquid oxygen readily dissolves liquid fluorine.
References.
1 A. F. de Fourcroy, L. N. Vauquelin, and B. R. Seguin, Ann. Chim. Phys., (1), 8. 113, 183,
1791; (1), 9. 7, 29, 1791 ; T. de Saussure, ^'fi., (1),171. 259, 1809; Nicholson's Jour7i.,26. 161, 1810 :
W. Allen and W. A. Pepys, Phil. Trans., 97. 267, 1807; R. Kirwan, ib., 71. 7, 1781 ; T. Thomson,
Ann. Phil, 16. 161, 1820 ; Compt. Bend., 12. 1048, 1841 ; J. B. Biot and F. J. Arago, Mem,
Inst., 71. 7, 1781 ; H. Davy, Elements of Chemical Philosophy, London, 1812 ; R L. Dulong and
J.J. Berzelius, Ann. Chim.'Phys., (2), 15. 386, 1820 ; J. B. A. Dumas and J. B. J. D. Boussingault,
ib., (3), 3. 257, 184] ; H. Bufif, Pogg. Ann., 22. 242, 1831 ; H. V. Regnault, Mem. Acad., 21. 121,
1847 ; Compt. Bend., 20. 975, 1845 ; J. M. Crafts, ib., 106. 1662, 1888 ; A. Jaquerod and
A. Pintza, ib., 139. 129, 1904 ; A. Jaquerod and F. L. Perrot, ib., 140. 1542, 1905 ; A. Leduc, ib.,
113. 186, 1891 ; 123. 805, 1896; Ann. Chim. Phys., (7), 15. 29, 1898; Becherches sur les gaz,
Paris, 1898 ; J. Thomsen, Zeit. anorg. Chem., 12. 1, 1896 ; Lord Rayleigh, Proc. Boy. Soc, 50.
448, 1892 ; 53. 134, 1893 ; Phil. Mag., (6), 15. 746, 1908 ; (6), 21. 465, 1911 ; (6), 22. 563, 1911 ;
E. W. Morley, On the Density of Oxygen and Hydrogen and on the Batio of their Atomic Weights,
Washington, 1895; Zeit. phys. Chem., 17. 87,1895; 20. 130, 1896; P. von Jolly, Wied. Ann.,
372 INORGANIC AND THEORETICAL CHEMISTRY
6. 620, 1879 ; J. P. Cooke, Amer. Ch^m. Journ., 11. 509, 1890 ; J. Giesen, Ann. Physik, (4), 10.
830, 1903 ; R. W. Gray, Journ. Chem. Soc., 87. 1601, 1905 ; P. A. Guye, Journ. Chim. Fhys., 5.
203, 1907 ; A. F. O. Germann, Journ. Phys. Chem., 19. 437, 1915 ; A. Jaquerod and M. Tourpaian,
Arch. Sciences Oen^.ve, (4), 31. 20, 1911.
2 S. von Wroblewsky, Gompt. Rend., 102. 1010, 1886; K. Olszewsky, Sitzber. Akad. Wien, 12.
72, 1884 ; 14. 198, 1886 ; J. Drugman and W. Ramsay, Journ. Chem. Soc, 77. 1228, 1900 ;
G. le B&s,Chem. News, 115. 146, 1917; J. Dewar,i6.,73. 40, 1896; Proc. Boy. Soc, 73. 251, 1904;
A. Ladenburg and C. Krugel, Ber., 32. 46, 1415, 1899 ; E. C. C. Baly and P. G. Donnan, Journ.
Chem. Soc., 81. 907, 1902 ; J. K. H. Inglis and J. E. Coates, Journ. Chem. Soc, 89. 886, 1906.
» E. Hantzschel, Ann. Physik, (4), 16. 565, 1905 ; L. L. Grunmach, Sitzber. Akad. Berlin,
679. 1906.
* H. Markowsky, Ann. Physik, (4), 14. 742, 1904 ; A. Bestebneyer. ib., (4), 15. 423, 1904 ;
O. E. Mayer and P. Springmuhl, Pogg. Ann., 1^. 536, 1873 ; A. von Obermayer, Sitzber. Akad.
Wien., 71. 281, 1875 ; K. Schmitt, Ann. Physik, (4), 30. 398, 1909 ; P. Kleint, Verh. deut. phys.
Oes., 7. 146, 1905; Lord Rayleigh, Proc Boy. Soc, 67. 137, 1900; W. Sutherland, Phil Mag.,
(5), 36. 507, 1893; (6), 14. 1, 1907 ; 0. Volker, Land olt-Bor ostein's Physikalisch-chemische
Tabellen, Berlin, 99, 1912; K. L. Yen, Phil. Mag., (6), 38. 582, 1919.
« L. L. Grunmach, Ber. Akad. Berlin, 679, 1906 ; E. C. C. Baly and P. G. Donnan, Journ.
Chem. Soc, 81. 907. 1902.
« P. L. Dulong, Ann. Chim. Phys., (2), 41. 1131, 1829 ; S. R. Cook, Phys. Rev., (1), 23. 212
1906.
' E H. Amagat, Ann. Chim. Phys., (2), 29. 37, 1893 ; H. K. Onnes and H. H. P. Hyndman,
Comm. Phys. Lab. Leiden, 78, 1902.
8 Lord Rayleigh, Zeit. phys. Chem., 52. 705, 1905; C. Bohr, Wied. Ann., 27. 459, 1886:
E. C. C. Baly and W. Ramsay, Phil. Mag., (5), 38. 323, 1894 ; A. Campetti, Atti Accad. Torino, 52
1893 ; A. Batelli, Nuovo Cimento, (5), 1. 81, 1901 ; W. Crookes, Phil. Trans., 172. 410,, 1882 ;
H. Ebert, Verh. deut. phys. Ges., 2. 104, 1900; A. Jaquerod and 0. Scheuer, Mem. Soc Phys. Geneve.
35. 659, 1908 ; Phil. Trans., 204. A, 351, 1905 ; Proc Roy. Soc, 69. 448, 1892 ; D. Berthelot,
Compt. Rend., 126. 954, 1898 ; 145. 180, 1907 ; A. Leduc, ib., 123. 743, 1896 ; A. Leduc and
P. Sacerdote, ib., 125. 297, 1897.
» W. Sutherland, Phil. Mag., (5), 43. 11, 83, 201, 1897 ; R. Threlfall and P. Martin, Chem.
News, 76. 283, 1897 ; Lord Rayleigh, Phil. Trans., 196. 205, 1901 ; 198, 467, 1092 ; M. Thiesen,
Ann. Physik, (4), 6. 280, 1901.
i» P. von Jolly, Pogg. Ann. Jubelband, 82, 1874.
^^ p. Giinther, Ueber die Warmeleitung von Sauerstoff, Stickstoff, und Wasserstojf, Halle, 1906 ;
S. Weber, Ann. Physik, (4), 54. 325, 1917 ; A. Winkelmann, Pogg. Ann., 156. 497, 1875.
12 H. Alt, Ann. Physik, (4), 13. 1010, 1904; H. V. Regnault, Mem. Acad., 26. 1, 1862;
L. Holbom and L. Austin, Abh. phys. tech. Reichsanst, 4. 131, 1905 ; W. Nernst and H. von
Wartenberg, Zeit. phys. Chem,., 56. 543, 1906; P. A. Miiller, Wied. Ann., 18. 94, 1883;
O. Lummer and E. Pringsheim, ib., 64. 555, 1898; G. N. Lewis and G. E. Gibson, Journ. Amer,
Chem. Soc, 39. 2554, 1917; G. N. Lewis and M. Randall, ib., 34. 1128, 1912 ; M. Pier, Zeit.
Elektrochem., 15. 536, 1909; A. Eucken, Ber. deut. phys. Ges., 18. 4, 1916 ; W. Nernst and P. A.
Lindemann, Zeit. Elekirochem., 17. 817, 1911.
13 L. Cailletet, Compt. Rend., 85. 1213, 1877 ; R. Pictet, ib., 85. 1214, 1877 ; S. von Wroblewsky,
ib., 97. 309, 1883; K. Olszewsky, ib., 100. 351, 1885; J. Dewar, Chem. News, 51. 27, 1885;
J. Natterer, Pogg. Ann., 62. 132, 1844.
1* M. W. Travers, G. Senter, and A. Jaquerod, Chem. News, 86. 61, 1902 ; T. Estreicher,
Zeit. phys. Chem., 49. 597, 1904 ; Phil. Mag., (5), 40. 458, 1895 ; H. Alt, Ann. Physik, (4), 13.
1010, 1904 ; Phys. Zeit., 6. 346, 1905 ; A. Eucken, Ber. deut. phys. Ges., 18. 4, 1916; L. L. Grun-
mach, Sitzber. Akad. Berlin, 679, 1906 ; K. Scheel, Zeit. angeiv. Chem., 32. 347, 1919.
" A. Eucken, Ber. deut. phys. Ges., 18. 4, 1916 ; W. Wahl, Proc Roy. Soc, 88. A, 61, 1913 ;
J. Dewar, ib., 85. A, 697, 1911 ; P. Juliusberger, Ann. Physik, (4), 3. 618, 1900.
1' I. Langmuir, Journ. Amer. Chem. Soc, 37. 1162, 1915.
1' M. Croullebois, Ann. Chim. Phys., (4), 20. 136, 1870 ; J. P. Homfray, Journ. Chem. Soc,
87. 1443, 1905 ; W. Ramsay and M. W. Travers, Proc Roy. Soc, 62. 225, 1897 ; G. D. Liveing
and J. Dewar, Phil. Mag., (5), 34. 205, 1892 ; J. H. Gladstone, Chem. News, 67. 94, 1893 ; Phil.
Trans., 159. 13, 1869 ; Proc Roy. Soc, 18. 49, 1869 ; C. and M. Cuthbertson, ib., 83. A, 151,
1909 ; L. Lorenz, Wied. Ann., 11. 70, 1880 ; E. Mascart, Ann. J^cok Norm. Sup., 1, 1877 ;
H. C. Rentschler, Astrophys. Jcmrn., 28. 348, 1908 ; W. Sellmayer, Pogg. Ann., 143. 272, 1871 ;
145. 399, 1872 ; 147. 386, 1872 ; P. Drude, Ann. Physik, (4), 14. 677, 1904 ; J. W. Bruhl, Zeit.
phys. Chem., 7. 2, 140, 1891 ; L. H. Siertsema, Proc Acad. Amsterdam-, 7. 294, 1899 ; 8. 5, 1900.
" K. Angstrom, Pogg. Ann., 94. 141, 1855 ; J. Plucker, ib., 107. 518, 1859 ; A. Wullner, ib., 135.
515, 1868 ; 137. 350, 1869 ; 145. 636, 1872 ; 146. 481, 1872 ; 147. 321, 1872 ; Weid. Ann., 8.
253, 1879 ; A. Schuster, ib., 7. 670, 1879 ; Phil. Trans., 170. 37, 1879 ; J. Plucker and W. Hittorf,
ib., 155. 23, 1865 ; W. Huggins, ib., 154. 139, 1864 ; ib., 170. 37, 1879 ; Proc Roy. Soc, 27.
383, 1879 ; 0. Runge and P. Paschen, Wied. Ann., 61. 641, 1897 ; C. Smyth, Phil. Mag., (5),
13. 330, 1882 ; G. Salet, Ann. Chim. Phys., (4), 28. 35, 1873.
19 A. PaalzofiF, Wied. Ann., 7. 130, 1879.
20 B. Reisman, Zeit. wiss. Phot., 13. 269, 1914 ; J. Stark, G. von Wendt, and H. Kirschbaum,
Phys. Zeit., 14. 770, 1913 ; J. Stark, ih., 14. 102, 779, 1913 ; H. Wilsar, ib., 14. 308, 1913 ; Ann
OXYGEN 373
Physik, (4), 39. 1351, 1912 ; C. Runge and F. Paschen, Wied. Ann., 61. 6*1, 1897 ; Astrophys.
Journ., 8. 70, 1898.
21 J. Janssen, CompL Rend., 101. 11, 649, 1885; 102. 1352, 1886; 106. 1118, 1888; 107.
672, 1888 ; N. Egoroff, ib., 101. 1143, 1885; G. D. Liveing and J. Dewar, Phil. Mag., (5), 26
286, 1888 ; J. Dewar, Chem, News, 67. 210, 1893 ; 0. C. Lester, Amer. Journ. Science, (4), 18.
147, 1904 ; K. Olszewsky, Wied. Ann., 33. 570, 1888 ; Monatsh., 8. 73, 1887 ; W. Burmeister,
Verh. deut. phys. Ges., 15. 589, 1913; J. TyndaU, Proc. Roy. Soc, 35. 129, 1883.
22 J. S. Townsend, Proc. Cambridge Phil, Soc, 9. 345, 1897 ; Phil. Mag., (5), 45. 125, 1898.
23 J. S. Townsend, Phil. Trans., 193. A, 129, 1900 ; J. Zeleny, ib., 195. A. 193, 1900 ; A. P.
Chattock, Phil. Mag., (5), 48. 401, 1899 ; J. Franck, Verh. deut. phys. Oes., 12. 613, 1910 ;
E. Salleo, Radium, 7, 362, 1910 ; A. Erickson, Phil. Mag., (6), 18. 32, 1909 ; P. Philips, Proc.
Roy. Soc, 83. A, 246, 1910.
24 A. L. Hughes and A. A. Dixon, Phys. Rev., (2), 10. 495, 1917 ; F. M. Bishop, ib., (2), 10.
244, 1917 ; K. T. Compton, ib., (2), 8. 412, 1916 ; J. Franck and G. Hertz, Verh. deut. phys. Ges.,
15. 34, 1913.
25 J. A. Flemmg and J. Dewar, Proc Roy. Soc, 60. 358, 1896 ; 61. 299, 316, 358, 1897 ;
F. Hasenohrl, Versl. Akad. Amsterdam, 137, 1900 ; M. Faraday, Phil. Mag., (3), 401, 1847 ;
J. Dewar, Proc Roy. Soc, 50. 247, 1892 ; A. Occhialini, Nuovo Cimento, (6), 7. 108, 1914.
2« R. Hennig, Wied. Ann,, 50. 485, 1893 ; G. Quincke, ib., 24. 347, 1885 ; 34. 401, 1888 ;
S. Hinrichsen, ib., 34. 180, 1888; J. A. Fleming and J. Dewar, Proc. Roy. Soc, 60. 283, 1896; 63.
311, 1898; E. Bauer, P. Weiss, and A. Piccard, Compt. Rend., 167. 484, 1918; H. K. Onnes
and A. Perrier, Comm. Phys. Lab. Leiden, 116, 1900.
27 R. W. Bunsen, Liebig'a Ann., 93. 21, 1855 ; L. W. Winkler, Ber., 22. 1764, 1889 ;
24. 3602, 1891 ; Zeit. phys. Chem., 9. 171, 1892 ; C. Bohr and J. Bock, Wied. Ann., 44. 318, 1891 ;
J. J. Fox, Trans. Faraday Soc, 5. 68, 1909 ; K. Dost and H. Grosse-Bohle, Mitt. Prilf. Wasser.
Abwasser, 168, 1906 ; O. Pettersson and K. Sonden, Ber., 22. 1439, 1889.
28 C. A. Seyler, Chem. News, 67, 87. 1893 ; A. H. GiU, Journ. anal. Chem., 6. 606, 1893 ;
W. E. Adenby and H. G. Becker, Proc Roy. Soc Dublin, 15. 385, 609, 1919.
29 G. Geffcken, Zeit. phys. Chem., 49. 269, 1904 ; A. Christoff, ib., 55. 622, 1906 ; J. S.
Maclaurin, Journ. Chem. Soc, 63. 737, 1893; C. G. MacArthurj " Jowr*. Phys. Chem., 20. 495,
1916 ; F. Clowes and J. W. H. Biggs, Journ. Soc Chem. Ind., 23. 358, 1904 ; E. E. 0. Lubarsch,
Wied. Ann., 37. 525, 1889; W. Timofejeff, Zeit. phys. Chem., 6. 151, 1890; S. Griewasz and
A. Walfisz, ib., 1. 70, 1887 ; M. G. Levi, Gazz. Chim. Ital, 31. 513, 1891 ; L. Carius, Liebig's Ann.,
94. 134, 1855 ; C. Bohr, Skand. Archiv. Physiol, 17. 104, 1905.
3» G. Neumann, Monatsh., 13. 40, 1892; T. Wilm, Bull. Soc Chim., (2), 38. 611, 1882;
T. Graham, Phil. Mag., (4), 32. 503, 1866.
31 L. Mond, W. Ramsay, and J. Shields, Proc Roy. Soc, 62. 290, 1897 ; C. Engler and
L. Wohler, Zeit. anorg. Chem., 29. 1, 1902.
32 E. Goldstein, Ber., 37. 4147, 1904; A. Magnus, Phys. Zeit., 6. 12, 1905; R. Lucas, Zeit.
Elektrochem., 11. 182, 1905.
33 H. St. C. Deville, Compt. Rend., 70. 756, 1870 ; A. Levol, ib., 35. 63, 1852 ; J. B. A. Dumas,
Ann. Chim.. Phys., (5), 14. 289, 1878 ; F. G. Donnan and T. W. A, Shaw, Journ. Soc Chem. Ind.,
29. 987, 1910.
34 S. Kern, Chem. News, 36. 20, 1877.
35 H. V. Regnault, Ann. Chim. Phys., (3), 14. 236, 1845 ; F. Leblanc, ib., (3), 16. 480, 1846 ;
E. H. Amagat, 16., (2), 29. 37, 1893.
36 T. de Saussure, Gilbert's Ann., 47. 113, 1814; J. Hunter, Phil Mag., (4), 25- 364, 1863 ;
J. Dewar, Compt. Rend., 139. 261, 421, 1904 ; L. Jouhn, ib., 90. 741, 1880 ; E. Goldstein, Ber.,
37. 4147, 1904 ; G. Craig, Chem. News, 90. 109, 1904 ; J. L. Baerwald, Ueber die Absorption von
Gasen durch Hohkohle bei tiefen Temperaturen, Freiberg i. B., 1906.
37 T. Graham, Compt. Rend., 63. 471, 1866 ; L. Troost, ib., 98. 1427, 1884 ; A. BartoH, Gazz.
Chim. Ital, 14. 544, 1884 ; M. Berthelot, Compt. Rend., 140. 1286, 1905.
38 E. Erdmann and F. Bedford, Ber., 37. 1184, 2545, 1904; A. Stock, ib., 37. 1432, 1904;
A. Stock and C. Nielsen, ib., 39. 3393, 1906; J. K. H. Inglis, Phil Mag., (6), 11. 640, 1906 ;
J. K. H. Tnglis and J. E. Coates, Journ. Chem. Soc, 89. 886, 1906.
§ 9. The Chemical Properties of Oxygen
Oxygen, so to speak, is the central pivot round which the whole of chemistry revolves. —
J. J. Berzelius.
Oxygen is entirely iinmatched among the rest of the elements both as regards the
number and the varied character of its compounds. — W. A. Tflden.
The great chemical activity of oxygen is well typified by the quaint remarks
which J. Priestley made about the gas. When a glowing splint of wood is plunged
into oxygen it bursts into flame ; the carbon of the wood is oxidized to carbon
dioxide (CO2). The inflammation of a glowing splint is often used as a test for
374 INORGANIC AND THEORETICAL CHEMISTRY
oxygen. A mixture of nitrogen and oxygen containing less than 28'29 per cent, of
the latter does not re-ignite a glowing splint, and if the mixture contains less than
about 16 per cent, of oxygen the splint will be extinguished. Oxygen alone has no
visible action on clear lime-water ; but after a spUnt has burnt in the gas, the clear
lime-water becomes turbid. Oxygen combines directly with most other elements,
particularly at elevated temperatures, forming oxides.
The direct combination of oxygen with some of the elements can be illustrated by
placing small dry pieces in deflagrating spoons, heating them until combustion begins,
and then plunging each into a jar of oxygen. The glowing piece of charcoal bums very
brightly and forms a gaseous oxide- — carbon dioxide, CO2. Sulphur burns with a lavender*
blue flame, forming gaseous sulphur dioxide- — SO 2 — which has the peculiar odour character-
istic of burning sulphur. The reaction is symbolized : S+02 = S02 ; sulphur dioxide is
soluble in water forming sulphurous acid — H2SO3 — which reddens blue litmus solution- —
H2O + SO2 =112803. Phosphorus burns in oxygen vigorously and brilliantly, forming
a white cloud of phosphorus pentoxide — P2O5. The reaction is represented : 4P+5O2
= 2P205. The phosphorus pentoxide dissolves in water, forming phosphoric acid- —
H3PO4. The reaction is written : P206 + 3H20 = 2H3P04. The phosphoric acid reddens
blue litmus. Metallic sodium treated in a similar way (spoon dry) burns with a bright
yellow flame and gives a white oxide which dissolves in water, forming a solution of caustic
soda. The solution turns red litmus blue. Calcium behaves similarly, but it burns with
an orange-red flame. A piece of burning m,agne-'^ium, ribbon plunged in oxygen bums with
an exceptionally brilliant flame. The white solid obtained is slightly soluble in water,
and the solution turns red litmus blue. To show the combustion of iron in oxygen gas, tie
a tuft of steel wool to the end of a stout iron wire by means of a piece of steel wire. Heat
the end of the wool in a Bunsen's flame, until incipient combustion begins, and quickly
plunge it into a jar of oxygen on the bottom of which a layer of water, sand, or asbestos
paper has been placed. The wool bums with dazzling scintillations, the product of the
reaction — iron oxide^ — falls to the bottom of the jar in fused globules. When cold, the oxide
of iron resembles a blacksmith's scale. It is called black or magnetic oxide of iron —
Fe304. The reaction is usually written: 3Fe+202 = Fe304. Experiments showing the
combustion of iron in oxygen date from G. C. Lichtenberg (1782).^ The oxide of iron so
formed is insoluble in water and has no effect on red or blue litmus.
Iodine, bromine, chlorine, fluorine, gold, platinum, and argon and its companions
do not combine directly with oxygen. The reaction between metals and oxygen
does not as a rule take place at ordinary temperatures, and heat is required. If
the oxide is unstable at the temperature necessary for reaction, it will not be formed
directly even though much heat be evolved in the formation of the oxide. Mercury
at a high temperature does not appear to react with oxygen since the oxide, if
formed, is immediately decomposed. The temperatures of formation and de-
composition of the oxide are not far apart ; with the oxides of silver and palladium,
the temperatures of formation and decomposition are probably much nearer even
than with mercury. Similar remarks apply to iodine and platinum oxides. The
elements, nitrogen, fluorine, chlorine, and bromine, absorb energy when they unite
with oxygen, and oxides can be formed indirectly or in some cases directly if the
temperature is very high — e.g. nitrogen oxide.
Oxygen combines indirexitly with all the elements excepting the argon group,
fluorine, and possibly bromine. "Oxygen," said C. L. Berthollet (1803), "seems
to take the lead of all substances in the extent and energy of its affinities," and,
with perhaps the exception of fluorine, these words are true to-day. The energy
which is degraded as heat when the different elements combine with oxygen, is a
distinctive characteristic. The amount of heat liberated in combining with a
gram-atom of the element is a rough indication of the avidity of the element for
oxygen. Table I shows the heat developed during the formation of a gram-
molecule of an Qxide ; 2 and also per gram-atom of the element united with oxygen.
If no remark is made as to the state of aggregation — solid, liquid, gas — a solid is to
be understood. If the metals be arranged in the order of their avidity or readiness
to combine with oxygen, cajsium, potassium, and sodium will be found at one end
of the series, while platinum and the argon family will be found at the other end.
If the heats of formation of the oxides of the elements be plotted against the atomic
weights, a periodic curve is obtained, corresponding approximately with the periodic
OXYGEN
375
curve obtained when many of the other properties of the elements are plotted
against the atomic weights.
The preparation of the oxides. — The methods for preparing individual oxides
are described when dealing with the respective elements. The following are common
enough to merit the designation general metJiods : (1) By calcining the metal while
freely exposed to air, e.g. tin gives stannic oxide, Sn-|-02->Sn02 ; (2) By calcining
the nitrate strongly, and subsequently washing the residue to remove the unde-
composed nitrate, e.g. with copper nitrate, Cu(N03)2, the action is represented :
2Cu(N03)2->2CuO+4N02+02 ; with chromium nitrate: 4Cr(N03)3->2Cr203
+I2NO2+3O2 ; (3) By calcining the carbonate, e.g. with calcium carbonate :
CaC03->CaO+C02. With barium and strontium carbonates it is better to mix the
carbonate with powdered carbon (lampblack) before calcination. (4) By pouring a
solution of alkali hydroxide or aqueous ammonia into a solution of the salt, washing
the precipitated hydroxide, and afterwards calcining it to drive off the water,
e.g. with ferric chloride and sodium hydroxide : FeCl3+3NaOH=Fe(OH)3
+3NaCl ; the subsequent calcination of the ferric hydroxide, Fe(0H)3, furnishes the
required oxide : 2Fe(OH)3->Fe203+3H20.
Table I.- — ^Heat Evolved or Absorbed in the Formation of the Oxides.
Cals.
1
Cals.
Cals.
Cals.
Cals.
Cals.
Oxide.
per gram
mole-
cule.
per gram
atom.
Oxide.
per gram
mole-
cule.
per gram
atom,
136-3
Oxide.
per gram
mole-
cule.
per gram
atom.
H,0 (gas) .
58-1
29-0
B2O3
272-6
Bi203
139-2
69-6
H2O (liquid)
69-0
34-5
AI2O3 .
392-6
196-3
SO2 (gas) .
69-3
69-3
H2O (solid)
70-4
35-2
TI2O
42-8
21-4
SO3 (liquid)
91-9
91-9
LigO
140-0
70-0
TI2O3 .
87-6
43-8
SeOa
571
57-1
NagO
100-9
50-4
CO (gas) .
29-2
29-2
WO3
65-5
32-7
K2O
98-2
49-1
C02(gas).
97-2
97-2
CI2O (gas)
17-9
-8-9
RbgO
94-9
47-4
SiOa
180-0
180-0
I2O6
45-3
22-6
CsgO
99-98
50-0
TiO.,
97-8
97-8
PbO
50-3
50-3
Ago
7-0
3-5
ZrOg
177-5
177-5
PbO
63-4
63-4
CU2O
43-8
21-9
N20(gas).
—17-5
-8-8
SnO
70-7
70-7
CuO
37-7
37-7
NO (gas) .
21-6
-21-6
Sn02
141-3
141-3
AU2O3
11-5
5-8
N02(gas).
20
-2-0
FeO
65-7
65-7
CaO .
131-5
131-5
P2O5
365-3
182-6
FeaOg
195-6
97-8
SrO
131-2
131-2
AS2O3 .
156-4
78-2
Fe304 .
270-8
90-3
BaO
133-4
133-4
AS2O5 .
219-4
109-7
CoO
64-1
64-1
BaOg
145-5
145-5
SbaOa .
166-9
83-4
NiO
61-5
61-5
MgO
143-4
143-4
Sb205 .
231-2
115-6
MnO
90-9
90-9
ZnO
84-8
84-8
HgO
21-5
21-5
MngO^ .
328-0
109-3
CdO
66-3
66-3
PdO
21-0
21-0
Mn02
.125-3
125-3
Hg^O
22-2
11-1
PtO
17-0
17-0
In reviewing the oxides of all the elements, it will be apparent that the pro-
portions of combined oxygen are not always the same. The elements, indeed, can
be arranged roughly into natural groups determined by the composition and pro-
perties of what D. I. Mendeleeff regarded as the different typical oxides they form.
This has been done in Table II.
Many of the elements form a number of different oxides, and in that case the
same element might fall into two or more different groups. This is well illustrated
by the family : iron, manganese, cobalt, and nickel. The RO oxides of these elements
are readily oxidized to sesquioxides of the type R2O3, and these same elements
also form still higher oxides of the type, RO2, and, in the case of manganese, there
is evidence of a yet higher oxide, Mn207. Under a pressure of about 12 atmospheres,
at 480°, oxygen 3 oxidizes the oxides of lithium, sodium, potassium, and barium to
peroxides of the type R2O2 ; and a small proportion of a peroxide is formed with
cobalt and nickel oxides ; lead oxide gives red lead PbsOi ; antimony oxide gives
376
INORGANIC AND THEORETICAL CHEMISTRY
the tetroxide Sb204 ; chromium sesquioxide gives chromium chromate 01304 or
CrCr04. Beryllium, calcium, strontium, zinc, cadmium, aluminium, boron, thallium,
silicon, zirconium, tin, bismuth, molybdenum, tungsten, uranium, and ferric oxides
do not change.
Table II.— Oxides of the Elements.
Oxide.
Corresponding group of elements.
Character of oxide.
MjO
(H), Li, Na, K, Cu, Rb, Ag, Cs, Au
Basic
MO
Be, Mg, Ca, Zn, Sr, Cd, Ba, Hg
Basic
M,03
B, Al, Sc, Ga, Y, In, La, Yb, Tl
The first oxide B2O3 is weakly acidic ;
the others are basic
MO,
C, Si, Ti, Ge, Zr, Sn, Ce, Pb
The first two are acidic ; the last one
is basic ; and the others are both
basic and acidic, becoming more
basic with increasing atomic weight
M,0,
N, P. V, Nb, Di, Er, Ta, Bi
These oxides are acidic ; BigOg is also
basic
MO3
0, S, Cr, Se, Mo, Te, W, U
Acidic and become less and less acidic
as the atomic weight increases ;
VO3 is also feebly basic
M,0,
F, CI, Mn, Br, I
Acidic. The highest oxide is repre-
sented by M2O7. No definite oxide
of Br or F is known
MO4
Fe, Ni, Co ; Ru, Rh, Pd ; Os, Ir, Pt
The first triad forms feebly basic
/
sesquioxides : RUO4 and OSO4 are
the only representatives of the
highest oxide. These are feebly
acidic. The lower oxides are feebly
basic
Oxygen is closely related to the elements of the sulphur, selenium, and tellurium
family. The changes in the physical characters of the oxides show regular grada-
tions with increasing atomic weight in harmony with the periodic classification.
Regidarities have been traced in the specific gravity, atomic volume, volatility,
stability, reactivity with water, heats of formations, etc. According to G. H.
Bailey * in the even series of Mendeleefi's table :
II.
BeO
CaO
SrO
III.
BaOa
ScaOa
YtaOg
IV.
CO2
TiOg
ZrO,
V.
N2O5
Nb,0.
VI,
CrOs
M0O3
WO3
VII.
VIII.
Fe04
RUO4
OSO4
The oxides of the first four groups are so stable that they undergo no decomposition
at temperatures below 1750° ; nitrogen pentoxide decomposes below 50°, and,
further on, the oxides are more stable the higher the atomic weight — uranium
oxide, UO3, appears to be an exception. In the horizontal series, the stabihty of
the oxides decreases from left to right as the atomic weight increases. The be-
haviour of the RO4 oxides of the eighth group is also in keeping with these generalities.
Osmium octoxide is more stable than ruthenium octoxide, and there is a doubt
about the existence of the corresponding iron compound.
Many substances are oxidized at ordinary temperatures — e.g. nitric oxide, a
colourless gas, oxidizes to reddish-brown nitrogen, peroxide, NO2, in air at ordinary
temperatures ; the alkali metals ; ferrous and manganous hydroxides ; etc. In
some cases the oxidation is so vigorous that the heat developed inflames the mass.
This is the case, for instance, with hydrogen phosphide, P2H4 ; silicon hydride,
Si2H6 ; zinc ethyl, Zn(C2H5)2 ; etc. Some of the metals in a very fine state of
subdivision are oxidized in air — e.g. pyrophoric iron, nickel, cobalt, etc. In some
cases the oxidation is specially stimulated by exposure to sunlight— e.^. lead sulphide,
PbS, becomes lead sulphate, PbS04 5 carbon chloride, C2CI6, forms a mixture of
OXYGEN 377
carbonyl chloride, COCI2, and trichloroacetyl chloride, CCl3.CO.Cl ; phosphorus
trichloride, PCI3, forms phosphoryl chloride, POCI3 ; etc. The presence of the
noble metals, and of cobalt, nickel, etc., in a fine state of subdivision, may
enormously accelerate the speed of oxidation.
The influence o£ water in chemical reactions. — Water plays an important role
in many reactions. If water be dropped on to a mixture of iodine with one-sixth of
its weight of aluminium powder, the reaction proceeds so rapidly as to inflame the
mass. There are, however, cases in which the minute trace of water vapour which
is present in an imperfectly dried gas, controls the reactivity of the gas. For
instance, H. B. Baker (1886) ^ showed that dry sulphur, dry phosphorus, and dry
carbon burn with great difficulty or not at all in dry oxygen. Similarly, H. B.
Dixon (1880) showed that carbon monoxide reacts with oxygen with greater
difficulty, if it be thoroughly dried — e.g. if moist, a mixture of the two gases readily
detonates, but not if dried. Numerous other reactions have been notified which are
arrested if the reacting materials be dried. In fact, many perfectly dried substances
often appear to be chemically inert, whereas they react vigorously if a trace
of moisture be present.
The fact is quite old. Near the end of the eighteenth century, for instance,
T. Bergmann 6 noticed that the " regulus of manganese" remains bright in dry air,
but not in moist air ; the illustrious C. W. Scheele also noted in 1786 that pyro-
phorus will not oxidize in air dried by quicklime, and he inferred that " the water
usually present in the atmosphere is the chief cause of the burning of pyrophorus."
Mrs. Fulhame, too, in her remarkable brochure, An Essay on Combustion with a
view to a new art of dying and fainting (London, 1794), showed " beyond the power
of contradiction " that water is necessary for the reduction of the metallic oxides,
and for the oxidation of the metals. She found, for example, that gold chloride
cannot be reduced by hydrogen gas if moisture be excluded. The efiect of moisture
is not to promote the reduction by breaking up the salt into minute particles, nor
by condensing the gas and so bringing the hydrogen into closer contact with the
metallic oxide ; for, if either of these views were correct, ethereal and alcoholic
solutions of the metallic salt should prove as effective as water. This is not the
case. Neither ether nor alcohol promotes the reduction if water be absent. Mrs.
Fulhame believed that the reaction — oxidation or reduction — took place in two
stages. In the first place, carbon monoxide decomposed the water, forming carbon
dioxide and liberating hydrogen; thus — C0+H20=C02+H2 (nascent); finally,
the nascent hydrogen united directly with the free oxygen, reforming water —
2H2 (nascent) -j-02=2H20. Consequently, the oxygen which unites with the
carbon monoxide to form carbon dioxide is not obtained directly from the oxygen
gas mixed with the carbon monoxide, but from the water. H. B. Dixon ^ developed
quite an analogous theory as a result of an important investigation on the oxidation
of carbon monoxide, and he submitted that carbon monoxide is oxidized by steam
with the liberation of hydrogen,^ and that the hydrogen then unites with oxygen to
reform steam. These results make it probable that steam does really undergo
" a cycle of chemical reactions whereby it gives up oxygen to carbon monoxide and
returns to its original state." H. B. Dixon also proved that other gases like hydrogen
sulphide, ethylene, formic acid, ammonia, pentane, and hydrogen chloride will
determine the explosion of carbon monoxide and oxygen ; while sulphur dioxide,
carbon disulphide, carbon dioxide, cyanogen, and carbon tetrachloride are quite
ineffective. Hence, he inferred that not only steam, hut all substances which will
form steam under the conditions of the experiment, are capable of determining the
explosion.
M. Traube (1882) observed that traces of hydrogen peroxide are generally
formed during the oxidation of carbon monoxide, and he suggested that hydrogen
peroxide, H2O2, is an intermediate product in the oxidation of hydrogen or carbon
monoxide, ; and that the water acts catalytically by acting as a link in the cyclic
changes: CO+H20+02=C02+H202 ; and CO+H202=C02+H20. Dixon,
378 INOKGANIC AND THEORETICAL CHEMISTRY
however, showed that the hydrogen peroxide is most probably a by-product in the
oxidation of moist carbon monoxide. In the oxidation of hydrogen, M. Traube
supposed that the water acted catalytically :
H , HOH O
H'^HOH'^0
HOH, HO HO ,H_HOH
HOH'^HO' *^^ h6^H~H0H
In reality we have not got much further than Mrs. Fulhame (1794) in working out
the mechanism of this reaction. This gifted woman said :
Water is essential both to the reduction and oxygenation of bodies, and is always
decomposed in these operations. ... In every instance of combustion water is decomposed,
•and one body oxygenated by the oxygen of the water, while another is restored to its
combustible state by the hydrogen of the same fluid.
It must be added that the water as a catalytic agent does not necessarily accelerate
the speed of all reactions. W. S. Millar, for example, found that the speed of de-
composition of diazoacetic ester by picric acid in alcohol solutions is retarded in a
marked degree if a small amount of water be present, and this the more with isobutyl
alcohol than with ethyl or methyl alcohol as solvent.
The physiological action of oxygen. — In 1667, Robert Hooke ^ clearly demon-
strated before the Royal Society that a continual supply of air is necesssary for the
maintenance of life. In an experiment on a dog with its ribs and diaphragm
removed, and described as "an experiment made by Mr. Hooke of preserving
animals alive by blowing through their lungs with a bellows," he emphasized
It was not the subsiding or movelessness of the lungs that was the immediate cause of
death, or the stopping of the circulation of the blood through the limgs, but the v)ant of a
sufficient supply of air.
In 1674, J. Mayow showed that one constituent of air is alone active, and further
that this constituent is the same as that on which ordinary combustion depends.
He called the active constituent spiritus nitro-cereus, which was later identified
with oxygen. From his experiments with mice, etc., he concluded :
It is manifest that air is deprived of its elastic force— decreased in volume- — by the
breathing of animals very much in the same way as by the burning of a flame ; and,
indeed, we must believe that animals and fire draw particles of the same kind from the air.
The importance of oxygen in the maintenance of animal life was emphasized by M. de
Condorcet's term Vair vital — ^the life maintaining constituent of air.^o It is the only
gas capable of supporting respiration. There are, however, a few micro-organisms —
e.g. the mould mucor racemosus, and the butyric acid ferment — which are killed by
oxygen, and they are able to live and multiply without air. L. Pasteur called them
anaerobic organisms. E. Weinland found that intestinal worms can normally
exist in the absence of oxygen ; and A. Piitter found that the leech can live two
days without oxygen.
The oxygen is carried by the blood to the various tissues in the body, and the
waste products are carried away by the same liquid. The circulating blood is
oxidized in the lungs of land animals, and in the gills of water animals. In insects,
the blood is oxidized in a system of ramifying tubes called tracheoe in which the air
is periodically changed by muscular movements and diffusion. Fish are dependent
upon the air dissolved in water for the oxygen they need for respiration. According
to E. A. Birge and C. Juday, the dissolved oxygen is deficient in the lower layers
of water in 129 inland lakes of Wisconsin. This is attributed to the thermal stratifi-
cation of the water owing to the greater specific gravity of the colder water hindering
vertical circulation, and the depletion of the oxygen in the colder layers by the
respiration of animals and plants, by the direct oxidation of dead organic matter,
and by the decomposition due to the action of bacteria. A deficiency has also been
noted in the lower layers of certain tidal waters which is similarly caused by
OXYGEN 379
stratification due to the greater specific gravity of the under-run of sea water
hindering vertical circulation.
Warm-blooded animals die very rapidly in an atmosphere containing no oxygen.
A man at rest becomes suddenly unconscious after about ,ten breaths of such an
atmosphere ; and very small animals, such as a mouse or a sparrow, in which the
breathing is far more rapid than in a man, are killed within a few seconds, death
being much faster than by drowning. When pure oxygen is breathed no noticeable
effect is produced for many hours. As A. L. Lavoisier first showed, and many sub-
sequent observers have also found, breathing pure oxygen causes no increase in the
oxidation processes within the body. Paul Bert found, however, that oxygen at
a pressure of over three atmospheres has a rapid poisonous effect on warm-blooded
animals, accompanied by diminution of oxidation processes. This poisonous action
is also produced in living organisms of all kinds. Pure air of which the pressure is
raised so high as to give the same partial pressure of oxygen has the same effect.
More recently, Lorrain Smith showed that exposure for two or three days to pure
or nearly pure oxygen produces inflammation of the lungs ; and the higher the
partial pressure the sooner the inflammation appears. It was formerly supposed
that poisonous organic matter is exhaled in the breath along with CO2. AH recent
investigation has shown that this view is without foundation. " It has been often
asserted," adds L. Hill, " that there is some organic poison exhaled with the breath.
I have carefully sifted the evidence on which this assertion is based, and find that
there is none worthy of evidence."
Air normally contains nearly 21 per cent, of oxygen by volume. When the
oxygen is reduced to about 17 '5 per cent, the flame of a candle or oil-lamp is ex-
tinguished. A man or animal is, however, not appreciably effected by so small a
diminution in the oxygen percentage. On the other hand, if the atmospheric
pressure, and consequently the partial pressure of oxygen, be diminished to a third
the man or animal is greatly affected and soon dies, while the flame continues to
burn almost as well as before. Roughly speaking, the flame responds to the 'per-
centage of oxygen in the air, while the animal responds to the 'partial 'pressure of
oxygen. H. C. Dallwig, A. C. Kolls, and A. S. Lowenhart (1915) found that the
flame of a candle is just extinguished when the partial pressure of oxygen is 116'4
mm. of mercury when this lowering is produced by adding nitrogen to air, whereas
the flame is first extinguished at 19 "8 mm. if the lowering is effected by reducing the
total pressure of the atmosphere. Pure oxygen can be breathed for many hours
without harm, and is used in a pure state in mine-rescue apparatus and in resusci-
tating persons poisoned by carbon monoxide. During the war the continuous
administration of air containing an increased percentage of oxygen was used with
striking success in the treatment of lung-inflammation caused by poison-gas ; and
similarly enriched air is now coming into extensive use in medical cases of other
kinds.
Uses of oxygen. — Mixtures of liquid oxygen and petroleum are violently explo-
sive. It is said that a lighted candle falling into a bucket of liquid oxygen in 1903
'* sent G. Claude to the hospital in a very pitiable condition." Liquid air, or rather
liquid air rich in oxygen, furnishes an explosive — called oxylignite — when mixed
with charcoal, or cotton wool. 3-cm. cartridges charged with one part of carbon,
one part of petroleum, and eight parts of liquid oxygen were tried experimentally
in cutting the Simplon tunnel. The cartridges were exploded by an electric fuse,
or a mercury fulminate cap. The chief objection is that the cartridges must be
used within a few minutes after being charged, or the oxygen will evaporate. This
objection might be an advantage under some circumstances, since a mis-fired shot
becomes harmless in a very short time. The cartridges must also be prepared
immediately before use, so that there are no dangers during transport.
The temperature of the hydrogen flame burning in air at 0^ is, according to
P. Mahler,ii 1960° ; of the carbon monoxide flame, 2100° ; and of the acetylene flame,
2350°. Oxygen is used in conjunction with hydrogen for the oxy-hydrogen blowpipey
380 INOKGANIC AND THEORETICAL CHEMISTRY
and with acetylene for the oxy-acetylene blowpipe used in welding, metal-cutting, etc.
Thick steel plates can be cut by directing a stream of oxygen on the heated metal.
Metal cutting or welding by the oxy-acetylene or oxy-hydrogen blowpipe has proved
to be a remarkable labour-saving process. It has been estimated that 70 per cent.
of oxygen consumed in the United Kingdom is used in cutting metals in shipbuilding
and repairing yards, steel works, and engineering shops. Oxygen is used in the
ventilation of submarines, etc., and for medical purposes ; it is employed in the
oxidation and thickening of oils to be used in making varnishes and linoleum. It
is sometimes used to hasten the maturing of spirits ; and the oxidation of alcohol
by mycoderma aceti in vinegar manufacture. Oxygen has been recommended in the
bleaching of paper-pulp, etc., where a fine stream of oxygen is said to efEect a saving
in the consumption of bleaching powder.12 It has also been proposed to use carbu-
retted oxygen — a safe mixture of oil and oxygen — as a motive gas for engines, for
illuminating purposes,i3 in organic " combustion analyses," etc.
The determination of oxygen in a gas. — The property of rekindling a glowing
spUnt is possessed by only one other gas — nitrous oxide — and the two gases are
distinguished by a bubble of nitric oxide — oxygen gives red fumes, nitrous oxide
does not. The methods for measuring the amount of oxygen in a gas depend on
its absorption by various liquids and solids. For instance : (1) a solution of cuprous
chloride in hydrochloric acid. The colourless solution becomes greenish-brown
owing to the formation of cupric oxychloride. The exhausted solution is restored
by keeping it in contact with copper shavings away from air. (2) An alkaline
solution of pyrogaUol freely absorbs oxygen forming a dark brown liquid — if the
solution be saturated with oxygen, some carbon monoxide may be formed. This
solution was used by J. von Liebig in 1851. 1* (3) Clean moist copper absorbs
oxygen. The absorption soon ceases owing to the formation of a film of oxide ; this
can be washed off by an ammoniacal solution of ammonium carbonate. The gas
containing oxygen is introduced into a vessel containing copper shavings and the
ammoniacal solution, the liquid is displaced and the copper absorbs the oxygen,
the return of the ammoniacal liquid displaces the gas. (4) Clean sticks of phosphorus
are sometimes employed for absorbing the gas. (5) A solution of chromous chloride
in hydrochloric acid ; and (6) an alkaline solution of ferrous tartrate, also absorb
oxygen.
The atomic weights of hydrogen and oxygen. — The early determinations of the
combining ratios of hydrogen and oxygen by H. Cavendish (1781), A. L. Lavoisier
and M. Meunier (1788), M. Monge (1788), J. Dalton (1803), and W. H. Wollaston
(1814) are of great historical interest ; but the results are not considered accurate
enough to be worthy of consideration in deducing best representative values of
these data. Several methods have been used in evaluating the ratio H : 0. The
accurate determination of this ratio has proved to be of extreme difficulty. The
subject has been well discussed by B. Brauner,!^ J. SebeHen in his Beitrdge zur
Geschichte der Atomgewichte (Braunschweig, 1884), and by F. W. Clarke in his
A Recalculation of the Atomic Weights (Washington, 1910).
I. Gravimetric methods. — In these methods of determining the ratio H : O, the
hydrogen and oxygen may each be weighed separately, and the water also weighed.
More usually two of these three quantities are determined, and the third estimated
by difference. The various methods include: (i) Those in which the hydrogen
and oxygen are weighed, and the water estimated by difference ; 1^ (ii) those in
which the oxygen and water are weighed, and the hydrogen estimated by difference ; i'
(iii) those in which the hydrogen and water are weighed, and the oxygen estimated
by difference ; 1® and (iv) those in which the hydrogen, oxygen, and water are all
weighed — synthese complHeA^ According to F. W. Clarke, the best representative
value is 0=15*8779 if H=l ; and H=l-00769 if 0=16.
II. Volumetric methods. — Volumetric methods include those in which the volumes
of the hydrogen and oxygen are measured and the water estimated by difference ;
and those in which either the volume of the oxygen or of the hydrogen, or both are
OXYGEN 381
measured and the resulting water weighed. Determinations of the ratio of the
combining volumes of hydrogen and oxygen 20 (Cap. Ill), give as the best representa-
tive value of the ratio H : 0=10077 : 16.
. III. Gas densities. — Determinations of the relative densities of hydrogen and
oxygen 21 furnish for hydrogen the atomic weight 1*00777 ; and for oxygen, 15"8767.
IV. Physico-chemical tnethods. — A number of other methods 22 have been
employed. For example, the method of critical constants by A. Leduc and P. Sacer-
dote, Lord Kayleigh, A. Jaquerod and 0. Scheuer, and by D. Berthelot furnished
H : 0=1*00777 : 16 ; the method of limiting densities by P. A. Guye, D. Berthelot,
and A. Leduc furnished H : 0=1*00775 : 16 ; and the method of molecular volume
by A. Leduc furnished H : 0=1*0076 : 16.
V. Indirect determination. — J. Thomsen determined the amount of ammonia
required to saturate a given amount of hydrogen chloride, then, given the atomic
weights of nitrogen and chlorine (0=16), the atomic weight of hydrogen can be
computed.
Dry hydrogen chloride was passed into a weighed flask containing water coloured with
litmus and weighed- — 5*0363 grms. of hydrochloric acid were absorbed ; dry ammonia was
passed in until the liquid was almost neutral and the flask again weighed. The excess of
ammonia or acid was determined by titration with standard acid or alkali — 2*3523 grms.
of ammonia were used to neutralize 5*0363 grms. of hydrogen chloride. Hence, HCl : NH3
= 5*0363 : 2-3523. If the atomic weight of chlorine be 35*457 and of nitrogen 14-044, the
atomic weight of hydrogen is 0-9989. The uncertainty as to the value of the atomic weight
of nitrogen here affects that of hydrogen.
J. S. Stas 23 determined the relation of silver to ammonium chloride and bromide.
Given the atomic weights of nitrogen, chlorine, and bromine (0=16), the atomic
weight of hydrogen follows. The mean of Stas' results with the chloride and
bromide gave H : 0=1 : 15*9229 ; or 1*00598 : 16. J. Dewar and A. Scott (1887)
tried to use the substituted ammonias — e.g. triethylamine, N(C2H5)3 — in place of
ammonia, but the difficulties involved in purifying the triethylamine make the
method undesirable.
References.
1 E. von Lippmann, Chem. Zta., 32. 161, 1908 ; Abhandlungen und Vortrdge, Leipzig, 2. 307 ;
1913.
2 J. Thomsen, ThermochemiscJie Urdersuchungen, Leipzig, 2. 395, 1882.
^ J. Milbauer, Chem. Ztg., 40. 587, 1916.
4 G. H. Bailey, Journ. Chem. Soc, 65. 315, 1894.
6 H. B. Baker, Journ. Chem. Soc, 47. 349, 1886 ; Proc. Chem. Soc, 18. 40, 1902 ; H. B.
Dixon, Proc. Boy. Sac, 37. 56, 1884 ; B. A. Bep., 593, 1880.
* T. Bergmann, Opuscula physica et chimica, Upsala, 2. 206, 1780 ; C. W. Scheele, CrelVs
Ann., 1. 483, 1786.
' H. B. Dixon, Phil. Trans., 175. 630, 1884 ; Journ. Chem. Soc, 49. 95, 1886 ; M. Traube,
Ber., 15. 666, 1882 ; 18. 1891, 1885.
8 W. R. Grove, Phil. Trans., 138. 617, 1847 ; H. L. Buff and A. W. Hofmann, Liebig's Ann.,
113. 129, 1860 ; Journ. Chem. Soc, 12. 273, 1860 ; A. Naumann and C. Pistor, Ber., 18. 2894,
1885 ; L. Maquenne, Bull. Soc Chim., (2), 39. 308, 1883 ; Compt. Bend., 96. 63, 1882 ; A. Gautier,
ib., 142. 1382, 1462, 1906 ; 0. Boudouard, Bull. Soc Chim., (3), 25. 484, 1901 ; 0. Hahn, Zeit.
phys. Chem., 42. 705, 1903 ; 44. 513, 1903 ; 48. 735, 1904 ; W. S. Millar, ib., 85. 129, 1913.
^ R. Hooke, Phil. Trans., 2. 539, 1667 ; J. Mayow, Tractalus quinque medico-physici, Oxonii,
1674.
10 A. L. Lavoisier, Mem. Acad., 355, 1780 ; 185, 1789 ; W. Kiihne, Zeit. Biol, 35. 43, 1897 ;
36. 425, 1898 ; E. Weinland, ib., 42. 55, 1901 ; A. Putter, Zeit. allgem. Physiol, 6. 217, 1907 ;
7. 16, 1907 ; P. Bert, CompL Bend., 74. 620, 1872 ; L. Pasteur, ib., 52. 344, 1861 ; 56. 416, 1863 ;
J^tudes sur la biere, Paris, 1876 ; E. A. Birge and C. Juday, Wisconsin Geol Nat. Hist. Sur., 22.
26, 1911 ; J. W. Sale and W. W. Skinner, Journ. Franklin Inst., 184. 675, 1917 ; H. C. Dallwig,
A. C. Rolls, and A. S. Loevenhart, Journ. Biol Chem., 20. 32, 1915 ; Amer. Journ. Physiol, 36.
356, 1915 ; 39. 77, 1915 ; E. C. Schneider and L. C. Havens, ib., 36. 380, 1915; C. le N. Foster
and J. S. Haldane, Investigation of Mine Air, London, 144, 1905; J. S. Haldane, Beport on the
Health of Cornish Miners, London, 1904; L. HiU, Journ. Physiol, 32. 225, 486, 1905 ; Journ.
Inst. Min. Eng., 43. 285, 1913; E. F. W. Pfliiger, Pfiugers Archiv., 10. 350, 1870; J. Barcroft,
The Bespiratory Function of the Blood, Cambridge, 1914; J. S. Haldane, Lung-Irritant Oas
382 INORGANIC AND THEORETICAL CHEMISTRY
Poisoning and its Sequelce, London, 1919; J. S. Haldane and L. Smith, Journ. Physiol, 32. 231,
1S07 ; 25. 331, 1900 ; I. L. Smith, Journ. Physiol, 24. 19, 1899 ; P. Bert, La pression barometrique,
Paris, 1878. I am indebted to Dr. J. S. Haldane for kindly revising this section.
1^ P. Mahler, Contribution a V etude des combustiblet^, determination industrielle de leur 'puissance
calorifque, Paris, 1893.
12 L. T. Thome, Journ. Soc. Chem. Ind., 8. 83, 1889.
" E. Tathara, Brit. Pat. No., 13763, 16138, 16142, 1889.
** J. von Liebig, Liebig's Ann., 77. 107, 1851 ; G. W. Jones and M. H. Meighan, Journ. Ind.
Eng. Chem., 11. 311, 1919.
•^ B. Brauner, R. Abegg's Handbuch der anorganischen Chemie, Leipzig, 2. i, 4, 1908.
18 Lord Ravleigh, Proc. Roy. Soc, 45. 425, 1889.
1' J. J. Berzelius and P. L. Dulong, Ann. Chim. Phys., (2), 15. 386, 1820 ; J. B. A. Dumas,
ib., (3), 8. 189, 1843 ; 0. L. Erdmann and R. F. Marchand, Journ. prakt. Chem., (1). 26. 461,
1842; T. Clark, Phil Mag., (3), 20. 341, 1842,- T. Hilditch, Chem. News, 49. 370, 1884;
W. Dittmar, Proc. Roy. Soc. Edin., 18. 320, 1891 ; W. Dittmar and J. B. Henderson, Proc. Phil
Soc. Glasgow, 22. 33, 1891 ; A. Leduc, Compt. Rend., 115. 41, 1892 ; Ann. Chim. Phys., (7), 15.
48, 1898.
18 J. Thomsen, Ber., 3. 927, 1870 ; E. H. Keiser, ib., 20. 2323, 1887 ; J. D. van der Plaats,
Ann. Chim. Phy-i., (6), 7. 499, 1886 ; J. P. Cooke and T. W. Richards, Amer. Chem. Journ., 10.
81, 191, 1888 ; E. H. Keiser, ib., 10. 249, 1888 ; W. A. Noves, ih., 11. 155, 1889 ; 12. 441, 1890 ;
13. 354, 1891 ; G. S. Johnson, Chem. News, 59. 272, 1889. *
i» E. H. Keiser, Amer. Chem. Journ., 13. 253, 1891 ; 20. 773, 1898 ; W. A. Noyes, Journ.
Amer. Chem. Soc., 29. 1718, 1907 ; 30. 4, 1908 ; E. W. Morley, Amer. Chem. Journ., 17. 267,
1895 ; Proc. Amer. Assoc., 11. 185, 1891 ; On the Density of Hydrogen and Oxygen, and on the
Ratio of their Atomic Weights, Washington, 1895.
*'' A. von Humboldt and J. L. Gay Lussac, Journ. Physique, 60. 129, 1805 ; J. A. C. Chaptal
and C. L. Berthollet, Ann. Chim. Phys., (1), 53. 239, 1805 ; A. Scott, B. A. Rep., 668, 1887 ;
Nature, 37. 439, 1888 ; Proc. Roy. Soc, 42. 396, 1887 ; Phil Trans., 184. 543, 1893 ; F. P. Burt
and E. C. Edgar, ib., 216. A, 393, 1916 ; P. A. Guye, Journ. Chim.. Phys., 15. 208, 1917 ; E. W.
Morley, Amer. Chem. Journ., 10. 21, 1888 ; Proc Amer. Assoc, 39. 161, 1891 ; S. Young, Nature,
37. 390, 416, 1888 ; A. Leduc, Compt. Rend., 115. 311. 3892.
21 J. B. A. Dumas and J. B. J. D. Boussingault, Ann. Chim. Phys., (3), 3. 257, 1841 ; H. V.
Regnault, ib., (3), 14. 211, 1845; (3), 15. 512, 1845; P. von Jolly, Wied. Ann., 6. 520, 1879;
G. Agamenone, Atti Accad. Lincei, (4), 1. 665, 699, 1885 ; Lord Rayleigh, Proc Roy. Soc, 43.
356, 1888 ; 50. 448, 1892 ; 53. 134, 1893 ; J. M. Crafts, Compt. Rend., 106. 1662, 1888 ;
A. Leduc, ib.. 111. 262, 1890; 113. 186, 1891 ; 116. 1248, 1893 ; 117. 1072, 1893 ; Recherches sur
Zfts gaz, Paris, 1898 ; A. Jaquerod and A. Pintza, Compt. Rend., 139. 139, 1905 ; P. A. Guye
and E. Mallet, ib., 138. 1034. 1904; A. Jaquerod and O. Scheuer, ib., 140, 1384, 1905;
R. W. Gray, Journ. Chem. Soc, 87. 1607, 1905 ; P. A. Guye, Journ. Phys. Chim., 5. 213, 1907 ;
J. P. Cooke, Amer. Chem. Journ., 11. 509, 1889; E. W. Morley, Proc Amer. Assoc, 39. 163,
1890 ; J. Joly, Proc Roy. Soc Dublin, (2), 6. 534, 1890 ; J. Thomsen, Zeit. anorg. Chem., 11. 14,
1896; 12.4,1896.
22 Lord Rayleigh, Phil Trans., 196. 205, 1901; 198. 417, 1902; 204. A, 351, 1905;
P. A. Guve, Journ. Phys. Chim., 3. 346, 1905 ; 15. 208, 1917 ; Bull Soc Chim., (3), 33. 1, 1905 ;
D. Berthelot, Seances Soc. Phys., (3), 7, 143, 1898 ; CompL Rend., 126. 954, 1030, 1415, 1501, 1898 ;
P. A. Guye, ib., 144. 976, 1907 ; A. Jaquerod and O. Scheuer, Compl. Rend., 140. 1384,
1905; A. Leduc, Journ. Phys., 152, 1897; Recherches sur les Gaz, Paris, 1898; A. Leduc and
P. Sacerdote, Compt. Rend.,\2%. 218, 1853, 1898.
2« J. S. Stas, Bull Acad. Belgique, 10. 294, 1860; Mem. Acad. Belgique, 35. 48, 1865; 43.
62, 1882 ; J. Thomsen, Zeit. phys. Chem., 13. 398, 1894 ; L. Meyer and K. Seubert, Ber., 27.
2770, 1894 ; J. Dewar and A. Scott, Proc Roy. Soc, 35. 347, 1887.
§ 10. The Origin of the Terms : Acid, Alkali, Base, Salt
However convenient the classification of oxides into acids and bases might be for an
elementary presentation of chemistry, a glance from the vantage ground of facts not
usually referred to in elementary courses, shows such classifications to be imperfect and
arbitrary to a degree.— D. Carnegie (1894).
The early chemists appear to have gradually learned to arrange certain sub-
stances into two groups according as these substances possessed certain qualities
in common with vinegar or with wood ashes. The former were called acids {acidus,
acid) and the latter alkalies (Arabian, alkali, ashes of a plant), because the alkalies
were generally obtained by calcining various materials and reducing them to ashes.
The word acid was probably first used in a concrete sense for vinegar, and it then
came to be used for certain substances which tasted " sharp " or sour like vinegar —
OXYGEN 383
the acid of soured wine ; the term alkali was used for crude potash ; and salt has
been used from the earliest times for culinary salt. Aristotle employed the term
salt for the evaporated lixivium of wood ashes. i Disocorides and Pliny employed
the same term for crude soda, and generally for substances which could be recovered
from their solution in water by evaporation. About the time the works of " Basil
Valentine " were written, the vitriols were regarded as metallic salts, and the term
salt came to be employed for that constituent of a substance which could not be
destroyed by calcination, and among the alchemists the term was used to represent
the principle of solidity. H. Boerhaave, T. Bergmann, and R. Kirwan used
solubility in water as one criterion for salts, but this led to the separation of substances
of a similar nature into separate groups.
Although the three terms — acid, alkali, and salt — were first applied to specific
substances, their meanings have changed so that they no longer designate the names
of things, but are employed as generic or class names to indicate what certain things
will do ; otherwise expressed, they are the names of certain chemical functions.
The generic term only becomes specific when an adjective is affixed — e.g. sulphuric
acid.
The great solvent or corrosive action of the acids was well known to the ancients.
This is emphasized by Pliny's story of Cleopatra and the pearls ; and Livy's and
Plutarch's fantastic story of Hannibal cutting a passage through the Alps by
dissolving limestone rocks by means of vinegar. In his Reflections upon the
hypothesis of alcali and acidum (London, 1684), Robert Boyle summarized the
properties of acids as substances which (1) have a sour taste ; (2) dissolve many
substances (corrosive) ; (3) precipitate sulphur from alkaline solutions of sulphur ;
(4) change the tint of many vegetable blue colours (e.g. blue litmus) red ; and
(5) lose their acid characteristics when brought into contact with the alkalies.
H. Boerhaave (1732) ^ divided the acids into acida vegetantia, or those derived
from plants ; and acida fossilia, or those derived from mineral substances. Soon
afterwards, J. van Helmont (1736) called the latter, acides mineraux, and included
them in the class containing sulphuric, hydrochloric, and nitric acids.
J. B. van Helmont (1640), F. Sylvius de la Boe (1659), N. Lemery (1675) and
H. Boerhaave (1732), at first, applied the term alkali to bodies which effervesced with
acids, but R. Boyle recognized as alkalies certain substances which do not act in
this manner. R. Boyle considered the alkalies to be substances which (1) possessed
detergent and soapy properties ; (2) dissolved oils and sulphur ; (3) restored
vegetable colours reddened by acids ; and (4) had the power of reacting with acids
to produce indifferent substances. The idea connoted by the term base is much
older than the word ; base stands in generic relations with alkali. F. Sylvius
de la Boe recognized the distinction between acids and bases in 1659, although the
idea was familiar to chemists before his time ; this is emphasized by the fact that
acids and alkalies have a strong disposition to unite chemically. In 1744, G. F«
RouUe employed the word base for any substance which unites with an acid to
form a salt, and which gives to the salt une forme concrete ou solide ; the term is
now usually applied to oxidized bodies with properties which are complementary
to the acids ; the term includes the earths, alkalies, metallic oxides (calces), and
all substances which produce salts by reacting with acids. It was soon found that
some substances with alkaline qualities did not melt or change when heated, did
not effervesce with acids, and were almost insoluble in water — these substances were
called earths.
With these criteria, it is possible to classify the oxides formed by burning carbon,
sulphur, phosphorus, sodium, iron, etc., in oxygen into acidic, basic, or neutral
oxides :
IDIC OXIDES.
Basic oxides.
Neutral oxides.
Carbon
Sulphur
Phosphorus
Sodium
Calcium
Magnesium
Iron
Copper
Tin
384 INORGANIC AND THEORETICAL CHEMISTRY
The properties of acids and alkalies or bases were thus opposed to one another ;
for when mixed together, the one neutraUzed the activity of the other. Although
each component of the mixture is itself pungent or corrosive, the final product is
usually mild and inoperative. J. B. van Helmont, about 1640, used the term satura-
tion, and a few years later, 0. Tachen gave one of the first rational definitions of a
salt, for he said that all salts can be resolved into an acid and an alkali, and very
soon, the term salt came to be used for the products of the interaction of acids and
alkalies or bases. John Mayow, in his essay On the combination of contrary salts
(Oxford,. 1669), recognized the dual nature of salts, and showed that although the
acid and alkali, when they meet, unite together to form a salt, yet they do not
destroy one another since both may be afterwards recovered from the salt. According
toN. Lemery (1675), during saturation, the base is " cloyed or filled with acid," and
a sel sale was defined as an alkali charged with an acid. G. E. Stahl (1723) called the
substance which united with the acid to form sodium chloride, 7nateria ilia quce
soli corpus prcebet. H. Boerhaave (1732) stated that an alkah is understood to be
saturated with an acid when a point is reached at which the product is neither acid
nor alkaline ; the resulting product was called a neutral salt — salia^ sic dicta jam
neutra. H. Boerhaave also spoke of salia alcalina, and of salia acida ; and he
regarded the vitriols as semi-metals.
It was soon recognized that many substances could not well be grouped with the
acids and bases although they possessed qualities characteristic of acids or bases.
Thus aluminium ammonium sulphate — alum — forms a solution with water which has a
sour taste, deprives sodium hydroxide of its alkaline qualities, and turns blue litmus
red ; copper sulphate reddens blue litmus ; sodium carbonate and sodium borate
turn red litmus blue, etc. Conversely, substances may be grouped as acids and bases,
even though they have no action on litmus, e.g. silicic acid, H2Si03, has no action on
blue Utmus, and yet it is an acid ; similarly, copper oxide, CuO, is a base without
action on red litmus.
What is the source o£ the acidity of acids ? — Otherwise expressed, why are the
acids acidic, and the bases alkahne ? J. J. Becher (1669), G. E. Stahl (1723), and
other early chemists postulated the presence of a principle of acidity — all acids in
common were supposed to be impregnated with more or less of a primordial or
primitive acid. Eor example, F. Sylvius de la Boe and J. F. Meyer 3 supposed both
acids and alkalies owed their pecuhar properties to the presence of a common principle
which the former termed fiery matter and the latter acidum pingue. When the
alkalies came to be divided into caustic and effervescent, it was assumed that
the transformation of caustic into effervescent Hme was due to the effervescent
lime transferring its acidum pingue to the effervescent alkali. G. E. Stahl
assumed that acids, alkalies, and salts contained one common ingredient — namely,
the primitive acid ; and that the three were transmutable inter se by adding or sub-
tracting primitive acid. The alkalies were supposed to contain less primitive
acid than acids or salts. T. Bergmann postulated a principle of acidity and a principle
of alkalinity, but he admitted that while " chemistry was not able to extract these
two universal principles, there is not the least doubt that they are different and oppo-
site to each other." Not till 1755 did J. Black * demonstrate qualitatively and
quantitatively that the loss of fixed air changes an effervescent into a caustic
alkali, and the union with fixed air changes a caustic into an effervescent alkali.
N. Lemery considered it to be self-evident that the ultimate particles of acids had
sharp edges or hooks which gave the acids their peculiar properties. His proof was :
I hope nobody will dispute whether an acid has points or not, seeing that it is demonstrated
by every one's experience, that an acid pricks the tongue Hke anything keen and finely cut ;
but a demonstration and convincing proof that an acid consists of pointed parts is that not
only all acid salts do crystallize with edges, but all dissolutions of different things caused
by acid liquors, do assume this figure in their crystallization.
This argument is invalid, for the crystals of nearty all compounds — whether acidic
OXYGEN 385
alkaline, or neutral — have sharp edges. The lesson is obvious. What is self-
evidently an hypothesis must not be advanced as if it were an inviolable fact.
A. L. Lavoisier, following J. Mayow (1669) and C. W. Scheele (1777), renounced
the peculiar fancies involved in the terms primordial acid, principle of acidity, etc.,
and ascribed acidity to the universal presence of oxygen, the acid-producer. Conse-
quently, Lavoisier's oxygen is the acidifying principle under another guise. J. Mayow
got very near to the same theory of acidity in 1669. This hypothesis can now be tested.
Befebences.
1 H. Kopp, Geschichte der Chemie, Braunschweig, 3. 1-61, 1845 ; E. J. Mills, Phil. Mag., (4),
37. 461, 1869 ; G. 0. Foster, ib., (4), 29. 262, 1865 ; (4), 30. 57, 1865 ; D. Vorlander, Journ.
prakt. Chem., (2), 87. 84, 1913 ; R. Meyer, ib., (2), 87. 280, 1913 ; A. W. WiUiamson, Joiirn.
Chem. Soc, 17. 421, 1864 ; Phil. Mag., (4), 29. 466, 1865.
2 H. Boerhaave, Elementa chemice, Lugduni Batavorum, 1732 ; J. Hellot, M6m. Acad., 36,
1736 ; J. B. van Hehnont, Oriiis medicmce, Amsterdam, 1648 ; F. Sylvius de la Boe, Opera
omnia, Paris, 1671 ; N. Lemery, Cours de chimie, Paris, 1675; G. E. Stahl, Amfiihrliche Beirach-
lung and zulanglicJier Beweiss wn den Saltzen das dieselben aiis eincr Zarten Erde mit Wasser innig
verbunden bestehen, Halle, 1723 ; G. F. Rouelle, Mem. Acad., 347, 1731; 97, 1744; 672, 1752;
0. Tachen, Hippocrates chemicus, Venice, 1666.
^ J. F. Meyer, Chymische Versuche, Hannover, 1764.
* J. Black, Experimental Essays, London, 1764.
§ 11. Acids
Acid is rather the name of a function than the name of a substance. — ^Mills.
In his study of the properties of oxygen, A. L. Lavoisier noticed that when certain
elements were burnt in oxygen, the resulting oxide forms an acid with water — e.g.
carbon, sulphur, and phosphorus. Hence, Lavoisier jumped to the conclusion
(1777) that " oxygen is an element common to all acids, and the presence of oxygen
constitutes or produces their acidity." He also considered oxygen to be the
essential constituent of all acids. The very name oxygen, given to this element, was
derived from Greek words signifying " the generative principle of acids " — o^v^,
sour, and yeiVo/xat, I produce — because " one of the most general properties of
this element is to form adds by combining with many different substances ; " hence
also the German term for oxygen Sauerstoff, meaning "acidifying stuff." In his
Considerations generates sur la nature des acides, and his Memoire sur Vexistence de
Vair dans Vacide nitreux (1777), A. L. Lavoisier considered that the difference in
the various acids depended on the nature of the substance or substances united with
the oxygen ; he called the non-oxygenated part of an acid a simple or a compound
acidifiahle base. The mineral acids are usually oxygenated compounds of simple
acidifiable bases — carbon, sulphur, nitrogen, phosphorus — the vegetable and animal
acids — tartaric and oxalic acids — are oxygenated compounds of the compound
acidifiable bases. Lavoisier's theory of acidity made him unprepared to find water
or oxidized hydrogen exhibiting no signs of acidity, and there is a possibiHty that
his hypothesis prevented his discovering the composition of water. L'analogie,
said Lavoisier,! m'avait jporte invincihlement d conclure que la combustion de Vair
imjiammable devoit egaleynent produire un acide. The difference between analogy
and fact, added Berthollet, is just the difference between probability and certainty.
For a time, le principe oxygene was almost a fetish with the French chemists ; but,
with increasing knowledge, it was found that Lavoisier's oxygen theory of acids
led to confusion and error, and it was gradually abandoned by chemists when it was
recognized that :
1. Some oxides form alkalies, not acids, with water. — E.g. sodium, potassium,
and calcium oxides. As Humphry Davy expressed it, " the principle of acidity of
the French nomenclature might now likewise be called the principle of alkalescence."
2. Some acids do not contain oxygen. — In 1785, J. C. de la Metherie 2 had
maintained as a paradox that oxygen does form a necessary constituent of acids.
VOL I. 2 c
386 INORGANIC AND THEORETICAL CHEMISTRY
This idea was ridiculed by A. L. Lavoisier, but C. L. Berthollet showed, in 1787,
that hydrocyanic (prussic) acid is a compound of carbon, nitrogen, and hydrogen,
but contains no oxygen ; and he also came to a similar conclusion with regard to
hydro-sulphuric acid. But for some time Lavoisier's reputation had more weight
than BerthoUet's facts. In 1810-11, Humphry Davy proved that hydrochloric
acid is a compound of hydrogen and chlorine and that no oxygen could be detected
in the compound. In 1813, H. Davy also proved that hydriodic acid contained
hydrogen and iodine, but no oxygen. Hence, added H. Davy, " acidity is not
connected with the presence of any one element ; " and he appears to have regarded
the acidity of a substance as a kind of resultant whose direction is hydrogen.
It must be added, however, that when these substances are thoroughly dried so
as to remove all traces of water they do not show acidic qualities. Water is always
present when these substances manifest their acidic properties, and water is itself
a compound of hydrogen and oxygen. Hence, a very good case might be made
out for an extension of Lavoisier's hypothesis, but solutions of ammono-bases and
ammono-acids in liquid ammonia would then have to be considered because this
solvent has no oxygen.
As a result of H. Davy's work, the acids came to be classed as hydracids — acids
containing no oxygen ; and oxyacids — acids formed from acidic oxides. In 1815,
H. Davy suggested the possibility that hydrogen, not oxygen, gives the acid
characters to the acids ; but he did not rush to the other extreme and say that all
hydrogen compounds are necessarily acids. In the same year, 1815, P. L. Dulong's
study of oxalic acid led him to the view that there is no essential difference between
oxygenated and non-oxygenated acids. He supposed oxalic acid to be a com-
pound of oxj^gen with carbon dioxide, and salts of oxalic acid to be formed by
replacing the hydrogen of the acid by the metals. In this way, hydrogen and the
metals were opposed to the salt-forming radicles.
There is no one property which we can use as an absolute criterion or decisive
test of acidity. In a crude sort of way, it can be said that acids usually have a sour
taste, are usually corrosive, redden the blue colour of vegetable substances {e.g. litmus) ;
and contain hydrogen, part or all of which can be replaced when the acid is treated with
a metal, metallic oxide, hydroxide, or carbonate. Acids are known which have a sweet
taste — e.g. amidophosphoric acid ; acids are known which are not corrosive ; and
alkalies have hydrogen replaceable by a metal as is shown when, say, aluminium is
treated with alkali lye. Alum, as indicated above, does not contain replaceable
hydrogen, and it would not therefore be classed as an acid, although it is sour,
corrosive, and colours blue litmus red. Sodium bisulphate has a sour taste, is
corrosive, reddens blue litmus, and contains replaceable hydrogen, but it is not
usually regarded as an acid because of its mode of formation. Again, methane,
CH4, is not considered to be an acid although it has hydrogen replaceable by a metal,
and the resulting compound is not called a salt, e.g. zinc methide, Zn(CH3)2.
However, we are yet far from a satisfactory definition of acids, although, as we
shall see later, a fair definition can be made in terms of the ionic hypothesis in spite
of the fact that so far as practical applications are concerned, definitions in terms of
the ionic hypothesis are not very different from those under consideration, the
difference is then rather a question of nomenclature.
Naturally the student delights in clear, sharp-cut definitions, and teachers of
science have many temptations to frame definitions and draw boundary lines which
do not exist in nature. " Definitions," said R. Hunter, " are the most accursed of
all things on the face of the earth."
References.
1 A. L. Lavoisier, Mein. Acad., 471, 1781 ; C. L. Berthollet, Ann. Chim. Phys., 3. 63, 1789.
* J. C. de la M6therie, Essai analytique aur Vair pur, et les dijfirentes especes d'air, Paris, 1785 ;
C. L. Berthollet, Ann. Chim. Phys., (1), 1. 30, 1787; P. L. Dulong, Mem. Acad., 23, 1815; Ann.
Phil, 7. 231, 1815; H. Davy, Phil. Trans., 90. 191, 1800; 99. 39, 460, 1809; 100. 231, 1810;
101. 1,1811; 105.219,1816.
OXYGEN 387
§ 12. Salts
The relation of acid and base in a salt is one of the main grounds of all theoretical
reasonings (on chemical combination).- — ^W. Whewell.
There is perhaps no inquiry which has thrown so much light on a multitude of combina-
tions which the want of method had permitted to be confounded, as Rouelle's observations
on the characters of the salts.' — C. L. BERTHOLiiET (1803).
In modern chemistry the term salt is a descriptive term applied to a distinct
class of substances, and not to any particular individual. A salt is produced by
replacing all or part of the hydrogen of an acid by a metal or basic radicle. For
instance, zinc displaces the hydrogen of sulphuric acid : ZnH-H2S04=ZnS04H-H2,
forming the salt zinc sulphate. C. Gerhardt i regarded salts as corps hinome, liable
to double decomposition ; and J. J. Grilfin regarded them as compounds of two
radicles. Hence J. von Liebig (1838), and C. Gerhardt (1843) defined acids to be
" salts of hydrogen " :
804^ — Bivalent Radicle. CI — Univalent Radicle.
Hydrogen sulphate (sulphuric acid) H2SO4 Hydrogen chloride (hydrochloric acid) HCl
Zinc sulphate .... ZnSOi Zinc chloride .... ZnCl2
Sodium sulphate . . . Na2S04 Sodium chloride . . . . NaCl
Salts of the binary acids (i.e. acids compounded of two elements like hydrochloric
acid, etc.) are usually named by dropping the prefix hydro and changing the termina-
tion -ic into -ide. Thus the acids just named furnish chlorides, fluorides, etc.
To show what chlorides, etc., are in question, the name of the corresponding element
(or elements) is introduced in an adjectival sense. Thus we have sodium chloride,
potassium chloride, calcium chloride, etc. The names of the elements are thus used
adjectivally in the same sense that the words stone, brick, and wood prefixed
to house are adjectival, and indicate the kind of house in question. Some
radicles free from oxygen, e.g. CN or Cy, behave as if they were single elements.
Thus hydrocyanic acid— HCN— is treated as if it were a binary acid, and its salts are
accordingly named cyanides.
The salts of the ternary acids [i.e. acids with three elements) are named by
changing the -ic termination of the acid into -ate, or the -ous termination of the acid
into " -ite," and adding the word so obtained to the name of the base or bases form-
ing the salt. The sulphuric acid forms sulphates — e.g. sodium sulphate ; nitric acid,
nitrates — e.g. calcium nitrate ; sulphurous acid, sulphites—e.g. ammonium sulphite ;
perchloric acid, perchlorates — e.g. potassium perchlorate ; hypochlorous acid, hypo-
chlorites— calcium hypochlorite ; carbonic acid, carbonates — e.g. calcium carbonate,
etc. Hence, some years ago, the name of the basic element used to be modified to
give it an adjectival form : hydric chloride ; potassic chloride ; calcic chloride ; etc.
This system has been abandoned unless it is desired to distinguish between -ous and
-ic compounds — e.g. ferrous chloride and ferric chloride, etc. Consonant with the
definition that acids are salts of hydrogen, nitric acid, HNO3, has been called
hydrogen nitrate ; hydrochloric acid, HCl, hydrogen chloride ; sulphuric acid,
H2SO4, hydrogen sulphate.
In normal salts all the displaceable hydrogen of the acid is replaced by the base.
For instance, sodium sulphate — Na2S04 — is a normal salt because all the replaceable
hydrogen of sulphuric acid is displaced by sodium. In acid salts only part of the
replaceable hydrogen has been displaced, acid sodium sulphate — NaHS04 — contains
half the replaceable hydrogen of sulphuric acid, and half as many equivalents of
sodium as normal sodium sulphate. If an acid contains two or more replaceable
hydrogen atoms, it does not follow that all need be displaced by the same element.
These ideas can be illustrated graphically — sulphur sexivalent :
HO^S^O HO^S^O NaO-^S^O KO-^^^O
Sulphuric acid. Acid sodium sulphate. Normal sodium sulphate. Sodium potassium sulphate.
About 1754, F. G. Rouelle 2 distinguished the neutral salts of a given base from
the acid salts with an excess of acid, and basic salts with an excess of base, and he
388 INORGANIC AND THEORETICAL CHEMISTRY?
showed the action of some of these salts on vegetable dyes. A. Bamne (1770)
objected to F. G. Rouelle's classification — neutral, acid, and basic salts — for he
maintained that the neutral salts are the only true salts, and that the basic and acid
salts are mixtures of neutral salts respectively with an excess of bases or acids. The
question was discussed later, whether the proportion of acid and base in the two salts
is constant so that there are but two combinations — those with a maximum and those
with a minimum proportion qf acid — or whether combinations exist with intermedi-
ate proportions dependent on the circumstances under which the salts are formed.
C. L. Berthollet (1803) advocated the latter alternative, but this was not generally
accepted, and guided by the law of multiple proportions the two salts are usually
considered to be combinations of acid and base in two definite and fixed proportions.
When deviations from these proportions are observed, it is supposed that a mixture of
normal and acid salts, or of the normal or acid salt with an excess of uncombined
acid or base, is in question.
Sometimes the term " hydrogen " is used in place of " acid " for the acid salts, and
sometimes the prefix bi- or di- is appended to the term for the acid in the salt. Thus,
acid sodium sulphate is also called sodium hydrogen sulphate, sodium bisulphate,
as well as mono-sodium sulphate, etc. Originally, the Latin prefixes bi-, etc., were
applied to the name of the acidic and the Greek prefixes di-, etc., to the name of the
basic part of a salt, so that sodium disulphate would not mean the same as bisulphate.
Colloquially, the prefixes bi-, etc., are used for the acid salt — e.g. sodium bicarbonate,
NaHCOs ; sodium bisulphite, NaHSOs, ^tc. — possibly because a term like sodium
hydrogen carbonate appears to be pedantic outside the lecture room.
The normal salts are sometimes called neutral salts in the sense that all the
hydrogen has been neutralized or displaced from the acid. These salts, however,
are not necessarily nevtral to litmus — thus normal zinc and copper sulphates react
towards litmus as if they were acids ; borax, sodiimi nitrite, and normal sodium
carbonate react as if they were alkalies. The confusion in the use of the term acid
may also be noted. It can be used as a noun to denote a particular class of compounds
now under discussion ; and also as an adjective to represent a certain quality or
property — e.g. the behaviour towards blue litmus — characteristic of the class acids.
Accordingly, some acid salts are acid to litmus, e.g. sodium hydrogen sulphate ;
others are alkaline, e.g. sodium hydrogen carbonate, acid potassium tellurate ;
others again are neutral, e.g. disodium hydrogen phosphate. Usually the normal
mercurous, mercuric, cupric, chromic, ferric, stannous, stannic, antimonious, and
bismuthous salts with the common acids have an acid reaction — redden blue litmus ;
while the borates, carbonates, chromates, hypochlorites, nitrites, phosphates, sili-
cates, sulphides, and sulphites have an alkaline reaction — turn red litmus blue.
The mode of defining a neutral salt as a compound obtained by mixing an acid
with a base until the product is neutral to litmus is ambiguous and therefore objec-
tionable ; the close relation between salts which are and salts which are not neutral
to litmus does not allow them to be distinguished from one another. Hence, J. J.
Berzelius proposed to retain the term neutral for those salts like potassium and
sodium sulphates which are neutral to litmus, and to take no account of the behaviour
of the other metallic salts towards litmus, but rather be guided by their analogy
with the salts which are neutral to litmus. This extension of the term neutral is
not free from objections. When the term neutral salt is now employed, it is usually
understood to be synonymous with normal salt.
It is sometimes necessary to use the prefixes mono-, di-, tri- ... to discriminate
between the different salts of one acid. Thus with phosphoric acid — phosphorus
quinquevalent :
H0\
HO-^P=0
HO/
NaO\
HO-^P^O
HO/
NaO\
NaO )P=0
HOA
NaO\
NaO^P=0
NaO/
Phosphoric
Monosodium
phosphate.
Disodium
phosphate.
Normal or tri-sodium
phosphate.
OXYGEN 389
It would be a mistake to assume that all the hydrogen of an acid is replaceable
by a base. Thus, so far as we know, hypophosphorous acid — H3PO2 — has only one
of its three hydrogen atoms replaceable by a metal. No one has ever prepared
Na2HP02. Following the recommendation of J. von Liebig (1857) : The number
of atoms of hydrogen in one molecule of an acid which are replaceable by a
metal, or a radicle, is termed the basicity of the acid. Thus hydrochloric acid—
HCl — is monobasic because each molecule of hydrochloric acid contains one replace-
able hydrogen atom ; sulphuric acid — H2SO4 — is dibasic ; phosphoric acid — H3PO4 —
is tribasic ; and ferrocyanic acid — H4FeCy6— is tetrabasic because the four hydrogen
atoms can be replaced by equivalent atoms of the basic elements^ — say four of potas-
sium, two of calcium, etc. Hypophosphorous acid — H3PO2 — is monobasic because
only one of the three hydrogen atoms can be replaced by a metal.
Alcohols and ethers. — The normal alcohols, methyl alcohol, CH3OH ; ethyl
alcohol, or simply alcohol, C2H5.OH ; etc., are related to water in that one hydrogen
atom of water is replaced by a univalent hydrocarbon radicle. When both hydrogen
atoms of water are replaced by univalent hydrocarbon radicles, the so-called ethers
are formed ; thus, (CH3)20 represents methyl ether or methyl oxide ; and ethyl ether or
ethyl oxide or simply ether is represented by (C2H5)20 ; and methyl ethyl ether,
by(CH3)(C2H5)0. Graphically,
H\Q C2H5\^pj CgHg-^^ ^^3^0 ^2H5\q
Water. Ethyl alcohol. Phenol. Methyl ethyl ether. Ethyl ether.
The hydrogen of the residual OH-group in the alcohols can be replaced by a metal
— e.g. sodium dissolves in ethyl alcohol with evolution of hydrogen. When the
solution is heated to 200° in a current of hydrogen to drive ofi the excess of alcohol,
the white powder which remains is sodium alcoholate, C2H5.0Na, which seems to
suggest that water and alcohol have acidic properties in that they contain replace-
able hydrogen. These compounds are included in the subject mattfer of organic
chemistry.
References.
1 J. J. Griffin, The Radical Theory in Chemistry, London, 1858 ; C. Gerhardt, Compt. Rend.,
17. 312, 1843; J. von Liebig, Liebig's Ann., 26. 113, 1838.
2 F. G, Rouelle, Mem. Acad., 347, 1731 ; 97, 1744 ; 572, 1754 ; A. Baum6, Chymie expiri-
mentale et raisonnie, Paris, 1773; C. L. BerthoUet^ Essai de slatique chimique, Paris, 1803.
§ 13. Neutralization
Neutrality measured by means of a colouring matter is but a hypothesis. — J. S. Stas
(1866).
A solution of sulphuric acid, like other acids, colours blue litmus red ; and a
solution of sodium hydroxide, like other alkalies, colours red litmus blue. It is
possible to mix the acid with the alkali so as to furnish a solution which neither
tastes nor reacts towards litmus like sulphuric acid or like sodium hydroxide. If
too much acid be present the litmus will be coloured red, and blue if too much alkali
be present. The mixture on evaporation furnishes a crystalline solid which neither
colours blue litmus red nor red litmus blue. The colour of a violet solution of
litmus is not affected. The product of the reaction is said to be neutral, and the
process of neutralization consists in adding an acid to an alkali, or an alkali to an
acid, until a neutral substance is obtained. The result of the reaction is called a
salt. The salt contains the metal of the alkali, and the radicle of the acid. The
litmus used to determine the point of neutralization is called the indicator. Several
other indicators besides litmus are available ; e.g. phenolphthalein furnishes a pink
coloration with alkaline solutions, and is colourless with acids and neutral solutions ;
390 INORGANIC AND THEORETICAL CHEMISTRY
methyl orange is yellow with alkalies, pink with acid, and orange with neutral solu-
tions. Poirrier's soluble blue gives a blue colour with acids and with carbonates, and
red with alkaUes. The petals of white flowers usually change to yellow when placed
in alkaline solutions and back to white when placed in acids ; red and purple flowers
usually become green or greenish-blue in alkaline solutions, and back to the original
or a brighter red colour in acids ; yellow flowers are usually not afltected by acids
or alkalies. Yellow turmeric becomes reddish-brown in alkaline solution. Litmus
and phenolphthalein solutions can each show the presence of about IS parts of
hydrochloric acid per million parts of solution, while methyl orange will show 3'7
parts of this acid per niilUon.
It will be observed that the determination of the neutral point is here referred
arbitrarily to the behaviour of litmus, because when compounds are arranged into
classes, acidic, alkaUne, and neutral, the members of the different classes are not the
same when different indicators are used. For instance, salts of the heavy metals —
alum, ferric chloride, ferrous sulphate, etc. — are usually acid to both litmus and
phenolphthalein as indicators, and neutral to methyl orange ; borax, sodium and
potassium bicarbonates, and sodium citrate are alkaline to litmus and methyl orange,
but neutral to phenolphthalein ; while sodium phosphate is neutral to litmus and
phenolphthalein, and alkaline to methyl orange. When free phosphoric acid is
titrated with sodium hydroxide, it appears to be a monobasic acid if methyl orange
(i.e. the commercial dimethylaminoazobenzene sulphonate) be used as indicator ;
to be a dibasic acid with litmus or phenolphthalein ; and to be a tribasic acid with
Poirrier's soluble blue. Again, free boric acid does not affect methyl orange, but it
reacts acid with both litmus and phenolphthalein after adding an equal volume of
dilute sodium chloride solution ; and an aqueous solution of potassium sulphite is
neutral to phenolphthalein, but it turns violet litmus blue. It therefore follows that
acids and alkalies have only a relative existence. This was emphasized by
J. Freind ^ in 1709, for he pointed out that the corrosive and colorimetric properties
of acids are often shared aUke with the alkalies, so that a substance which is termed
an alkali if referred to one body, might be called an acid by the very same writers
if referred to another body ; and added : "In vain we endeavour to fix the boun-
daries which separate each kind."
Other properties of acids and alkalies have been employed to determine the neutral
point and also the point where a normal salt is formed when an acid is treated with
an alkali and conversely. The index of refraction, electrical conductivity, and the
freezing point may be cited in illustration. E. Cornec (1909) 2 showed that when the
lowering of the freezing point of a solution is plotted with the composition for all
proportions of acid and base, the minimal points in the curve correspond with sharply
defined salts and bends occur where partly stable compounds occur. Thus, when
JiV-sodium hydroxide is added in gradually increasing quantities to JiV-hydrochloric
acid, the lowering of the freezing point falls regularly from 1*885" to 0-890° when
equivalent proportions are present, and then regularly rises to 1*705°, the value for
sodium hydroxide. The curve. Fig. 5, thus consists of two straight lines which inter-
sect at a point corresponding with sodium chloride. Sulphuric acid and sodium
hydroxide give a well-defined minimum corresponding with Na2S04, but there is no
sign of the formation of an acid salt, NaHS04, in the solution. Hence, it is inferred
that the acid sulphate dissociates completely in solution into the normal sulphate and
free acid : 2NaHS04->Na2S04+H2S04. Phosphorous acid, H3PO3, behaves like a
dibasic acid giving a minimum point with Na2HP03 ; hypophosphorous acid, H3PO2,
behaves like a monobasic acid ; arsenic and phosphoric acids with sodium hy-
droxide give minima corresponding respectively with Na3P04 and Na3As04, and bends
or terraces in the curve for phosphoric acid correspond with the partial formation
of NaH2P04 and Na2HP04 ; analogous bends in the arsenic acid curve. These
two acids with ammonia give curves with minima corresponding respectively with
(NH4)2HP04 and (NH4)2HAs04. Chloric acid, HCIO3, and perchloric acid, HCIO4,
give graphs characteristic of monobasic acids ; selenious acid, H2Se03, dithionic
OXYGEN
391
acid, H2S2O6, and carbonic acid, H2CO3, all behave as dibasic acids ; hypophosphoric
acid, H4P2O6, and pyrophospboric acid, H4P2O7, behave like tetrabasic acids.
As a corollary from Richter's law of proportionality that acids and alkalies unite
in constant proportions to form salts, it follows that when two neutral salt solutions
mutually decompose one another, the newly formed products are also neutral,
because the amount of base neutralized by a certain weight of one acid is also neutral-
ized by an equivalent weight of another acid. In illustration, when a neutral
aqueous solution of sodium chloride is added to an equivalent solution of silver
nitrate, the solution remains neutral after the precipitation of the silver chloride. It
also follows from Richter's law that if one metal be precipitated by another metal
from a neutral salt, the neutrality is maintained. T. Bergmann (1785) knew that
when one metal is precipitated by another from a neutral salt solution, the neutrality
is not disturbed, which he interpreted in terms of the phlogiston theory by assuming
that the quantities of two metals which are united with the same amount of acid
contain the same amount of phlogiston. Richter's law of neutraUty — Neutrali-
tdtsge.s'etz — is obviously a special case of the law of reciprocal proportions
20
• 5 I
ioa
0-6'
\,
s
\,
/
\
s.
/
\
\,
/
s
\,
y
/
/Vn
r
0 10
Acid
20 30 40 50 60 70 60 90 100
N
^o/
.ro.e-
-J
0 3°
I
'~
/
N
X
-/Va
%P
/
^
0^
i
^
Mfo.
^Na^PO^
(HCi)
Base (NaOH)
10 20 30 40 50 60
Acid (H3 PO^)
70 80
Base
90 100
(NaOH)
Fig, 5.— Lowering of the Freezing Point of
Solutions of Sodium Hydroxide and
Hydrochloric Acid.
Fig, 6.- — Lowering of the Freezing Point of
Solutions of Sodium Hydroxide and
Phosphoric Acid.
which, as previously shown, was recognized a few years later. J. J. BerzeHus (1827)
unfortunately confused the work of Wenzel and Richter on this subject, and the
mistake was continued by later writers up to about 1841.
Neutralization versus hydrolysis. — The process of neutralization of a basic
hydroxide by an acid is attended by the formation of a salt and water. We shall find
later that some salts — e.g. zinc sulphate, sodium carbonate, potassium cyanide,
etc. — are partially decomposed — i.e. hydrolyzed — by water into acid and base.
The action of water on such a salt or base is thus an example of an opposing
reaction ; hydrolysis is opposed to neutralization :
H2S04+Zn(0H)2:
Neutralization— >
:ZnS04+2H20
<— Hydrolysis.
In some cases, however, the amount of hydrolysis is inappreciable, and the process
of neutralization is so complete that it can be employed for measuring the quantity
of acid or base in a given solution. Hence, the chemical action of water as a solvent
can be neglected in many chemical reactions, but in other cases the solvent is of
prime importance, when it determines the nature of the compound formed. Probably
most of the examples of normal salts, which furnish acid or alkaline solutions, are
hydrolyzed by water, and the acidic or alkaline properties of the aqueous solutions
are due to the corresponding products of hydrolysis.
Acidimetry and alkalimetry. — A standard solution containing a known amount
of acid or base per litre is prepared, and just sufficient of this solution is added to
neutralize a solution of a given base or acid. The volume of the standard solution
392 INORGANIC AND THEORETICAL CHEMISTRY
required for the purpose is noted. It is possible to calculate the amount of " chemi-
cally pure " substances present in the given solution from the volume of the standard
solution required for the neutralization. A solution containing one equivalent
weight of the acid or base element or compound expressed in grams per litre is
called a normal solution, written " iV-solution," and a solution containing one-tenth
the concentration of a normal solution is called a decinormal solution, written " j-^N-
solution." The equivalent weight of a base is that quantity which just completely
neutralizes one molecular weight of a monobasic acid ; and the equivalent weight
of an acid is that quantity which contains unit weight of replaceable hydrogen.
Thus 36" 47 grams of HCl per litre gives a normal solution of hydrochloric acid ;
and 49*04 grams of H2SO4 per Utre gives a normal solution of sulphuric acid. Here
the molecular weight of the latter acid is 98'08, and the acid is dibasic, for it contains
two replaceable hydrogen atoms ; and, by definition :
Ti . , X p .J Molecular weight of acid
Equivalent of acid = — - — , —,
^ Basicity of acid
that is, the equivalent of sulphuric acid is 98'084-2=49"04. A normal solution
of sodium hydroxide contains 40 grams of NaOH per litre, and a litre of a normal
solution of any acid so far considered will just neutralize a Utre of normal solution
of any base.
.Examples. — (1) Suppose that a 50 c.c. burette be charged with a normal solution of
sodium hydroxide, and suppose that the amount of HCl in 500 c.c. of a dilute solution of
hydrochloric acid be in question- — acidimetry^ — pipette 50 c.c. of the acid into a beaker and
add a few drops of litmus. The alkali solution is run from the burette into the beaker until
the addition of but one more drop of acid is needed to change the red litmus to blue. Suppose
that 42 c.c. of the normal sodium hydroxide has been run from the burette. The argument
runs : The neutralization NaOH + HCl =NaCl+H20 shows that 40 grams of sodium
hydroxide correspond with 36*47 grams of HCl ; and 1000 c.c. of NaOH has 40 grams of
sodium hydroxide, which is equivalent to 36"47 grams of HCl. Consequently 42 c.c. of
the standard sodium hydroxide solution is equivalent to 1'53 grams of HCl per 50 c.c. of
the given acid, or 15'3 grams of HCl are present in 500 c.c. of the given acid.
(2) Suppose that 42 c.c. of a decinormal solution of sulphuric acid were required to just
neutralize 50 c.c. of a solution of potassium hydroxide, how many grams of potassium
hydroxide would be contained in a litre of solution ? A normal solution of sulphuric acid —
^^2804, molecular weight 98- — contains 49 grms. of the acid per litre, and a -ji^^-solution
contains 4'9 grms. per litre, and this is equivalent to 5*6 grms. of potassium hydroxide per
litre. Hence 52 c.c. of the ^j^N-H^^O^ are equivalent to 0*235 grm. of KOH per 50 c.c. of
the given solution. Ansr. 4*7 grms. of potassium hydroxide per litre.
Similar remarks apply to the determination of alkalies — alkalimetry — by stan-
dard solutions of the acids. This process of analysis is called volumetric analysis
in contradistinction to gravimetric analysis^ which involves several weighings during
each determination. In volumetric analysis, the preparation of the stock of standard
solution may involve one or two weighings ; the stock of standard solution may serve
a great number of analyses. Experimental details are discussed in laboratory
text-books.
W. Ostwald employed the term mol as an abbreviation for gram-molecule, that is, a
weight of a compound equivalent to the molecular weight expressed in grams — e.g. one
gram-molecule or one mol of sulphuric acid, H2SO4, is 98 grms. ; and 147 grms. of absolute
sulphuric acid contains 147-^98=1*5 mols or gram-molecules. F. Fichter (1914) pro-
posed val for a gram-equivalent, that is, a weight of a compound numerically the same as
the equivalent weight expressed in grams- — e.g. one gram-equivalent or one val of sulphuric
acid is 49 grms. A millimol is equivalent to a milligram molecule ; and a millival to a
milligram equivalent, etc. The t-erm mol is in fairly common use, but val is not used in
place of equivalent weight.
Rbfeeences.
^ J. Freind, Prcelectiones chymicPy Amstelodami, 1709 ; Chymical Lectures^ London, 1737.
2 E. Comec, Compt. Rend., 149. 676, 1909; 153. 341. 1911; Bull Soc. Chim., (4), 5. 1081,
1121, 1909.
OXYGEN 393
§ 14. Bases
We maintain that hydrogen is an essential not an accidental constituent of all acids and
alkalies.— J. P. Cooke (1876).
Philosophically, acids and bases ought to be regarded as salts. — A. Naquet (1864).
The term base — Greek y3ao-ts, a base — was originally intended to express the idea
that the metal or metallic oxide was the more important constituent, the foundation,
or base so to speak, of a salt. This idea was dropped when it was recognized that
the acidic constituent of a salt is just as important as the basic constituent. The
idea persists in chemistry text-books where the salts are described under the basic
element.
As a first approximation to a satisfactory definition, it was said that a base
is a substance which reacts with an acid to produce a salt and water. For
instance, zinc oxide reacts wHh sulphuric acid to produce zinc sulphate and water :
ZnO+H2S04==H20+ZnS04. Sodium hydroxide reacts with sulphuric atid to
produce sodium sulphate and water: 2NaOH+H2S04=2H20+Na2S04. The
oxides of the non-metallic elements are usually but not always acidic, and the
oxides of the metals are usually but not always basic.
H. Zeitler (1917) ^ illustrates the formation of water during the union of an acid and base
to form a salt by placing a stick of dry alkali hydroxide in a jar of dry hydrogen chloride.
After a short time, the glass is bedewed, and the alkali is covered with crystals of the alkali
chloride.
As a rule, the bases include the oxides and the hydroxides of the metals, but for
convenience, certain groups of elements are called bases, although they form salts
by direct addition or combination without the separation of water, e.g. ammonia —
NH3, hydroxylamine— NH2OH, hydrogen phosphide— PHg, etc. Thus, gaseous
ammonia and hydrogen chloride form ammonium chloride : NH3+HC1=NH4C1.
Liquid ammonia dissolves but does not colour phenolphthalein, and it is an open
question whether it should be called a base. However, the aqueous solution of
ammonia probably forms ammonium hydroxide, NH4OH, which does behave like
the regular bases in this respect : NH40H+HC1=NH4C1+H20.
The definition of a base indicated above is highly unsatisfactory because it in-
volves the definition of an acid, and we have just acknowledged that a satisfactory
definition of an acid is not yet possible. Hence the definition of a base defines the
unknown in terms of the unknown — ignotum per ignotum. Alkali and base are not
synonymous terms. Every alkali is a base, but every base is not an alkali. The
alkali oxides form very soluble hydroxides with marked basic properties, e.g. potassium
hydroxide. The oxides of the alkaline earths form sparingly soluble hydroxides
with less marked basic properties, e.g. calcium hydroxide. The other oxides, as a
rule, do not react directly with water, and the hydroxides are made indirectly. An
oxide cannot be classed as acidic or basic unless it can be shown to produce
corresponding salts. These facts are sometimes summarized in a scheme
resembhng :
Examples.
iVery
soluble .... Alkali oxides
Sparingly
soluble . . . Alkaline earth oxides
uttsca < - j j^^ ^^^ react directly with water Iron and copper oxides
(Hydrides of certain non-metals and their derivatives . . . Ammonia, phosphine
The process of synthesizing a salt from an acid and base can be reversed. In 1803,
J. J. Berzelius and W. Hisinger 2 showed that aqueous solutions of the salts are resolved
into their proximate constituents — acids and bases — by the passage of an electric
current ; and they demonstrated that during the electrolysis of an aqueous solution
of a salt, the acid accumulates about the positive and the base about the negative
pole. This is readily illustrated by electrolyzing an aqueous solution of a neutral
394 INORGANIC AND THEORETICAL CHEMISTRY
salt in a U-tul>e. The liquid is coloured violet with litmus. In a short time, the
liquid about the negative pole becomes blue, and that about the positive
pole red.
Peroxides. — We have seen how barium oxide, BaO — barium bivalent — when
heated under certain conditions forms barium peroxide — Ba02. The peroxides
contain a higher proportion of oxygen than the normal oxides. Barium oxide with
sulphuric acid forms barium sulphate and water : BaO+H2S04=BaS044-H20.
It is therefore a base. Barium peroxide forms barium sulphate, water, and oxygen
when likewise treated with sulphuric acid : 2Ba02+2H2S04=2BaS04+2H20-f-02.
The hypothetical salt, Ba(S04)2, ^ot BaS04, corresponds with barium peroxide.
Hence, barium peroxide is not a basic oxide. If Ba(S04)2 or a related salt could
be prepared, then barium peroxide would, by definition, be a basic oxide.
G. H. Bailey 3 showed that the tendency of the typical oxides of MendeleefE's
table is to form higher oxides — peroxides — without regard to the stability of the
oxide ; with the even series, this tendency is greater in a given (vertical) family of
elements the higher the atomic weight ; and in the horizontal series, also, there is a
tendency to associate with oxygen in passing from left to right, i.e. with increasing
atomic weight. In the odd series, taken vertically, there is less tendency to form
peroiddes as the atomic weight of the positive element increases, and the attraction
for oxygen grows feebler as the atomic weight increases in passing from left to
right, although there is a disposition for the attraction to become more marked
as the extreme right of the period is reached. The constitution and properties
of the peroxides is discussed later.
Amphoteric oxides. — Lead dioxide or peroxide, Pb02 — lead quadrivalent — can
be regarded as a basic oxide because it forms the corresponding salt — PbCl4 — with
hydrochloric acid. But Pb02 also forms salts — ^plumbates — with bases, e.g. potassium
plimibate, 0=Pb=(0K)2. Hence, a substance may be both acidic and basic
according to circumstances. Aluminium hydroxide — aluminium tervalent — is a
base, because, when treated with an acid, it forms a salt — aluminium chloride,
AICI3 — and water :
A1(0H)3+3HC1=A1C13+3H20
But aluminium hydroxide when treated with a base, say, sodium hydroxide, also
forms a salt — sodium aluminate, Al(0Na)3 — and water :
Al(OH)3+3NaOH=Al(ONa)3+3H20
Hence, aluminium hydroxide acts towards an acid like a base, and towards a base
like an acid. Such oxides can be called intermediate oxides, or amphoteric oxides —
from the Greek d/xf^orcpo?, both. Zinc oxide is an intermediate oxide. Stannic
oxide, 0=Sn=O— tin quadrivalent— forms stannic sulphate, S04=Sn=S04,
and also sodium stannate, 0=Sn=(0Na)2 ; hence, stannic oxide is also an inter-
mediate oxide.
Basic salts. — On comparing the graphic formulae of the hydroxides of sodium
(univalent), lead (bivalent), and bismuth (tervalent) :
OH /^H ■
Na-OH Pb<l;S^ Bi^OH
^^ \0H
Unlacidic base. Biacidic base. Teracidic base.
with the graphic formula for mono-, di-, and tri-basic acids we naturally inquire
if the OH or hydroxyl group can be replaced by acid radicles one by one so as to fur-
nish what would be called uni-, bi-, and ter-acidic bases. In the particular examples
just selected, salts corresponding with Pb(0H)N03 and with Pb(N03)2 ; or
T>K/OH r>u^N03
^*^<N03 ^^<N03
Basic lead nitrate. Normal lead nitrate.
OXYGEN 395
are known. The former is called basic lead nitrate^ the latter normal lead nitrate,
or simply lead nitrate. Similarly, Bi(0H)2N08, hasic bismuth nitrate, and normal
bismuth nitrate, Bi(N03)3, are known. The basic salts are thus intermediate in
composition between the normal salts and the basic oxides ; they are usually
derived from the more feeble bases— MgO, ZnO, PbO, CuO, 61263, AI2O3, etc. As
a rule, the basic salts unite readily with other salts to form complexes or double
salts (q.lK).
The basic salts are usually prepared by the action of water or of bases — potassium
hydroxide, aqueous ammonia, etc. — on solutions of the normal salts. Some basic
salts form well-defined crystals, others are more or less amorphous, ill-defined, mud-
like precipitates about which doubts can be raised whether they are really homo-
geneous chemical individuals, for their composition varies with the conditions under
which they are formed, with the amount of washing the precipitate has suffered,
and even with the way the solutions have been mixed. A great many basic salts
have been reported which are probably mixtures or partially decomposed compounds.
Accordingly, there have been many differences of opinion as to the basic salts of
many of the elements ; as one writer has expressed it, " the principle employed in
selecting which are true individuals and which are mere mixtures has been left to
individual taste." The phase rule, to be described later, furnishes a rational basis
which can often be employed in deciding which precipitates are mixtures and
which are compounds.
There is need for a clear understanding of the term baf>ic : NagO represents a basic com-
pound (oxide) ; HNO3 represents a monobasic compound (acid) ; and BiO.NOj represents
a basic compound (salt). Similarly, HNO3 represents an acid ; CUSO4 has an acid reaction ;
NaHCOg represents an acid salt ; and NaOH represents a monoacid base. Two of these
have an acid and two an alkaline reaction.
The basic salts are usually, not always, less soluble in water than the corresponding
normal salts.
References.
A H. Zeitler, Zeit. phjs. chem. Unterr., 30. 35, 1917.
2 W. Hisinger and J. J. Berzelius, Gehlen's Joum., 1. 147, 1803.
3 G. H. Bailey, Journ. Chem. Soc, 65. 315, 1894.
§ 15. Hydroxides and Anhydrides
Acids and alkalies are compounds having the same general molecular structure, and the
differences between acids and alkalies, and, we might add, the differences between individual
acids and individual alkalies, depend on the nature of their radicles. — J. P. Cooke (1876).
We have seen that sulphur dioxide and phosphorus pentoxide form acids with
water :
S02 -f
H2O =
- H2SO3;
and P2O5 +
3H2O -.
= 2H3PO4
Sulphur
Sulphurous
Phosphorus
Phosphoric
dioxide.
acid.
pentoxide.
acid.
The water in these compounds has completely lost its identity, and it is generally
supposed to produce a new class of bodies called hydroxides. Every element,
excepting fluorine and the argon family, appears to form one or more hydroxides,
directly or indirectly. The heats of formation of a few hydroxides 1 from their ele-
ments and water are indicated in Table III. If the heat of formation of the oxide.
Table I, be deducted from these values, the heat of conversion of the oxide into the
hydroxide will be obtained. For the formation of these hydroxides from these
elements add on the heat of formation of the corresponding amount of liquid water,
viz. H2O— 68-36 Cals., and JH^O— 34'18 Cals.
396
INORGAKI/J AND THEORETICAL CHEMISTRY
Table III. — Heats of Formation of some Hydroxides.
Hydroxide.
Cala.
Hydroxide.
Cals.
Hydroxide.
Cals.
KOH
68'99
TlaOa.tcHjO . 86*34
Mn(0H)2
94-77
NaOH
67-69
Sn(0H)2 .
i 68 09
MnO(OH)2
116-33
Cu(0H)2 .
37-52
SnO(OH)2
i 133-50
Fe(0H)2 .
68-28
AuaOsa^HaO
13-19
Pb(0H)2 .
125-16
FeaOg.xHgO
191-16
Ca(OH)a .
146-47
H,PO,
200-06
Co(OH)2 .
63-40
Sr(OH)2 .
146-14
H3ASO4
11309
CoaOa.ajHgO
149-38
Ba(OH), .
146-50
HaSbO*
114-39
Ni203.a;H20
120-38
Mg(OH), .
148-96
Sb(0H)3
83-71
Pd(0H)2 .
22-71
Zn(OH)a .
82-68
Bi(0H)3
68-87
PdO(OH)2.
30-43
Cd(0H)8 .
65-68
H3POS
125-16
PtO(OH)2 .
17-88
Al,Os.a;HjO
388-92
H2SO4
124-56
Pt(0H)2 .
19-22
Tl(OH)
22-73
TeO(OH)a
77-18
The oxides from which the acids are produced do not contain the elements of
water. They are called anhydrides, or acid anhydrides — from the Greek a,
without ; v8(Dp, water. Thus SO2 is not only called sulphur dioxide, but also sul-
phurous anhydride ; and P2O5 is not only phosphorus pentoxide, but phosphoric
anhydride. An anhydride can be regarded as the residue left when the elements
of water are removed from the oxyacids. Thus sulphuric acid, H2SO4, less water,
gives sulphuric anhydride — SO3 — also called sulphur trioxide ; sulphurous acid,
H2SO3, less water, gives sulphurous anhydride — SO2. It is generally supposed that
sulphurous anhydride in combining with water forms a compound containing quadri-
valent sulphur and two univalent hydroxyl — OH — ^groups, that is, S0(0H)2. The
reaction is symbolized :
0=S=0 + H- 0H-> 0=S<
OH
OH
(sulphurous acid)
and sulphuric acid is considered to be a compound containing sexivalent sulphur
and two hydroxyl groups, S02(0H)2. The reaction is symbolized :
}=0 + H-OH->^>S<
OH
OH
(sulphuric acid)
If the acids be regarded as salts of hydrogen, it can be argued that water is a basic
oxide which unites with an acid anhydride to form a salt, e.g. SO3 (acidic oxide)
+H2O (basic oxide) =H2S04 (salt), by analogy with S03-f-K20=K2S04. Isaac
Newton called water a salt. It is easy to show that with the regular definitions
of acid and base, the fame of venerable sulphuric acid — the mother of acids — can
be attacked. Representing sulphuric acid as just indicated, the two hydroxyl
groups can be replaced one by one with other acid radicles, e.g.
o^s<,ci
where the hydrogen of the hydrochloric acid, HCl, is displaced by the bivalent
radicle SO2, " analogous with the formation of magnesium chloride, MgCl2, by the
substitution of the hydrogen of hydrochloric acid by the bivalent atom Mg " in
magnesium hydroxide, Mg(0H)2. This shows how the definitions of acid and base,
if not applied with care, may lead into a bewildering labyrinth. Enough has been
said to show that an acid aiihydride with water forms an acid, and with a base
it forms a salt : Zn0H-S03=ZnS04 (zinc sulphate). Sulphurous acid can also
be regarded as sulphurous hydroxide — S0(0H)2 ; and phosphoric acid — phosphorus
quinquevalent — as phosphoric hydroxide — P0(0H)3. The basicity of an acid is
generally supposed to correspond with the number of hydroxyl groups it contains.
The hydrogen of the hydroxyl groups is supposed to be the displaceahle hydrogen
OXYGEN 397
referred to in the definition of acids. Monobasic hypophosphorous acid — H8PO2 — is
supposed to be H2P0(0H) ; or
.H /^
HO-PO<„ ; also written 0=P^H
^ \0H
because there is only one displaceable hydrogen atom per molecule. The hydrogen
atoms directly united to the phosphorus atoms are not supposed to be replaceable
by the bases, but the hydrogen of the single hydroxyl group is displaceable.
The basic oxides are sometimes called basic anhy^des, and they too form
hydroxides with water, e.g. calcium oxide, CaO — calcium bivalent — with water forms
calcium hydroxide, Ca(0H)2 :
Ca=0 + H-OH -> Ca<^^
From this point of view water can be regarded as hydrogen hydroxide, H— OH,
analogous with K— OH, potassium hydroxide, and Na— OH, sodium hydroxide.
Water itself behaves in some respects as if it were an acid, and in others as if it were
a base. In view of the regular definition, it could be reasonably argued that if
water be an acid, sodium hydroxide, NaOH, is an acid salt, and sodium oxide, Na20,
a normal salt
^>0 ^*>0 ^*>0
jj^U g.^^ Na^^
Water. Sodium hydroxide. Sodium oxide.
Excluding certain carbon compounds, the hydroxides of the non-metallic elements
are usually but not always acids, and the hydroxides of the metals are usually but
not always bases. The term hydroxide is generally reserved for compound of the
basic oxides with water ; and the term anhydride is usually reserved for the acid
anhydrides. The compounds of the basic anhydride with water (hydroxides) were
once called hydrates — e.g. potassium hydroxide was called potassium hydrate, etc.
The term hydiated salt is applied more or less vaguely to compounds which contain
the elements hydrogen and oxygen in the proportion required to form water — com-
bined water.
References.
1 J. Thomsen, Thermochemische Unterauchungen, Leipzig, 2. 395, 1882.
§ 16. The Polar Theory of Chemical Combination.
Strife between opposite tendencies is the parent of ail things. — Heracleitus (c. 4.50 B.C.).
Nature is constantly labouring after repose by the balance and neutralization of contrary
tendencies ; and so far as polar forces enter into her economy, she seeks harmony by means
of discord, and unity by opposition. — W. Whewell (1840).
Every chemical action is fundamentally an electrical phenomenon. . . . Electricity is
the first cause of all chemical action.— J. J. Berzelius (1812).
In some early conjectures on chemical affinity, Isaac Newton (1714) assumed
that electrical and chemical phenomena were both due to attractive forces acting
at insensibly small distances. He said :
The attraction of electricity reaches to sensible distances and so has been observed by
vulgar eyes ; but there may be others which reach to so small distances as to have hitherto
escaped observation. Possibly electrical attraction reaches to small distances, even without
being excited by friction.
Newton seems to have regarded the intensity of chemical affinity to be inversely
proportional to the composition of compound particles. The more complex the
398 INORGANIC AND THEORETICAL CHEMISTRY
aggregates the weaker their affinity. Newton, however, did not consider that the
facts were sufficiently well known to justify further conjectures, for he said :
We must learn from the phaenomena of nature what bodies attract one another, and
what are the laws and properties of the attractions before we inquire the cause by which
the attraction was performed.
H. Davy's electrical theory of chemical afllnity (1807). — After W. Nicholson
and A. Carlisle (1800) had decomposed water, and J. J. Berzelius and W. Hisinger
(1803) had decomposed salts by the electric current, chemists began to suspect that
electrical and chemical forces were closely related. H. Davy took up the subject
about 1806, and in his paper On some chemical agencies of electricity (1807), he
showed that sulphur and copper can be charged with opposite electricities by
friction — the former negatively, the latter positively — just as A. Volta (1792)
proved that when two metals touch one another, they develop electricity — each
assuming an electric charge of opposite sign to the other. H. Davy tried to show
that the chemical activity of a substance is dependent upon its electrical condition,
for he said :
As the chemical attraction between two bodies seems to be destroyed by giving one of
them an electrical state different from that which it naturally possesses ; that is, by bringing
it artificially into a state similar to the other, so it may be increased by exalting its natural
energy. Thus, whilst zinc, one of the most oxidizable of metals, is incapable of combining
with oxygen when negatively electrified in the cu^cuit, even by a feeble power ; silver, one
of the least oxidizable, easily unites to it when positively electrified ; and the same thing
might be said of other metals.
Davy argued that in the act of combination, the reacting substances by contact
acquire electrical charges of opposite signs ; and that chemical combination is
accompanied by a neutralization or exchange of electricities of opposite signs between
the combining substances. Thus, an acid unites with an alkali because the former
acquires an electronegative and the latter an electropositive charge ; oxygen acquires
an electronegative charge, and it unites with metals which acquire an electropositive
charge ; similarly, electronegative sulphur unites with electropositive copper pro-
ducing electrically neutral sulphide. When the quantity of electricity which is
neutralized in the act of combination is restored, the compound is decomposed, and
the original products are reproduced. As a climax, H. Davy virtually said that
chemical affinity is nothing but electrical energy ; for example, in his Elements of
Chemical Philosophy (London, 1812), he said :
Electrical effects are exhibited by the same bodies when acting as masses, which produce
chemical phenomena when acting by their particles ; it is not therefore improbable that
the primary cause of both may be the same, and that the same arrangement of matter, or
the same attracting powers which place bodies in the relations of positive and negative^ —
i.e. which render them attractive of each other electrically, and capable of communicating
attractive powers to other matter^ — may likewise render their particles attractive, and enable
them to combine when they have full freedom of motion,
Davy did not follow up his ideas about the relations between electrical disturbance
and chemical decomposition, on the theoretical side, but he applied the principle
as an instrument of decomposition, and solved some questions of the very greatest
importance to the growing science, for it led him to the isolation of the alkali metals
— ^potassiimi and sodium.
The idea of acidity involves two concepts — (a) an antagonistic force which
is reciprocated by the alkalies ; and (b) a great tendency to unite with bodies
generally. In 1809, A. Avogadro published a paper entitled, Tdees sur Vacidite et
Valkalinite.^ H. Davy's experiments on electrolysis suggested to A. Avogadro the
idea of a chemical force which is polar at the moment of action, and which not only
determines the union of an acid and alkali, but also chemical changes generally.
According to Avogadro :
All the phenomena are easily explained if we consider acid and alkali antagonism as
purely relative properties, only becoming somewhat absolute when referred to a middle
OXYGEN 399
substance A which has the acid antagonism with reference to B, and which may possess
the alkaline antagonism with reference to a third substance C. What are then termed
acids and alkalies are merely bodies which have the acid or alkali antagonism in
respect of certain other bodies whose position in the scale is approximately indicated by
certain properties, such as inability to affect vegetable blues. . . . The degree of acidity
or alkalinity of a compound depends upon the degree of those properties in its constituents.
. . . Of two substances in the act of combination, one always plays the part of acid, and the
other of alkali ; and it in this» antagonism which constitutes the tendency to combination,
or afifmity properly so called.
Accordingly, said A. Avogadro, different substances can be arranged in series, the
position of each marking its true affinity to any predecessor or successor. Oxygen
and sulphur would come first in the series, hydrogen and carbon last, with the
neutral salts in the middle of the series. The measure of chemical antagonism is
electric heterogeneity or oxygenicity. A substance is the more oxygenic the less it
is oxidizable.
The essence of polarity is the contrast of opposing qualities such as is exhibited
by the so-called north and south poles of a magnet where unhke poles attract and
like poles repel one another ; by the two states of static electricity — ^positive and
negative ; and by the phenomena of electrolysis — as interpreted by C. J. T. von
Grotthus (1805), H. Davy (1806), and M. Faraday (1834) 2— where the atoms or
radicles have opposite polarities to that of the electrode about which they accumulate.
Atoms with opposite polarities combine readily, while those with the same polarity
have little or no tendency for union. " In every part of nature," said K. W. Emerson,
" we meet with polarity. . . . An inevitable dualism besets nature so that each
thing is a half, and suggests another to make it a whole." J. J. Berzelius considered
that " the form of crystalline substances presupposes an effort on the part of the
atoms to touch one another by preference at certain points, and this shows that
the particles probably exhibit an electric or magnetic polarity."
J. J. Berzelius' electrochemical theory (1819-48). — J. J. Berzelius further
expressed his view that the electrical charges on the particles were the ccntrolUng
factors in chemical reactions. BerzeHus' views were described in his Essai sur la
theorie des proportions chimiques et sur V influence chimique de Velectricite (Paris, 1819),
and he took quite a different view from H. Davy as to the way the electrical charges
on the particles produce chemical action. While H. Davy considered the electrical
charges to be the consequence of contact or of mutual action between heterogeneous
particles, Berzelius believed that each elementary atom is endowed with two kinds
of electricity and has in consequence two electrical poles ; these poles differ in
strength so that the resultant effect is to make each atom appear as if it were
positively or negatively electrified. Thus, Berzelius distinguished electropositive
and electronegative elements according to which charge prevailed ; the kind of
charge carried by an element was determined by the appearance of the element at
the positive or negative pole when a compound of" the element was electrolyzed.
The varying degrees of chemical affinity were supposed to imply that different
substances were charged with varying quantities of electricity. Consequently
the elements were arranged in series according to the magnitude of the charge, and
Berzelius thus obtained an electrochemical series with the alkali metals at one end
of the series, and oxygen at the other :
Potassium
Sodium
Zinc
Lead
Gold
Fluorine
Nitrogen
Sulphur
Oxygen
400 INORGANIC AND THEORETICAL CHEMISTRY
Oxygen was supposed to be the most electronegative substance of aU, and was
assumed to be always electronegative. In order to explain the variations of
chemical aflQjiity with temperature, the electric polarity was further supposed to
vary with temperature. To explain why sulphur has a greater affinity for oxygen
than for, say, gold, when both oxygen and sulphur are electronegative, BerzeUus
assumed that the absolute quantity of positive electricity on sulphur is much greater
than on gold, and since elements attract one another by their contrary poles, sulphur
exerts a stronger attraction for oxygen than for gold. Thus, an element might be
positively polar with some elements, and negatively polar with others — sulphur,
for instance, is positive with oxygen, and negative with hydrogen and the metals.
In the case of the caustic alkaUes, too, said J. J. Berzelius, water plays the part of
an acid ; and when it unites with an acid (anhydride), it plays the part of a base.
According to J. J. Berzelius, chemical combination consists in the attraction
of the dissimilar poles of the reacting units, and, in consequence, the neutralization
of opposite electric charges. If opposite electrical charges be exactly balanced,
an electrically neutral compound, chemically inactive, is supposed to be formed.
Berzelius explained double decomposition by his electrochemical theory in these
words :
Every chemical action is an electrical phenomenon dependent upon the electrical polarity
of the particles, and everything which appears to be the result of chemical affinity is really
due to the electrical polarity of some bodies being stronger than that of others. When the
compound AB is decomposed by a substance C, the affinity of C for A is greater than that
of B for A, and C must possess a stronger electrical polarity than B. . . .If two bodies
AB and CD react so as to produce two new bodies AD and BC, it follows that the electric
polarities in the latter pair of bodies are better neutralized than in the former.
These statements may or not may be true, but the argument is not sound, for it
is assumed that because B is displaced by C from its combination with A, the affinity
of C for A is greater than that of B f or A ; it is further assumed that affinity and
electrical polarity are the same, and to state that this proves that the electrical
polarity of C for A is greater than B for A is arguing in a circle.
When, say, sodium unites with oxygen to form the base sodium oxide, Na20 ;
or sulphur with oxygen to form the acid anhydride sulphur trioxide, SO 3, primary
compounds, or compounds of the first order, are formed ; these primary compounds
are made up of atoms having opposite polarities, thus :
NagO cto SO3 CO2
The electrical attractions are not supposed to be always exactly neutralized during
the formation of these primary compounds ; the basic oxides were supposed to have
an excess of positive electricity and the acid anhydrides an excess of negative
electricity ; the excess causes a further attraction between the acidic and basic
radicles resulting in the formation of compounds o£ the second order, for example •
H2O+SO3 NagO+SOg Ca^O+COg KgO+ClgOg
and these again might similarly form compounds o£ higher orders, for example,
said Berzelius, alum must be looked upon as the product of a reaction between
aluminium and potassium sulphates, the former acting as a negative, the latter as
a positive radicle :
K2S04+Al2(S04)3=K2S04.A]2(S04)3
similarly the union of an anhydrous salt with water may be regarded as a combina-
tion of the positively charged anhydrous salt with negatively charged water :
K2S04Al2(S04)3+24H20=K2S04.Al2(S04)3.24H20 •
Present formulae.
Berzelius' formulse.
. Na2S04
NaaO.SOa
. H2SO4
HjO.SOa
. CaCOa
CaO.COg
. KCIO3
K2O.CI2O6
OXYGEN 401
Similarly, potassium chloride, KCl, acts as a base towards platinum tetrachloride,
PtCl4, and these two salts unite to form a compound with a composition correspond-
ing with 2KCl.PtCl4. Berzelius' ideas were embodied in the chemical formulae in
use about 1820. For example :
Sodium sulphate .....
Sulphuric acid .....
Calcium carbonate .....
Potassium chlorate .....
The idea of compounds of different orders is fairly old, for, as previously shown,
Isaac Newton had as clear ideas on this subject as Berzelius. Joseph Black has
told us in his Lectures on the Elements of Chemistry (Edinburgh, 1. 281, 1803) :
The older chemists gave the name mixt to chemical compounds consisting of two ingre-
dients which we have never been able to reduce to simpler ingredients. Particles of a mixt
compound with particles of another mixt formed particles of a compound ; the union
of two compounds formed a decompound ; the union of two decompounds formed a super-
compound ; etc.
Chemists did not long use the terms indicated by J. Black, although there is a
tendency to retain the idea of J. J. Berzelius' compounds of different orders, e.g.
A. Werner (1893) used the idea in his theory of the ammino-compounds.
The main objections to Berzelius' hypothesis are as follows : In the first place,
contrary to Berzelius' assumption that different substances are charged with vary-
ing quantities of electricity, M. Faraday proved that on electrolysis, definite and fixed
quantities of electricity are associated with the atoms of matter, although the atoms
of the same kind of matter in different compounds, on electrolysis, might be charged
with different yet definite quantities of positive or negative electricity. Secondly,
after Avogadro's hypothesis had been established, BerzeHus' theory was thought
to be incompatible with such a comparatively simple reaction as 2H2+02->2H20,
for the compound nature of oxygen is due to different electrical charges on the
component atoms of the molecules ; at first sight this does not agree with the supposed
identity of the resulting two molecules of water. Berzelius accordingly denied the
diatomic nature of the elementary gases. Thirdly, J. B. A. Dumas (1834) showed
that the hydrogen atoms in compounds like CH4 can be replaced one by one by atoms
of chlorine. J. J. Berzelius had postulated that hydrogen is an electropositive
element, and chlorine an electronegative one, as exemplified by hydrogen chloride.
Here in Dumas' substitutions, a negative element can be exchanged for a positive
element without fundamentally altering the chemical character of the resulting
compounds. It would be easy to modify Berzehus' theory to meet these
difficulties.
J. J. BerzeHus entangled his electrochemical theory with other hypotheses as
to the structure of compounds which ultimately brought about its fall ; but the
electrochemical theory should be considered on its merits apart from the subsidiary
hypotheses. " There is life after death in the case of a good doctrine," said
I. Remseninl903 ; "and the spiritual part of the electrochemical hypotheses of Ber-
zelius, so to speak, lives to-day as the doctrine that atoms of the elements carry
electric charges which are the cause of their chemical activity." H. Davy himself
pointed out that his statement of the mode of action of the electrical forces is so
general that half a dozen essentially different schemes might be devised, each in
agreement with the hypothesis that " the forces termed chemical affinity and
electricity," as Faraday expressed it, " are one and the same ; " or that " chemical
affinity is a consequence of the electrical attractions of particles of different kinds
of matter." This hypothesis, developed by H. von Helmholtz (1881), R. Abegg
(1906), J. Stark (1908), and J. J. Thomson (1914), is that which is generally accepted
to-day, and will be described later.
VOL. I. 2d
402 INORGANIC AND THEORETICAL CHEMISTRY
Refebenoes.
1 A. Avogadro, Joum. Phys., 68, 142, 1809 ; E. J. Mills, Phil Mag., (4), 37. 461, 1869.
2 Attempts have been made to read into Heracleitus' doctrine of contraries a foreshadowing
of the doctrine of polarities — F. Lassalle, Die Philosophie Herakleitos' des Dunkeln von Ephesos,
Berlin, 1858.
§ 17. Binary and Unitary Theories of the Constitution of Acids and
Salts
We understand a phenomenon historically when we are clear in our minds concerning
the external conditions and habits of thought from which it sprung ; and when its mainsprings
of action and its purposes, as well as the effects which have proceeded from it, are distinctly-
traceable.— P. Cabus (1892).
The fundamental idea in Lavoisier's system is tlie dualism or polarity of com-
pounds. The acids were regarded as compounds of the acidifiable bases with oxygen,
and salts were compounds of acids with oxygenated compounds of the metals or
radicles. Lavoisier's oxygen theory was accepted by J. J. Berzehus (1815), and by
J. L. Gay Lussac (1816). The former adapted the dualistic theory to his electro-
chemical hypothesis of chemical combination. He said :
Assuming that every chemical compound is solely dependent upon two opposing forces
— positive and negative electricity- — every compound must be composed of two parts held
together by their mutual electrical forces. Hence, every compound body, whatever be
the nimaber of its constituents, can be separated into two parts one of which is positively
and the other negatively electrified. Sodium sulphate, for example, is obtained not from
sulphur, oxygen, and sodiiun, but from sulphuric acid and soda each of which can itself be
separated into positive and negative constituents.
All acids and salts were accordingly supposed to have a binary or dualistic structure.
Berzelius assumed that there is no redistribution of the atoms during the formation
of sulphuric acid by the action of water, H2O, on sulphur trioxide, SO 3 ; and
accordingly he represented the constitution of sulphuric acid by the formula H2O.SO3.
J. J. Berzehus (1815) did not at first accept Davy's demonstration (1810-15) that
some acids — e.g. hydrochloric acid — are free from oxygen, because it disturbed the
uniformity of his dualistic or binary system of chemical combination, but between
1820-5, he abandoned this prejudice, and recognized the existence of non-oxygenated
acids, although he fought for his dualistic or binary hypothesis to the end of his
life. He died on August 7th, 1848.
T. Graham's theory of acids (1833). — In his Ueher die neueren Gegenstdnde der
Chymie (Breslau, 1796), J. B. Kichter described some experiments which pointed
to the generalization : The same constant weight of Lebensluftstcff (oxygen) is
combined with those weights of metal which are required to saturate a constant
amount of acid ; otherwise expressed, the quantities of the various bases required
to saturate a constant amount of acid contain the same weights of oxygen. J. J.
Berzelius substantiated "this discovery of the meritorious investigator[J. B. Richter"
by experiments described in his paper entitled, Versuch die bestimmten und einfachen
Verhdltnisse aufzufinden nach welchen die Bestandtheile der anorganischen Natur mit
einander verhunden sind (1811-2) ; and about 1826, chemists generally held the
opinion that the metal oxides contain one atom of the metal to one atom of oxygen,
and that one molecule of the metal oxide united with one molecule of the acid to
form a molecule of neutral salt. T. Graham, however, proved that this hypothesis
is erroneous, for in his Researches on the arseniates, phosphates, and modifications of
phosphoric acid (1833), he demonstrated the existence of three distinct acid hydrates
of phosphoric oxide, P2O5 — then written PO5. Keeping to the modern notation,
the three hydrates had respectively 3, 2, and 1 molecules of water per molecule
of phosphoric oxide, and corresponded with the formulae :
P2O6.3H2O P2O6.2H2O P2O5.H2O
Orthophosphorlc acid. Pyrophosphoric acid. Metaphoephoric acid.
OXYGEN
403
These hydrates were respectively regarded as *' terphosphate, biphosphate, and
phosphate of water." Graham further showed that the molecules of water in these
three acids could be replaced one by one with basic oxides so that three sodium salts
of orthophosphoric acid, with its three molecules of basic water, are possible ; two
with pyrophosphoric acid, with its two molecules of basic water ; and one with
metaphosphoric acid, which has only one molecule of basic water. He exhibited
the constitution of the phosphoric acids and their salts of soda in tabular form :
Phosphoric acid
Biphosphate of soda
Phosphate of soda .
(.Subphosphate of soda
I Pyrophosphoric acid
Bipyrophosphate of soda
Pyrophosphate of soda
j Metaphosphoric acid
\Metaphosphate of soda
First class
Third class
Oxygen in the
Water.
Soda
Acid
0
3
5
1
2
5
2
1
5
3
0
5
0
2
6
1
1
6
2
0
5
0
1
5
1
0
5
The result of Graham's work was to show (i) that acids may contain the equiva-
lent of n molecules of water which can be replaced by basic radicles to form salts ;
and (ii) the number of molecules of acid required to form a neutral salt is not neces-
sarily equal to the number of molecules of the base, as was supposed to be the case
from the experiments of Richter and Berzelius.
J. von Liebig's theory of polybasic acids (1838). — J. von Liebig followed up
Graham's work, and in a paper entitled, Ueher die Constitution der organischen
Sauren (1838), he adduced examples proving that the molecules of all acids are
not equivalent to one another ; in other words, acids may be mono-, di-, tri-, . . .
basic, according as the acid contains one, two, three, . . . molecules of water which
can be replaced by the corresponding number of molecules of the base. This is
sometimes called Liebig's theory of polybasic acids. Liebig also said when two
and more than two molecules of the base combine with one molecule of the acid,
and only one molecule of water is separated during the operation (that is, fewer
than the number of equivalents of the fixed base), a basic salt is produced. J. von
Liebig proved that the products of decomposition of organic acids and salts are
different under different conditions, and thus demonstrated the fallacy of the then
prevalent assumption that the products of decomposition of a compound prove that
they are present as such in the original compound. Every theory based on processes
of decomposition, said J. von Liebig, is incomplete and insufficient.
Again, J. von Liebig showed that when lime is neutralized with sulphuric or
hydrochloric acid, the same amount of water is formed. According to Berzelius'
dualistic theory, in the one case water was present in the sulphuric acid ready
formed ; and in the other case, the water is produced during the reaction. It
cannot be supposed that there is any essential difference between these two reactions,
because in both it is most probable that the metal of the lime replaces the hydrogen
of the acid, and the hydrogen of the acid combines with the oxygen of the lime to
form water. Hence, there is no difference in kind between the action of the so-
called oxyacids and hydracids. J. von Liebig thus brought the reactions between
most of the acids and bases under one common scheme : Acid+base=salt4- water.
Acids, said J. von Liebig, are particular compounds of hydrogen in which the latter
can be replaced by the metals.
Neutral salts are those compounds of the same class in which the hydrogen is replaced
by an equivalent of the metal. Those substances at present called anhydrous acids acquire
the property of forming salts with metallic oxides, for the most part, only on the addition
of water ; or they are compounds which decompose the oxides at a high temperature. . . .
At ordinary temperatures no salt can be produced without water, and the constitution of
the salts is analogous to that of the hydrogen compounds which we call acids. The principle
of Davy's theory is that the capacity of saturation of an acid is dependent upon the hydrogen,
or upon part of the hydrogen, which it contains, so that if the other elements of the acid are
404 INORGANIC AND THEORETICAL CHEMISTRY
collectively called the radicle, the composition of the radicle does not possess the slightest
influence on this capacity.
On grounds of probability and convenience, J. von Liebig thus advocated a
hydrogen theory of acids analogous to that previously suggested by H. Davy and
P. L. Dulong. To recast the words of J. J. Berzelius, sodium sulphate is a com-
pound of sulphur, oxygen, and sodium, but there is nothing to show that it is com-
posed of a basic oxide, Na20, with an acid anhydride, SO3. Acids are regarded as
combinations of simple or compound radicles with replaceable hydrogen ; and salts
are derivatives of the acids formed by replacing the hydrogen of the latter by metals
or equivalent radicles. This form of the hydrogen theory of acids was opposed by
J. J. BerzeUus to the end of his life (1848), because he saw in it evidence against his
own dualistic view of the composition of acids and salts. J. von Liebig's theory is
strictly dualistic in that the replaceable hydrogen in acids is contrasted with the
acidic radicle.
A. Laurent and C. Grerhardt's unitary hypothesis.— While the hydrogen theory
of acids was developing, Berzelian dualism was fighting a losing battle with organic
chemistry, as described previously ; and it received its severest blow with the
advent of A. Laurent and C. Gerhardt's unitary hypothesis, where it was shown
that the molecules of a compound are to be regarded as simple edifices, as J. B. A.
Dumas expressed it, and not double buildings ; molecules are capable of modifica-
tion by the exchange of one of their constituent elements for another. The unitary
hypothesis denies the existence of separate and opposing components in any
particular compound ; it represents acids and salts by similar formulae to the hydro-
gen theory, but it does not insist upon their containing any definite compound,
radicle, or their being composed according to any particular type ; although it
does place compounds with analogous properties in the same class to express the
influence of each element on the united properties of the compound.
CHAPTER IX
WATER
§ 1. The Cycle of Water in Nature
<^
I am the daughter of Earth and Water
And nursling of the Sky ;
I pass through the pores of the Ocean and Shores ;
I change, but I cannot die. — The Cloud.
Water is widely distributed in nature in its three states of aggregation — steam
or aqueous vapour, liquid water, and solid ice or snow. It has been estimated that
three-fourths of the surface materials on the crust of the earth is water. Animals
and plants contain a large proportion of combined water — e.g. fish contains the equi-
valent of about 80 per cent. ; beef, 60-62 per cent. ; the human body, 70 per cent. ;
aquatic plants between 95 and 99 per cent. ; and ordinary land plants, 50-75 per
cent. A great many rocks contain water — combined and absorbed. Clay, for
example, contains up to 14 per cent, of combined water.
Water plays a vital part in the nutrition of animals and plants ; indeed, it is
absolutely indispensable to animal and vegetable life. It is universally employed
as a solvent, and it is utilized as a thermal agent in refrigeration and heating, where
it is valuable on account of the magnitude of its heat of fusion, and its specific heat ;
it is also similarly used in steam engines. In short, water is employed for countless
purposes by man ; and it is the cause of the most striking phenomena in nature.
The circulation of water in nature — the water cycle. — All the water on the earth
passes through a remarkable cycle of changes. The heat of the sun leads to the
evaporation of water from the seas, etc. ; water vapour is only 0*62 times as heavy
as an equal volume of air, and consequently it rises into the upper regions of the
atmosphere, as well as diffuses into and mixes with the atmospheric air. The
temperature of the ascending vapour gradually decreases, and consequently a plane
must be reached where the air is saturated with moisture. The vapour will then
condense in the form of fine drops of water — mist or clouds. The fine drops coalesce
into larger drops. Ultimately, the condensed water must descend again to the earth
as dew, rain, snow, or hail. The wind distributes the vapour. The heat given up
during the condensation of the vapour is distributed or carried from the hotter
regions — where evaporation is fastest — to the colder regions — where the vapour is
condensed — thus helping to stretch the temperate regions nearer to the poles. The
water which is sprayed, as rain, etc., on the surface of the globe, does a certain
amount of physical and chemical work. On the chemical side, water helps in the
decomposition and weathering of rocks ; and on the physical side, it transports
matter in suspension from the higher to the lower levels. The soluble matters
ultimately collect in the seas.
Thus the water cycle involves : (1) evaporation from the oceans, seas, lakes,
etc. ; (2) condensation in the upper regions of the atmosphere as a fine mist of
distilled water where it collects as clouds ; (3) further condensation followed by
rain ; (4) percolation of the rain-water through the soil and its accumulation on an
impervious bed of rock, whence it is forced to the surface, as spring water, by the
pressure of the superincumbent water ; and (5) the collection of spring and surface
405
406 INORGANIC AND THEORETICAL CHEMISTRY
waters by the streams and rivers to be forwarded to the sea. The river thus returns
the water whence it came to commence anew the never-ending cycle. P. B. Shelley
has described the idea in a charming manner in his well-known poem — The Cloud.
It must be added that a relatively small proportion of the water which finds
its way into the ground is " fixed " by reacting with certain silicates and other
minerals forming hydrated siUcates, hydrated alumino-silicates, etc. — e.g. kaolinite,
Al2O3.2SiO2.2H2O ; selenite, CaS04.2H20 ; etc.
Rain-water. — When rain falls on the surface of the earth, part of it sinks deeply
imderground to reappear as spring water ; about 25 per cent, drains off directly
into streams and rivers ; a part is retained as the ground water of soils ; and a part
returns by evaporation directly to the atmosphere. Rain, in its journey through
the air, dissolves oxygen, nitrogen, carbon dioxide, chlorides, and ammoniacal
and nitrate nitrogen. It also carries down dust — organic and inorganic. Rain-
water, particularly if collected near the sea in high winds, contains sodium chloride ;
and if collected near towns, contains sulphur compounds — sulphur dioxide and
sulphuric acid — derived from the products of combustion of coal. When evaporated
to dryness, 10,000 parts by weight of rain-water will give about 0*34 part of solid
matter ; most of this consists of sodium chloride and organic matter. Rain-water
contains in solution 0'013 per cent, of dissolved nitrogen; 00064 per cent., oxygen;
and 0*0013 per cent., carbon dioxide. The rain which falls at the end of a shower
is less contaminated than that which falls at the beginning, because the atmosphere
is washed, so to speak, during the earlier part of the shower. According to
F. W. Clarke's important work. The Data of Geochemistry (Washington, 1916),
As a carrier of ammonia, nitric acid, sulphuric acid, and chlorine, rain-water performs a
function of the highest significance to agriculture, but whose geological importance has not
been generally recognized. Rain and snow collect these impurities from the atmosphere
in quantities which vary with local conditions, and redistribute them on the soil.
The average of a number of analyses of rain-water at Rothamsted (England) 1 gave
2*71 lbs. of ammoniacal nitrogen per acre per annum, and 1*13 lbs. of nitric nitrogen.
In most cases ammonia is in excess, but in the tropics the reverse seems to obtain.
At Barbados,^ 1*009 lbs. of ammoniacal nitrogen and 2 443 lbs. of nitric nitrogen
were annually deposited per acre. Similarly, 14'87 lbs. of chlorine as chlorides
were deposited annually per acre at Rothamsted, and 180*63 lbs. at Ceylon.
Spring and mineral water. — Directly the rain-water strikes the ground, it
begins to attack and dissolve various rocks, decaying organic tissue (humic com-
pounds), etc., forming surface and ground water. It is estimated that between 25
and 40 per cent, of the rainfall, in temperate regions, soaks into the ground. In
its journey underground the percolating water — underground water — loses most of
its organic matter and dissolves more or less mineral matters — compounds of calcium,
magnesium, and sodium ; carbon dioxide ; etc. The greater the depth to which the
water sinks the greater the amount of solid matter it can dissolve. Water under
great pressure is a powerful solvent. Sooner or later the water which has percolated
underground will be forced to the surface as spring water. If the spring water
holds an unusual amount of some particular constituent in solution which gives
it a marked taste, or some specific property, the term 7nineral water is applied.
Mineral waters do not necessarily contain a large excess of mineral matters in solu-
tion. The water from mineral springs is often named according to some special
constituent dissolved in the water, or from the locality of the spring. Fresh water
is a vague term applied to a natural water which does not contain much dissolved
impurity ; to natural water as distinct from salt or sea water ; etc.
Chalybeate waters contain ferrous carbonate — e.g. Tunbridge ; Buxton ; the Excelsior
Spring, Saratoga, N.Y. ; the Hot Springs of Arkansas ; Homberg ; etc. Sulphur waters
contain hydrogen sulphide and other sulphur compounds, alkaline sulphides, etc., e.g.
Baden ; Carlsbad ; Harrogate ; Bath ; Aachen ; the Red Sulphur Spring, Sharon, N.Y. ;
etc. The water of the Steamboat Springs in Nevada has borates and deposits a sinter
containing arsenic, antimony, mercury, lead and copper sulphides, as well as traces of gold
WATER 407
and silver. Saline waters contain salts of various kinds, for instance, magnesium sulphate
and chloride which give the water a bitter taste — e.g. Bath ; Epsom ; Seidlitz ; Friedrich-
shall i Ofen ; Cheltenham ; etc. Sodium sulphate and sodium carbonate — e.g. Marienbad ;
Carlsbad ; etc. Carbon dioxide {acid reaction) — e.g. Apollinaris (imitations of this and other
mineral waters are made artificially ; the natural water is bottled and exported). Carbon
dioxide with sodium carbonate (alkaline reaction.) — e.g. Vichy; Neuenahr; etc. Carbon
dioxide with sodium chloride- — e.g. Ems ; Neider-Selters ; etc. Sodium and other chlorides
— e.g. Homberg ; Aachen; Baden Baden ; Congress Spring, Saratoga, N.Y. ; etc. Some
waters contain iodine and bromine compounds- — e.g. Congress Spring, and Excelsior Spring,
Saratoga, N.Y. ; Woodhall Spa ; etc. Arsenic — e.g. Roncegno ; Levico ; etc. Lithia — e.g.
Congress Spring, Saratoga. N.Y. ; etc. Boric acid~e.g. Yellowstone Park ; Chaguarama
Valley (Venezuela) ; Tuscany ; etc. Silica- — e.g. the Hot Springs of Iceland, New Zealand,
Yellowstone Park, etc. Hard waters have calcium and magnesiiun carbonates and sulphates
in solution. The waters of some springs, particularly in volcanic districts, issue at an
elevated temperature, hence the term thermal waters, e.g. the Hot Springs of New Zealand
(about 60°) ; Teplitz (39°-49°) ; Vichy (32°) ; San Bernardino, California (40°-78°) ; etc.
For an extensive bibliography on the different natural waters, see C. Doelter, Handbuch
der Mineralchemie, Dresden, 3. i, 889, 1918 ; and H. von Fehling, Neues Handworterbuch
der Chemie, Braunschweig, 11. 745, 1915. For a bibliography of British mineral and
chemical waters, see W. H. Dalton, B. A. Rep,, 859, 1888.
River water. — Spring water collects in rivers and streams. Of tlie total rainfall
on all the land of the globe per annum— 29347 "^ cubic miles — J. Murray 3 estimates
that the rivers of the world discharge about a quarter — or 6524 cubic miles — into
the sea per annum. Kivers contain not only the dissolved and solid matters in
suspension furnished by spring waters, but also organic matter derived from plants
growing on the sides and bottom of the river, and also drainage from the villages
and towns through which the river passes. The river, in virtue of its greater volume
and force, carries along a considerable amount of suspended solids. Eiver water
also contains in solution matter dissolved from the land which it drains, and this
the more the further the river is away from its source. Thus, the Irwell near its
source has about 0*008 per cent, of dissolved solids, and at Manchester nearly 0*056
per cent. The waters of the Dee (Scotland), draining slate and sandstone rocks,
contains about 0*0056 per cent, of solid matter, about one-fourth of this being calcium
salts ; the Thames, draining chalk rocks, contains about 0*03 per cent, of solid
matter, two-thirds of which are calcium salts. F. W. Clarke estimates that
2,735,000,000 tons of solid matter in solution are annually carried to the ocean by
rivers. This does not include suspended matter.
Sea-water. — Just as spring water flows into the rivers, the rivers flow into the
sea carrying their dissolved salts, and suspended matters which have not been
deposited in transit. Consequently, the salts in sea-water have probably been all
derived from the land, and hence it has been said that sea-water holds the debris
of ancient continents in solution. Indeed, attempts have been made to estimate
the age of the sea from the time required for the accumulation of the salt it contains.
For example, it has been estimated that the rivers of the world discharge some
160 million tons of salt into the sea every year, and that the seas hold in solution
some 144 billion tons or 120 million tons per cubic mile, enough to cover the whole
of the present dry land with salt to a depth of 400 ft. Consequently, if these esti-
mates be somewhere near the mark, and if present conditions are not very different
from those which prevailed in former times, it must have taken at least 90 million
years to accumulate the amount of salt now present in the seas.* Estimates of the
volume of the water of the ocean vary from 302,000,000 to 323,722,150 cubic miles.
The vapour which rises from the sea by evaporation is almost pure water ;
hence, unless the dissolved matters are continuously removed, sea-water must be
gradually getting more and more salty. The sea in regions where there is a large
ramfall has less soluble salts than elsewhere. Sea- water contains a relatively large
proportion of soluble salts — the Atlantic contains from 3*301 to 3*737 per cent, of
solids in solution. The composition of the dissolved solids in a number of lakes and seas
is indicated in connection with NaCl. Where the evaporation is greatest we naturally
expect to find the greatest proportion of salts in solution. The water on the surface.
408 INORGANIC AND THEORETICAL CHEMISTRY
for example, usually contains more salt than water deeper down ; similarly, the
Mediterranean contains from 3836 to 4* 115 per cent, and the Indian Ocean from
3553 to 3668 per cent, of solids in solution ; whereas the White Sea contains
2-698 to 2-965 per cent. ; the Black Sea, 1826 to 2*223 per cent. ; and the Baltic,
with its numerous fresh-water tributaries, and less evaporation, contains between
03 and 0*8 per cent, of soUds in solution. Salts accumulate in land-locked or
partially land-locked seas and lakes much faster than in the sea. In illustration,
the Red Sea contains from 5-085 to 5854 per cent, of solids in solution ; Owens
Lake (California), 7-2 per cent. ; the Dead Sea contains 19215 to 25998 per cent. ;
the Great Salt Lake (Utah), 14994 to 23-036 per cent. ; the Caspian Sea, 1267 to
2850 per cent. ; and the Elton Lake (Russia), 265 per cent. These masses of water
behave as if they were exposed in a large evaporating basin, for the salts accumulate
in the water and are deposited in crystalline masses on the shores of the lakes as
the water evaporates. Average spring water contains ten times as much sihca as
sea- water. If all the salts in sea-water are derived from the land, it might be asked :
Where have the silica and calcium salts gone ? The deficiency is said to be
adequately explained by the abstraction of these substances from sea-water by the
marine animals and plants. Diatoms and sponges, for example, use silica to make
their skeletons and shells ; while corals and shell fish use calcium carbonate for
making their skeletons and shells. It has been estimated that a single oyster re-
quires the lime in about 50,000 times its weight of sea-water to make its shell. So
that while fresh supplies of silica and lime salts are being continuously poured into
the sea, the store is being steadily removed.
Potable and drinking water. — It is claimed that the natural waters in particular
localities contain impurities specially favourable to certain industries, and con-
versely in other localities. Hence, " in the brewing of malt liquors," said T. Berg-
man (1778), " the baking of bread, the bleaching of linen, dyeing, the preparing of
hides and skins, and in a number of other arts, the quality of the water employed
is of no small consequence, that unless one be chosen fit for the purpose, the whole
process fails." Potable water — that is, water fit for human consumption — is obtained
principally from rivers and lakes, and also from wells — artesian and otherwise.
The inorganic or mineral matters usually found in solution in natural water are not
directly injurious to health. The purification of water for towns and cities is a
very important practical problem for the chemist. The best process can be adopted
only after a careful study of the local conditions, and the nature of the impurities.
Water should be freed from pathogenic (disease-producing) bacteria, and from
suspended impurities. This is generally done by filtration through large filter beds
made from layers of sand and gravel, extending in some cases over an acre of ground.
In special cases, a Pasteur-Chamberlain's bougie (candle), made of unglazed and
porous earthenware, and shaped like a hollow candle, is arranged to screw on to the
water tap. The water is forced through the earthenware by the pressure of the
main and trickles through the aperture below. Bacteria, organic matter, etc.,
collect on the inside of the bougie as a slimy layer which clogs the filter. The
bougie must be frequently cleaned or replaced (1) to permit the free passage
of water ; and (2) to remove the layer of slimy organic matter which serves as a
medium for the growth of bacteria. In some cases the living organisms in water
are killed by the addition of minute traces of poison — ozone, sodium hypochlorite,
copper sulphate, etc. The salt last named also prevents the growth of green alga?
which are sometimes very troublesome.
To maintain the purity of the water supply up to the proper standard, it is
necessary to make (1) a periodical critical examination of the source from which
the water is obtained ; (2) regular bacteriological examinations for pathogenic
germs ; and (3) chemical examinations for nitrogenous organic matter — albumenoids,
etc. — upon which bacteria feed ; and for the products of bacterial life — free ammonia,
ammonium nitrate, and nitrate. The presence of these substances in water throws
it under suspicion.
WATER
409
References
1 R. Warrington, Journ. Chem. Soc.y 51. 500, 1887 ; 55. 537, 1889 ; N H. J. Miller, Jcmrn.
Agric. Science, 1. 286, 1905.
^ J. B. Harrison and J. Williams, Journ. Amer. Chem. Soc, 19. 1, 1897.
3 J. Murray, Scottish Oeog. Mag., 3. 65, 1888 ; P. W. Clarke, The Data of Geochemistry, Wash-
ington, 1916.
* J. Jolv, Tram. Roy. Dublin Soc, (2), 7. 23, 1899 ; B. A. Rep., 369, 1900 ; Geol. Mag., 344.
504, 1901 ; W. J. Sollas, Journ. Geol. Soc., 65. 41, 1909 ; H. S. Shelton, Science Progress, 9. 55,
1914 ; Chem. News, 99. 253, 1909 ; 102. 75, 1910 ; 112. 85, 1914 ; R. B. Dole, ib., 103. 289,
1911; K. Karsten, Ein£ neue Berechnung der mittleren Tiejen der Oceane, Kiel, 1894;
J. Murray, Scottish Geog. Mag., 3. 39, 1888.
§ 2. The Purification and Distillation o£ Water
Water is purified on a small scale by distillation. The water is boiled in a flask
or boiler, and the steam is condensed back to the liquid condition by passage
through a tube, about which a continual steam of cold water flows. To economize
space, the condensing tube is generally coiled as a spiral — called the worm — and
kept in a tank through which cold water continually flows. It might be added
that the counter-current principle applied to condensers did not originate with
J. von Liebig, but was employed by a French chemist in 1770, by C. Weigel
iu 1771, and by J. Gadolin.i The form of apparatus
sometimes employed in the laboratory with the ordinary
Liebig's condenser for distilling small quantities of
liquid, is easily modified so as to prevent the steam
coming in contact with rubber or cork stoppers, etc.
— nothing but glass. The condenser is fitted to the
distilling flask by ground joint ; rubber or cork
stoppers are not used. Much more compact con-
densing arrangements 2 are available with some of
the more recent vertical condensers. The form illus-
trated in Fig. 1, for example, has been recommended
for preparing water for bacteriological purposes. W^ater
as free as possible from ammonia should be employed ;
river water is therefore objectionable. A small amount
of volatile organic matter if present will be carried
over with the first rush of steam, and soluble matters
derived from the glass receiver and condenser may be
found in the distillate. Tubes of quartz glass, block-
tin, or silver for the condenser are better than glass,
since the water acts very much more vigorously on glass
than it does on quartz, tin, or silver. Distilled water
which has been kept some time in a glass bottle cannot be used satisfactorily in the
analysis of silicates, because it contaminates the silicate undergoing analysis with some
of the constituents to be determined. In very special cases silver, gold, and platinum
vessels have to be used as condensers and receivers. J. S. Stas 3 obtained water
free from volatile organic matter by mixing 4 or 5 per cent, of potassium permanganate
or manganate with the water in the boiler. Some potassium hydroxide was also added
to keep the solution tres alcalin. The distillate was mixed with aluminium sulphate,
or with potassium or sodium hydrogen sulphate, and again distilled to eliminate
ammonia, etc. G. A. Hulett recommended distilling the water twice — once after
the addition of sulphuric acid and potassium permanganate ; and the second time,
after the addition of baryta water so as to get rid of carbon dioxide. According
to W. E. Bousfield,* potassium hydrogen sulphate gives very satisfactory results
in keeping back ammonia and traces of basic impurities : and according to
Fig.
1. — Distillation Ap-
paratus
410 INORGANIC AND THEORETICAL CHEMISTRY
R. Bourdillon, phosphoric acid acts very much the same — about 0*5 per cent, of
potassium hydrogen sulphate will suffice.
The purest water so far prepared has an electrical conductivity of 0043x10"^
reciprocal ohms at 18°, and for ordinary conductivity experiments, conductivity
water — water with a conductivity of 10 "^ to 3 X 10 ~^ reciprocal ohms — is considered
satisfactory. The former can be prepared only by distillation in vacuo, and it cannot
be kept without absorbing impurities — carbon dioxide, ammonia, etc. — from the
air, and these quickly increase the conductivity of the water. Thus, F. Kohlrausch
and A. Heydweiller ^ found that a freshly prepared sample of water distilled in
vacuo had a specific conductivity of 005 X 10"^ to Oil X 10"^ ; and after prolonged
exposure to air, 0*66 XlO-^. Water which has been distilled in air always contains
carbon dioxide, and F. Kohlrausch says that such water can be freed from much of
this gas by passing through it a current of air which has been scrubbed in a tower of
soda-lime. It is now usual to pass a stream of scrubbed air through the condenser
in which the water is being condensed. The flask or bottle in which the conductivity
water is stored should be made of good Jena glass ; it should be fitted with a
paraffined cork ; it should be arranged with a stoppered syphon for draining ofi the
water as required ; and the air entering the bottle as the water is removed should
pass through a tube packed with soda-lime.
References.
1 R. A., Parfumerie moderney 12. 10, 1919.
2 F. Mvlius and F. Forster, Ber., 24. 1482, 1891 ; H. C. Jones and E. Mackay, Zeit. phys.
Chem., 22. 237, 1897 ; W. Marek, Journ. prakt. Chem, (2), 60. 681, 1899.
» J. S. Stas, Mem. Acad. Belgique, 35. 1, 1865 ; (Emres, Bruxelles, 1. 100, 536, 1894 ; Chem.
News, 4. 207, 1861 ; 15. 204, 1867 ; G. A. Hulett, Zeit. phys. Chem., 21. 297, 1896 ; A. A. Noyes
and W. D. Coolidge, Proc. Amer. Acad., 39. 190, 190S ; J. Kendall, Journ. Amer. Chem. Soc, 38.
2460, 1916 ; H. J. Weiland, ib., 40. 131, 1918.
* W. R. Bousfield, Journ. Chem. Soc, 87. 740, 1905; 101. 1443, 1912; R. Bourdillon, ib.,
103. 791, 1913 ; J. Walker and W. Cormaek, ib., 77. 5, 1902 ; H. Hartley, N. P. Campbell, and
R. H. Poole, ib., 93. 428, 1908; C. B. Clevenger, Journ. Ind. Eng. Chem., 11. 964, 1919.
5 F. Kohlrausch and A. Heydweiller, Zeit. phys. Chem., 14. 317, 1894; 42. 193, 1902;
F. Kohlrausch, Wied. Ann., 44. 583, 1891 ; Pogg. Ann. Erg., 8. 1, 1876.
§ 3, The Effect o! Temperature and Pressure on the Volume of Water
Not only in the matter of solutions, but in other more strictly physical relations, it
is a misfortune that the r6lp of a typical liquid was assigned to water.^ — G. F. Stradlinq
(1901).
Although at ordinary temperatures water is a clear limpid liquid, it forms a
crystalline solid — ice — at temperatures below 0°, under atmospheric pressures, and
it forms a gas — steam — at temperatures exceeding 100°. In 1803, J. Southern i made
some measurements to determine how much water was required to furnish one cubic
foot of steam at various pressures. A litre of liquid at 100° occupies 1696 litres when
it is changed to saturated vapour ; and at 0°, a litre of liquid forms 205,093 litres of
saturated vapour. The great expansion experienced when water passes into steam
has been suggested as a substitute for gunpowder for discharging projectiles ; and
in 1824, J. Perkins constructed a steam gun which gave results rivalling those
obtained with gunpowder.
A. Winkelmann measured the relative density of water vapour, standing in
equihbrium with the liquid, when the density of air is taken as unity :
Pressure
. 0-5
1-0
2-0
3-0
4-0 atm.
Temperature
. 81-7°
1000°
120-6°
133-90°
144-0°
Density
0-63357
0-64026
0-64838
0-65400
0-65860
The theoretical density on theassumption that the molecular weight of water is 18'02
is 0*6224 ; the observed densities agree with the assumption that the molecular
WATER 411
weight of water is between 18'33 and 19"06. This means that a small fraction of
the molecules in water vapour are polymerized.
0. Knoblauch, E. Linde, and H. Klebe 2 have calculated the specific volume of
water vapour, if it obeyed the ideal gas laws, to be
^ 18016 p
and the percentage deviations from the calculated (vq) and observed {v) values,
when 2^ is constant, are
Temperatut
■e . 100°
110°
120°
130°
140°
150°
160°
170°
180'
v-v. .
1-6
2-0
2-5
31
3-5
4-5
5-1
6-0
6-8
The ice and water molecule theory ol W. C. Rbntgen.— The physical properties
of water differ widely from those of most liquids, and the list of anomalous properties
is a long one. In order to explain these, W. C. Rontgen (1891) ^ assumed that
water is a mixture of two kinds of molecules which he called ice molecules and water
molecules. The ice molecules were supposed to form a mass more complex but less
dense than water molecules. Similarly, in I. Traube's theory of liquids, the existence
of what he called gasogenic molecules and Uquidogenic molecules is postulated. The
former would be represented by steam molecules, and the latter by water molecules.
If (H.20)n represents the ice molecules, (H20)^ the water molecules, and H2O the
steam molecules, then
m(H20)n^w(H20)^^(m+w)H20
Ice mols. Water mols. Steam mols.
Increase of temperature— > <— Decrease of temperature
Increase of pressure— > <— Decrease of pressure
The diminution in volume which occurs when ice changes into liquid water corre-
sponds with the passage of ice into water molecules, because the ice molecules have
the less density and occupy a greater volume than the sum of the volumes of the
corresponding water molecules.
The following hypotheses enable most of the so-called abnormal properties of
water to be explained : (1) Liquid water contains at least two kinds of molecules —
respectively called ice molecules and water molecules ; (2) Low temperatures favour
the accumulation of ice molecules ; (3) High pressures favour the accumulation of
water molecules ; pressure dissociates the more complex molecules ; and the
greater the pressure, the less the proportion of ice molecules in the liquid ; (4) The
passage of ice into water molecules is accompanied by a contraction, and conversely ;
(5) The passage of ice into water molecules absorbs heat and is therefore said to be
an endothermal reaction, and conversely for the reverse change ; (6) In spite of the
heterogeneous character of liquid water, this liquid freezes at a constant temperature
because there is a definite equilibrium concentration for each molecular species at
each temperature ; and (7) The proportion of ice molecules in the liquid is reduced
when a salt is dissolved therein. The anomalous properties of water are thus referred
to the presence of special kinds of molecules in liquid and solid water ; these mole-
cules, too, are invested with special properties to fit the facts. This is a dangerous
method of investigation, but there is an overwhelming mass of evidence to justify
the procedure.
There is a definite equilibrium concentration between the ice molecules and the
water molecules at each temperature, and C. S. Hudson * represents this by the curve
0, Fig. 2 ; the solubility of the ice molecules will also vary with temperature, as
represented by the curve S\ Fig. 2. Above the freezing temperature, the solu-
bility of the ice molecules in the water molecules is greater than the equilibrium
concentration and the solution is not saturated ; at temperatures below the freezing
point the liquid is supersaturated with ice molecules and the water can freeze ; at
412 INORGANIC AND THEORETICAL CHEMISTRY
the freezing point, the equilibrium concentration of the ice molecules is equal to
their solubiUty. The addition of soluble salts to the liquid lowers the eqmHbrium
concentration and solubihty of the ice mole-
cules, and thus lowers the maximum density
and freezing point of the liquid.
^ 1^ ,A\^X ^® thermal expansion of water. — It was
known to the Arabian writer Al-Khazini in
the twelfth century that water contracted when
cooled, and expanded when heated, and that
ice was specifically lighter than water. His
measurements were so accurate that they are
quite in accord with those adopted at the
0- present day. He found ice to have a specific
Temperature gravity of O'OGS, and hot water 0958 when
Fig. 2. — ^Equilibrium Concentration water at ordinary temperatures is taken as
and Solubility of Ice Molecules at ^^^ty G. Galilei also discussed the expansion
m^l^""* Temperatures (Diagram- ^^^ ^^^^ ^^^^ ^^^^^ ^3 j^^^^^^^ ^^^ the
™* ^ contraction which occurs when water is cooled.
The coefficient of thermal expansion a, or (dvldt)plv—t]ia,t is, the increase in
volume per unit volume per degree rise of temperature at a constant pressure —
of water, and of a few other liquids for comparison, is
Water.
. 0-00048
Mercury. Sulphuric acid. Alcohol.
0-00018 0-00063 0-00113
Ether.
000155
Benzene.
0-00125
CSa.
0-00119
The coefficient increases with a rise of temperature. J. Meyer found for water.
20°
0-000110
40°
0-000217
60°
0-000305
100°
0-000452
140°
0-000584
180°
0-000722
200°
0-000788
5'
«
? 4
These numbers agree with the values of G. A. Him, and G. Tammann and
K. Zepernick between 100° and 143°.
The observed expansion of water
with rise of temperature is assumed
to be the joint effect of (a) a de-
crease in volume due to the change
of ice into water molecules ; and
(b) the normal expansion due to
the further separation of the mole-
cules. The higher the temperature
the less the proportion of ice mole-
cules to be changed into water
molecules, and the thermal expansion
becomes proportionally greater.
Excluding water, the coefficients
_^___^__^^_______^_^^___^_^ of thermal expansion of about eleven
100 200 300 400 500 600 700 800 900 Hquids — including alcohol, carbon
Fig. 3.— Coefficients of Thermal Expansion of sulphide, etc. — which have been tried ,
Water at Different Temperatures and Pressures, diminish the higher the pressure. For
example, E. H. Amagat (1893)5
found the following values for the coefficient of expansion of ether ( X 10^) at
difEerent temperatures and pressures :
Pressure (atmospheres)
axl0« fromO° to 20°
axlO« from 138° to 198°
With water, E. H. Amagat found that the coefficient of thermal expansion
at a constant temperature for temperatures up to 50° increases with increasing
pressures as indicated in the first five vertical columns of Table I ; about 50°,
-^
^0
-/oo"
^
■--20
Z80»
-
__
—
^50
—
3a
E5£
\0
.ji
—
'z:^
--^
0
"li-
^
50
100
200
600
800
1000
1511
1445
1319
1045
958
900
^-
—
2156
1165
1008
890
WATER
413
the reverse obtains, for the coefficient of expansion decreases with an increase of
pressure, and water then behaves like other liquids which have been tried. Conse-
quently, water behaves like a normal liquid at temperatures exceeding 50°. E. H.
Amagat's and F. Auerbach's values for the effect of
pressure on the mean coefficient of thermal ex-
pansion of water are illustrated by Fig. 3. The
horizontal lines, Table I, represent the coefficient
of thermal expansion at different temperatures
when the pressure is constant. Up to a pressure of
3000 atm., the coefficient steadily increases as the
temperature rises, no matter what be the pressure
provided it be constant. E. H. Amagat's and G.
Tammann's results for the effect of pressure on the
volume of water at different temperatures are
illustrated in Fig. 4 — volumes are denoted by
ordinates, temperatures by abscissae
J. H. Vincent found the linear coefficient of
expansion of ice between —10° and 0° to be
0-0000507 ; and J. Dewar found the mean coefficient
of expansion of ice between 0° and —188*7° to be
0*00008099 — about a quarter of the value between 0°
and 10°, and half the value between 4° and 100°.
The Florentine Academicians discovered in 1670, 'oo'o
that there is a certain temperature at which liquid loooo
water possesses a maximum density, but the tem-
perature at which water acquires this state was ^^l^'-^^^Yo^tf^l^:^
not determmed with any degree oi precision until Different Temperatures.
J. A. Deluc 6 noticed that the anomaly obtained
when water is used as a liquid in thermometers corresponds with 5° as the
temperature of maximum density. Count Rumford (1805), J. G. Tralles
(1807), G. C. Hallstrom (1827), and C. M. Despretz (1836) made careful
measurements of this constant, and the more recent work of P. Chappius (1897)
and L. C. de Coppet (1904) gives numbers ranging from 3*980° to 3*983° for the
temperature of maximum density of water; while, according to J. D. van der
Waals, the temperature of maximum density is 4*18° in vacuo, and 4*08° under
the normal pressure of one atmosphere. E. H. Amagat found the temperature of
maximum density was lowered to 3*3° by a pressure of 41*6 atm., to 2*0° by a
pressure of 93*3 atm., and to 0*6° by a pressure of 144*9° atm. According to.
S. Lussana, the temperature of maximum density is lowered to 4*10—0*0225(^—1),
by a pressure of ^ atm.
10170
10150
10130
•Olio
10090
I 0070
1-0050
0030
Table I.
-The Coefficient of Thermal Expansion of Water at Pressures ranging
FROM 1 to 3000 Atm.
si
2^
*
Mean coefficient of thermal expansion x 10«.
^"S
0°-10**.
10°-20°.
20°-30«.
30»-40°.
40°-50°.
50»-60».
60»-70°.
70»-80».
soo-oo".
OO'-lOO".
1
14
149
267
334
422
490
556
100
43
165
265
345
422
485
548
^_
. —
—
200
72
183
276
350
426
480
539
600
.
400
125
221
298
363
429
478
627
575
626
673
600
169
250.
319
372
429
484
620
557
605
650
800
213
272
339
378
439
480
518
546
595
630
1000
259
293
343
396
437
474
512
554
681
610
2000
364
356
416
423
469
. —
.
. —
. —
. —
3000
391
420
433
440
469
■ —
' —
• — •
— •
" —
414
INORGANIC AND THEORETICAL CHEMISTRY
The water molecules en masse have a relatively greater density than ice mole-
cules, and hence occupy a smaller volume. During the melting of ice, however,
only a certain fraction of the ice molecules are changed into water molecules, and
the resulting liquid is a solution of ice molecules in water molecules. Any further
application of heat results in (a) a decrease in volume arising from the transformation
of the ice into water molecules ; and (6) an- increase in volume due to the joint
thermal expansion of both the ice and water molecules. The observed effect is the
difference between these two opposite effects. In passing from 0° to 4°, the ex-
pansion due to the thermal expansion is masked by the contraction due to the break-
ing down of a definite proportion of the ice molecules ; from 4° upwards, thermal
expansion overbalances the contraction due to the changing molecules ; while at
about 4°, the two effects just balance one another. The temperature of the maximum
density of water is reduced by increasing the pressure, because the proportion of
ice molecules in water is reduced.
C. M. Despretz's study of the effect of salt solutions on the temperature of
maximum density led him to the conclusion that the lowering of the temperature of
the point ofTnaximum density of water caused hy the addition of a soluble salt is directly
proportional to the concentration of the solute. F. Rosetti did not succeed in finding
any definite relation between the lowering of the temperature of maximum density
and the lowering of the freezing point produced by a dissolved salt, for although a
definite ratio was obtained for a given solute at different concentrations, a different
ratio was obtained with a second solute. Consequently, while the lowering of the
freezing point depends only on the concentration of the solute, the lowering of the
temperature of maximum density depends on the nature as well as on the concentration
of the solute. L. C. de Coppet noticed that for solutions of salts of a given family
of metals, the lowering of the temperature of maximum density is sensibly the same
for a given acid radicle. Each acidic and each basic radicle produces its own effect,
and the joint effect is the sum of the separate effects. R. Wright has established a
similar rule for the salts of the dibasic acids, and the salts of the alkaline earths.
Table II
• — The Volume
OF Water between —
10° AND
320° (Volume
AT 4° Unity).
Volume occupied at e° by one c.c. at 4°.
Tempera-
ture.
0
1
2
3
4
5
6
7
8
9
-1
1-00019
0
0013
0021
0031
0042
0055
0070
0088
0108
0131
0157
0
0013
0007
0003
0001
0000
0001
0003
0007
0012
0019
1
0027
0037
0048
0060
0073
0087
0103
0102
0104
0157
2
0177
0189
0221
0244
0268
0294
0320
0347
0375
0404
3
0435
0466
0497
0530
0563
0598
0633
0669
0706
0743
4
0782
0821
0861
0901
0943
0985
1028
1072
1116
1162
5
1207
1254
1301
1349
1398
1448
1498
1548
1600
1652
6
1705
1758
1813
1867
1923
1979
2963
2093
2152
2210
7
2270
2330
2390
2452
2514
2576
2639
2703
2768
2833
8
2899
2965
3032
3099
3168
3237
3306
3376
2447
2518
9
3590
3663
3737
3810
3884
3959
4035
4111
4188
4265
10X1
0433
0505
0601
0693
0794
0902
1019
1145
1279
1429
10X2
1590
177
195
215
236
259
283
308
34
38
10X3
42
46
51
' — '
—
"~-
' — ■
■""■
— ~
—
The variation of the volume of water with temperature has been the subject of
researches extending from the beginning of the nineteenth century up to the present
time. The more recent determinations are due to K. Scheel (1892),^ W, Kreitling
(1892), P. Chappius (1897), and to M. Thiesen, K. Scheel, and H. Dieselhorst (1900)
of the Physikalisch Technischen Reichsanstalt. The results obtained by the latter,
between 0° and 40°, are indicated in Tables II and III. The numbers in the sixth
WATER
415
and seventh decimal places are deleted, and the fifth decimal raised one unit when
the deleted figures exceed 50. The data from 40° to 100° are by M. Thiesen (1904) ;
those from 0° to —10° are mean values of data by J. J. Pierre (1845), H. Weidner
(1866), and F. Rosetti (1871) ; and those from 100° to 320° are mean values from
measurements by W. Ramsay and S. Young (1893), J. J. Waterston (1863), and
G. A. Hirn (1867).
Table III. — The Specific Gravity of Water between —10° and 320'
Gravity at 4° Unity).
(Specific
Specific gravity at e"
when the specific gravity at 4* is unity.
Tempera-
ture
d.
0
1
2
3
4
6
6
7
«
9
-1
0-99985
-0
9987
9979
9970
9958
9945
9930
9912
9892
9896
9843
0
9987
9993
9997
9999
. —
9999
9997
9993
9988
9981
1
9973
9963
9952
9940
9927
9913
9897
9880
9880
9843
2
9823
9802
9780
9752
9752
9707
9681
9654
9626
9597
3
9567
9537
9505
9473
9440
9406
9371
9335
9299
9262
4
9224
9186
9147
9107
9066
9025
8982
8940
8896
8852
5
8807
8762
8715
8669
8621
8573
8525
8475
8425
8375
6
8324
8272
8220
8167
8113
8059
8005
7950
7894
7830
7
7781
7781
7723
7666
7607
7489
7429
7368
7307
7245
8
7183
7121
7057
6994
6930
6865
6800
6734
6668
6601
9
6534
6467
6399
6330
6261
6192
6122
6051
5981
5909
10X1
9585
9510
9434
9352
9264
9173
9075
8973
8866
8750
10X2
6828
850
837
823
809
794
779
765
75
72
10X3
70
68
66
■ —
" — •
• —
■ — •
• — •
" — •
' —
The volume v of water for a temperature 6 between 0° and 33° can be repre-
sented by K. Scheel's formula v=Vo (l-O-46427<9+O'O585O5302_o.O7678986/3
+0'0g50024^4). The coefficient of expansion of ice is 0"04375. J. Duclaux as-
sumed that the expansion of water for temperatures between —10° and 150° is a
composite effect of two opposing forces : (1) the relation between the temperature
and the expansion of a liquid constituted of simple molecules which can be repre-
sented by a parabolic formula a-{-bd-\-cd^ ; and (ii) the increase in the volume due
to the polymerization of the constituent molecules whereby the volume of the liquid
becomes specifically greater as the number of polymerized molecules increases
on the falling temperature. Assuming that tTie expansion is almost proportional
to the number of polymerized molecules, it can be represented by Kn/T, where
K and n are constant, and T is the absolute temperature. Consequently, the ob-
served expansion is the sum of the two separate effects. The constants a, &, c, n,
and Jc can be evaluated from M. Thiesen, K. Scheel, and H. Dieselhorst's data,
so that the volume v of the water at 6°, when J'=2734-^, is
v=O-991833+O-OOO2252O8^+O'OOOOO28447502_|.
0-061695711
"273+^
A. Hess found the specific volume of ice at 0° to be 1*0236 ; and J. Dewar calculates
that at absolute zero, the specific volume of ice will be 0"9584.
M. Thiesen (1904) represented the relation between the specific gravity D of
water and the temperature 6, between 25° and 100°, by the empirical formula :
j^_.,_(^-3'982)2 (0-1-273) (350-
466,700
•6)
{9+67) (365-^)
while M. Thiesen, K. Scheel, and H. Dieselhorst have modified an older formula
416 INORGANIC AND THEORETICAL CHEMISTRY
of D. I. MendelcefE, and have represented the specific gravity of water by the
expression :
(^-4)2
118,932+1366-75^-4-13^2
for temperatures between 0° and 30°. If the water contains air in solution, its
specific gravity is reduced ; but above 20°, the effect is negligibly small ; the differ-
ence rises steadily from 0-0000025 at 0° to a maximum 00000034 at 8°, and then
steadily falls to 0*0000004 at 20°.
Both H. Kopp and I. Traube » have shown that the molecular volumes of liquids
— that is, the molecular weights divided by the respective specific gravities — are
additive properties in that they can be represented as the sum of the molecular
volumes of their components, and the results computed on this assumption agree
with the observed values. Water is exceptional in that the observed molecular
volume is larger than the computed result ; it is therefore inferred that the mole-
cular weight is larger than that represented by the simple formula H2O, and
I. Traube has shown that a molecule corresponding with (H20)3 or HgOa gives a better
agreement between theory and observation. According to D, Berth elot, the
molecular v of a liquid at the absolute temperature T is related with the critical
temperature Tc and critical pressure Pg (atm.) by the expression ?;=ir4 TJPci^Tc—T).
When applied to water at 16°, the molecular volume appears to be 25 in
place of 18 ; hence, the molecule is more condensed than corresponds with the
formula H2O.
The curve. Fig. 7, was obtained by plotting the volume of a given mass of
water at different temperatures ; it shows that water above 4°, like most liquids,
expands when heated and contracts when cooled ; but for temperatures below 4°
the curve is abnormal, for the water expands when cooled, and contracts when
heated. If the specific gravity of water at 4° be taken as unity, it follows that
water becomes specifically lighter when the temperature is raised or lowered
beyond this point. The temperature of maximum density of water, 4°, is often
taken as a standard, or unit of reference for specific gravity, etc.
The expansion of water when cooled from 4° to 0° is very small, but that
minute quantity has a very important bearing in nature. When the water on the
surface of, say, a lake is cooled, it contracts. The heavier cold water sinks, and the
warm water rises. This circulation cools the temperature of the whole body of
water down to 4° ; any further cooling results in the formation of specifically
lighter water. Accordingly, this remains on the surface, and circulation ceases.
Finally, as a result of this remarkable, and abnormal property, when the temperature
of the atmosphere falls to 0°, a surface film of ice is formed. Ground ice or anchor
ice may be formed at the bottom of the more shallow rapidly moving streams when
the cooling water is thoroughly mixed, and not allowed to settle in layers. If the
water did not expand as the temperature falls to 0°, the whole body of water would
freeze from below upwards and produce profound climatic changes, since the larger
amount of ice formed in winter would materially affect the temperature for the rest
of the year. The remarks do not apply to salt (sea) water which contracts as the
temperature is lowered down to the freezing point, but sea water shows a tempera-
ture of maximum density at —321°. Indeed, C. M. Despretz (1839) showed that
the temperature of maximum density of salt (sodium chloride) solutions is lowered
almost proportionally with the amount of salt in solution ; thus.
Per cent, sodium chloride . . .0 0*05 O'l 0-4 0-8
Temperature maximum density . . 4° 3° 1-8° —6*6° —16-6°
Similar results have been obtained with other salts, and S. Lussana (1895) has
shown that the effect of an increasing pressure is to still further reduce the tempera-
ture of maximum density. Thus, with solutions containing 0'5 per cent, of sodium
chloride, the temperature of maximum density is 335— 00177(^— 1), where p
WATER
417
denotes the pressure in atmospheres ; and with 1'44 per cent, solutions of the same
salt, the temperature of maximum density is 0*77— 0'011(^— 1).
In the act of freezing, water expands so that 100 c.c. of liquid water at 0° gives
approximately 110 c.c. of ice at the same temperature. In 1665, Robert Boyle^®
found the specific gravity of ice to be 0*903 because he observed that the volume
of water on freezing expanded 11 12 per cent. This result has no pretension to
exactitude because of the uncorrected errors due (i) to strains in the containing
glass during freezing ; (ii) to the probable presence of minute cracks in the arti-
ficially frozen ice ; and (iii) to the presence of varying amounts of dissolved eases.
Similar remarks apply to the early determinations indicated by L. Playfair and
J. P. Joule, who found a mean value of 0*9184. A more accurate determination was
made by C. Brunner (1845), who found the specific gravity of ice at 0° is 0*9180
±0-000039 ; G. Duvernoy gave 0*922 ; R. Bunsen (1870) gave 0*9165 ±0*00003 ;
A. Leduc (1906), 0*9176. J. Pliicker obtained 0*91580 ±0*000008 ; H. Kopp's not
very exact value (1855) is 0*907 ± 0*0007 ; L. Dufour's value (1860) is 0*9178 ± 0*0005 ;
J. von Zakrzewsky's (1892), 0*916710 at —0*7°. E. L. Nichols (1899) emphasized
the difference of the order of 1 or 2 parts in 1000 in the results obtained between
artificial ice 0*91615 and natural ice 0*91807. This does not mean that the two
forms of ice are different, but that the differences are due to strains set up in
the artificial ice in the act of freezing ; E. L. Nichols believes that these disappear
in time, and the specific gravity then attains
its final value. In 1901, H. T. Barnes and
H. L. Cooke found the specific gravity of
natural ice from the St. Lawrence river to
be 0*91661 ±0*00007; J. H. Vincent (1902)
obtained 0*9160. According to J. Dewar,
the specific gravity of ice at —188*7° is
0*92999 ; and he calculates that at absolute
zero, the specific gravity would be 0*9368 so
that ice can never be cooled until it has the
same specific gravity as water has at 100°.
The specific gravity of ice at 0° varies with
its mode of formation from 0*9159 to 0*9182 ;
the specific gravity of water at 0° is 0*999867.
Accordingly, ice floats on the surface of water.
The expansion of water during freezing is an
important factor. The expansion may burst the intercellular tissue of plants by
freezing the cell-sap ; the expansion may disrupt the fibres of flesh, so that the frozen
meat appears rather more pulpy than ordinary meat. If water freezes in pipes, the
expansion of water in the act of freezing may burst the pipe, and water will leak
when the ice thaws ; water freezing in the surface crevices of rocks, splits and widens
the fissures so that the surface crust of the rock appears to disintegrate during a
thaw. The debris collects as talus at the foot of the rocks, ready to be transported
by water to lower levels. Hence this simple force plays an important part in the
weathering and decay of rocks, building stones, etc., in countries exposed to
alternate frost and thaw ; and, adds J. Tyndall : " The records of geology are
mainly the history of the work of water."
In an old experiment of Rumford's usually cited to illustrate the low thermal
conductivity of water a piece of ice was weighted to keep it at the bottom of a
cylinder of water. It was then possible to boil the water by heating the cylinder
near the top, and this without melting the ice. The experiment further illustrates
what would happen if water did not exhibit the anomalous expansion on freezing.
Instead of the ice being buoyed up to the surface it would sink to the bottom of
lakes, etc. The warmer water would remain on the surface in summer so that the
ice would increase in winter and persist in summer until a great portion of the
water on the surface of the earth would be permanently frozen. As it is, the
VOL. I. ■ 2 E
i-OOZO
10016
<j
(j
I :: ::
■
hOOlZ
1
i^
-^n
tffipi
fH
hoooa
1-0004
j: :::::
:::::.'
;ii::::::
!; ;::|;;;
1-0000
m
,:M
: : : : Tern
oerature ::
Fig. 5. — Relations between the Voliune
and Temperature of Water.
418
INORGANIC AND THEORETICAL CHEMISTRY
temperature at the bottom of a body of fresh water cannot get below the temperature
of maximum density, for if the water be cooled further it rises to the surface and
is there frozen. The ice thus hinders the further cooling of the water which remains
in the liquid state.
The change in volume of water with increasing temperatures above 40° is
represented by a curve concave towards the temperature axis, while with salt
solutions this curve is flatter, approximating to a straight line with highly con-
centrated solutions. The curves for solution and solvent thus cross each other as
illustrated in Fig. 6. According to P. de Heen,ii foj. lithium, sodium, potas-
sium, and ammonium chlorides the points of intersection are respectively 30°, 55°, 50°,
and 35° ; for calcium, barium, magnesium, and aluminium chlorides, respectively
45°, 50°, 35°, and 37° ; for sodium and potassium sulphates, 60° ; for sodium and
potassium carbonates, respectively 67° and 65° ; and for potassium and ammonium
nitrates, 70°.
The compressibility of water. — In 1620, Francis Bacon subjected water to
pressure, by squeezing it in a lead sphere ; the water passed through the pores of
the metal. The Florentine Academicians tried a similar experiment with a gold
vessel in 1667, and came to the conclusion that water is incompressible. J. Can-
ton,i2 in 1762, was probably the first to establish the compressibility of rain-water,
sea- water, mercury, spirit of wine, and oil of cloves ; and, in 1764, he showed
that the compressibility decreases as the tem-
perature is raised. J. Perkins, in 1826, also
showed that the compressibility decreases as the
pressure is increased — quickly at first, and after-
wards more slowly. Experiments on this subject
were made by H. C. Oersted (1822), D. CoUadon
(1827), and by H. V. Regnault (1848), and have
been continued from that time to the present day.
The compressibility of a liquid is the fractional
change in volume, dv/v, which occurs per unit
change of pressure. The coefficient of compressi-
bility, P=—dvlvdp, is numerically equal to the
decrease in volume per unit volume produced by
unit change of pressure. The average compressi-
pressure which occurs on the application of p units of
where Vq is the initial volume, and v the volume at
/
(f>A
^
r'^
\ /
/
.^
4i
k<
/
y
/^'
^
Temperature
Fig. 6. —Diagrammatic Representa-
tion of the Thermal Expansion
of Water and Salt Solutions.
bility per unit of
pressure is {vo—v)lpvQ,
a pressure p. P. G. Tait found for fresh water the empirical formula : Average
compressibility=0-28/(36+P)(150+^). Some results are shown in Table IV
where the pressures are measured in tons per sq. in., showing that the com-
pressibility decreases : (a) with an increase of pressure : (6) with a rise of
temperature ; and (c) when the water has salts in solution — the ratio of the
compressibility of sea water and fresh water is nearly 0*92. The coefficient of
compressibility of water per atmosphere, near 0°, is given in the following table with
that of a few other liquids for comparison :
Water.
0-000050
Mercury.
0-0000038
Alcohol.
0-000093
Ether.
0-000164
Benzene. Carbon disulphide.
0-000085 0-000078
T. W. Richards and his co-workers find a compressibility of 42-1x10-6 at 20°
between 100 and 500 megabars ; 43-3x10-6 between 100 and 300 megabars ;
and 40-9x10-6 between 300 and 500 megabars. G. A. Hulett and T. Peczalsky
have also measured the compressibility of water. Liquids in general are but
slightly compressible. The volume of water is reduced but 0-00005th part of its
volume per atmosphere pressure at 0°. According to P. G. Tait,i3 this small com-
pressibility means that if sea -water were quite incompressible the average level of
the sea-water would be raised 116 feet higher than it is to-day, and about 4 per cent,
of the present land surface would be submerged. The compressibility of liquids,
WATER
419
including water, decreases with rising pressure, temperature constant ; because
the compressibility decreases faster than the volume. The compressibility of liquids
other than water increases with rise of temperature, that is, the thermal expansion
Table IV. — Compressibility of Water.
TemperaUire.
Fresh water.
Sea-water,
1 ton.
2 tons.
3 tons.
Iton.
2 tons.
3 tons.
0-4°
3-4°
11-8°
15-0°
0-00004770
4671
4415
4338
0 00004617
4521
4276
4219
0-00004510
4395
4163
4102
0-0000435
427
404
398
0-0000420
413
392
387
0-0000410
403
3835
378
decreases with rise of pressure as shown by the vertical column, Table V, for water
above 50°. This is the behaviour which would be expected from a liquid composed
of particles of constant volume, but separated by spaces which can be changed in
size by pressure and temperature. With water the compressibility first decreases
with a rise of temperature up to about 50°. This is shown by the vertical columns
in Table V below 50°— at a higher temperature, the compressibility of water in-
creases with rise of temperature as is the case with other liquids. The minimum
Table V. — The Volume of Water at Different Temperatures and Pressures.
Pressure
kilograms
Volume (CO.
per gram)
; volume at 0° unity.
per sq. cm.
-20°.
-15°.
-10°.
-5°.
0°.
5°.
10°.
15°.
20°.
25°.
0
1-0017
1-0006
1-0000
0-9999
10001
1-0007
1-0016
1-0028
500
0-9800
0-9783
0-9776
9782
0-9791
0-9800
0-9812
0-9825
1000
9606
9592
9584
9596
9609
9623
9638
9654
1500
0-9401
9413
9404
9407
9420
9435
9451
9467
9483
2000
0-9233
9240
9248
9257
9265
9281
9298
9315
9332
9349
2500
9083
9092
9102
9115
9131
9148
9166
9185
9203
9222
3000
8957
8966
8978
8999
9009
9026
9044
9063
9081
9100
3500
—
8860
8872
8884
8903
8923
8944
8964
8984
9005
4000
—
8764
8772
8784
8805
8823
8842
8860
8878
8897
4500
— .
. —
8680
8691
8713
8721
8749
8767
8785
8802
5000
. — .
8593
8604
8626
8643
8661
8678
8696
8714
5500
—
— .
. —
8548
8565
8582
8599
8616
8633
6000
. .
_
. — .
. ,
8480
8496
8513
8529
8545
8561
6500
. .
_
— .
.
8414
8429
8444
8460
8475
8490
7000
—
—
—
.
8356
8370
8384
8398
8412
8426
7500
—
. — .
—
—
. —
8309
8321
8334
8346
8358
8000
. —
. —
—
— .
— .
_
8262
8273
8284
8285
8500
— .
—
. — .
. —
. —
—
—
8208
8218
8228
9000
. — .
—
—
—
— .
. —
— .
8149
8157
8165
9500
—
. —
. —
. —
. —
—
. —
—
8099
8106
10000
• —
' — ■
—
• —
—
■ — ■
—
—
8046
8050
temperature is but slightly affected by a change of pressure, but it becomes
less and less pronounced with a rise of pressure until it has almost disap-
peared at a pressure of 3000 kilograms per sq. cm. Again, according to E. H.
Amagat (1877), the compressibility of ether at 13'7° under 11 atmospheres
pressure is 0*000168, and at 100°, 0*00056 of its volume per atmosphere.
W^ith water, the reverse obtains, and the compressibility falls to a minimum at
about 63°, and then increases. Thus, C. Pagliani and G. Vicentini (1884)
420
INORGANIC AND THEORETICAL CHEMISTRY
found the coefficient of compressibility, j3, of water, per atmosphere, at different
temperatures :
Temperature .... 0° 2-4° 493° 66-2° 77-4° 99-2°
j8xlO' 503 496 403 389 398 409
The inversion at about 60° is supposed to be the joint effect of {a) the diminution
which attends the passage of ice into water molecules, and (b) that due to the
squeezing of the molecules together. The higher the temperature the less the
disturbance arising from the former ; at about 63°, it becomes negligibly small.
Under very great compression, E. H. Amagat (1891) found that water behaves
like a normal liquid, showing that under great pressures virtually all the ice mole-
cules are probably transformed into molecules of one kind. Hence, as indicated
above, the temperature of maximum density of water is lower the greater the
pressure. Under great pressures, too, ice can be melted, because pressures prevent
the formation of or depolymerizes the ice molecules. At still greater pressures
further complications arise. The compressibility at different pressures, expressed
in kilograms per sq. cm. at 0° and 22°, found by P. W. Bridgman, are :
Pressure
500
1000
2000
4000
6000
7000
11,000
Compressibility at 0° .
0-0224
0-0414
0-0735
01195
0-1520
0-1644
—
Compressibility at 22° .
—
0-0383
00679
0-1137
0-1465
0-1600
0-2042
The disturbances have been traced by G. Tammann and P. W. Bridgman to the
transformation of ordinary ice into a number of different allotropic forms.
Table V, showing the volume water at different temperatures and pressures,
was compiled by P. W. Bridgman in an important memoir : Water, in the liquid
and five solid forms, under pressure (1912).
Some results are plotted in Fig. 7. The
pressure and volume at 0° are indicated on
each curve. Each curve is drawn to scale, but,
for the sake of compactness, the curves for the
different pressures have been brought together,
otherwise the curves would have been about ten
times their present distance apart. Up to 3000
kilograms pressure, the expansion at 0° is in-
creased with a rise of pressure ; but at higher
pressures, the expansion decreases with rising
pressure. There thus appears to be a maximum
in the curve between 0° and 20°. The relation
between the temperature and volume of a given
mass of a normal liquid at a constant tempera-
ture is illustrated by the curve A, Fig. 8, where
the expansion is greater the higher the tem-
perature. With an abnormal liquid like
Fig. 7.-Relation between Tempera- ^a^^''' ^^^^l Passing the minimum at the tem-
ture and Volume of Liquid Water perature of maximum density, it might be
at various constant Pressures. expected that there would be a limit to the
- -. I, increase in volume with decreasing temperatures
when atll the water molecules have been converted into ice molecules ; and the
liquid might be expected to behave in a normal manner and decrease in volume
with a fall of temperature. The volume temperature curve B, Fig. 8, would
represent the behaviour of such a liquid. With water, the exploration is prevented
by freezing, but such a curve has been actually realized by P. W. Bridgman with
water under a constant pressure of 1500 kilograms per sq. cm. — Fig. 8.
According to 0. Tumlirz (1909),!* the data obtained by E. H. Amagat for the
relation between the pressure p and volume v of liquids at the absolute temperature
T, can be represented by the equation {2)-\-V){v—h)=RT, where R and h are
WATER 421
constants for any given substance, and P, the internal pressure, is a function of
temperature only, and is evaluated from the experimental data. The results
apply very well for pressures up to 3000 kilograms per sq. cm., but not so well for
P. W. Bridgman's results up to 10,000 kilograms per sq.
cm. The values given by the formula for the compressibility
become small too rapidly at high pressures, so that the
observed compressibiUty remains larger than the values cal- i
culated by 0. Tumlirz's formula valid at lower pressures. -|
One possibility is that b is not constant, and that the molecules ^
themselves are compressible, apart altogether from the
closing up of the intermolecular spaces. D. Tyrer com- ' Temperature.
pared the coefficients of isothermal and of adiabatic com-
pressibilities of water between 1 and 2 atm. pressure, and ^^^- ^T^° n"^^J''"^f
X J i. r.o nr^o J i /^/^o ^i i" -l i /• i perature Curves of
found at 0 , 70 , and 100 , the former to be respectively Liquids
502-8x10-7, 452-9 XlO-7, and 418-8x10-7, and the latter
respectively 502-5x10-7, 424-5x10-7, 4290x10-7. If j8 be the ordinary
isothermal compressibility, i.e. — dv/vdp, and a the adiabatic compressibility,
Cp the specific heat at constant pressure, v the specific volume, T the absolute
temperature,
T. W. Richards and C. L. Speyers find the compressibility of ice between 100 and
500 megabars, at —7-03°, to be 0*000012, that is, about one-fourth the compressi-
bility of water at a neighbouring temperature. K. R. Koch found the elastic
modulus of ice to be 626 kilograms per sq. cm.
According to W. C. Rontgen and J. Schneider, and M. Schumann (1887), ^^ the
coefficient of compressibility, j8, of aqueous solutions of salts is less than that of
water, and this the more the more concentrated the solution. Thus, with solutions
of potassium chloride.
Per cent, potassium chloride . . 0 2-52 5'35 10*68 16-81
iSxlO^ 500 481 424 400 354
The facts are explained by the assumption that the proportion of ice molecules is
less in aqueous solutions than in pure water. W. C. Rontgen and J. Schneider
could not confirm M. Schumann's observation that the compressibiUties of aqueous
solutions of potassium, calcium, ammonium, and strontium chlorides are greater
than that of water, for they always found a lower compressibility with these solutions
than with water. G. de Metz also found that cane sugar also diminishes the com-
pressibility of aqueous solutions in a similar manner ; and T. W. Richards and
S. Palitzsch, that the compressibility of aqueous solutions of urethane at 20° rapidly
decrease with increasing concentration from 42"25xlO-5 for pure water to 38-91
Xl0~6 with 39'4 per cent, solutions; the compressibihty then increases at first
slowly and then rapidly with increasing concentration. The results agree in
showing that the first effect of dissolving anything in water is to dissociate the ice
molecules ; increasing the pressure or temperature acts in the same direction.
With normal liquids there is an increase in the compressibility with increasing
concentration. K. Drucker found that the compressibility of aqueous solutions
of organic acids likewise show a minimum in the compressibility curve. The
compressibility coefficient of solutions of salts in water usually increases as the
temperature rises. According to J. Guinchant (1901), with pressures up to 4 atm.
the volume of the dissolved substance does not change ; the observed change is
solely due to the medium.
The tensile strength of liquids.— About 1850, M. Berthelot i6 filled a glass tube
nearly full of liquid, removed the air, sealed the tube, heated the tube until the liquid
422 INORGANIC AND THEORETICAL CHEMISTRY
almost filled the interior, and allowed the whole to cool slowly to the ordinary tempera-
ture. The liquid continued to fill the tube so that the volume of water was ^^^th
larger than it should be for the given temperature ; with alcohol the volume was ~rd
and ether ^^^th larger. This represents the tensile strength of water to be 50 atm . , and
over 100 atm. in the case of alcohol and ether. R. H. Worthington found 17 atm.
for the breaking strain of alcohol. J. Stefan and 0. Tumlirz calculated a strength
of about 2000 atm. for alcohol on the assumption that the internal pressure is a
measure of the theoretical tensile strength. H. M. Budgett found a breaking
strength for water of nearly 900 lbs. per sq. in., or about 60 atm. 0. Reynolds
recorded the fact that when a liquid is flowing through a pipe with a constriction,
the velocity may be so high that the corresponding diminished pressure of the liquid
is sufficient to break it. He regarded the effect as a boiling of the liquid under
diminished pressure. S. Skinner and F. Entwistle regard the phenomenon in the
constricted tube as a true tensile rupture produced in the moving liquid, and in their
study of the effect of temperature on the rupture of water flowing through con-
stricted tubes, they found indications that the tensile strength of water becomes zero
at about 320°, a temperature approaching the critical point of water, and by forcing
the liquid through a capillary constriction until the speed in the capillary is sufficient
to produce rupture, they found that the tensile strength becomes zero in the neigh-
bourhood of 245°. J. Larmor found that if J. D. van der Waals' equation holds
in the liquid state, a negative pressure can subsist only at fjnd of the absolute critical
temperature, meaning that the tensile strength could subsist up to 538° K. or 265°.
Eefebences.
1 J. Southern, Phil. Mag., 17. 120, 1803 ; J. Perkins, Jmirn. Arts and Sciences, 7. 148, 1824.
2 O. Knoblauch, R. Linde, and H. Klebe, Mitt. Forsch. Ver. Ing., 21. 33, 1905 ; A. Winkel-
mann, Wied. Ann., 9. 208, 1880.
^ H. Whiting, A New Theory of Cohesion applied to the Thermodynamics of Liquids and Solids,
Cambridge, Mass., 1884 ; W. C. Rontgen, Wied. Ann., 45. 91, 1891 ; J. J. van Laar, Zeit. phys.
Chem., 31. 1, 1899 ; H. Witt, Oefvers Vet. Akad. Fohr., 57. 63, 1900 ; H. M. Vernon, Phil. Mag.,
(5),31. 387, 1891.
* C. S. Hudson, Phys. Rev., (1), 21. 16, 1905.
« E. H. Amagat» Ann. Chim. Phys., (6), 29. 68, 605, 1893 ; G. A. Him, ib., (4), 10. 32, 1866 ;
J. Meyer, Nernsfs Festschrift, Halle a. S., 278, 1912 ; G. Tammann and K. Zepemick, Zeit. phys,
Chem., 16. 659, 1895 ; J. H. Vincent, Proc. Roy. Soc, 69. 422, 1902 ; J. Dewar, Chem. News, 85.
277, 288, 1902; F. Auerbach, Physik in graphischen Darstellungen, Berlin, 1912 ; G. Tammann,
Ueher die Beziehungen zwischen den inneren Krdften und Eigenschaften der Losungen, Hamburg,
1907 ; P. W. Bridgman, Proc. Amer. Acad., 47. 382, 1911 ; E. H. Amagat, Compt. Rend., 116.
946, 1893.
® J. A. Deluc, Sur les modifications de V atmosphere, Paris, 1772 ; T. A. Hope, Ann. Chim. Phys.,
(1), 53. 272, 1805 ; L. de Coppet, ih., (7), 28. 12, 1903 ; (7), 3. 246, 268, 1894 ; Compt. Rend., 125.
533, 1897 ; 128. 1559, 1899; 131. 178, 1900; 132. 1218, 1901; 134. 1208, 1902; F. Rosetti, ^ww.
Chim. Phys., (4), 10. 461, 1867 ; (4), 17. 370, 1869 ; C. M. Despretz, ib., (2), 62. 5, 1836 ; (2), 70.
49, 1839 ; (2), 73. 296, 1840 ; Count Rumford, GilberVs Ann., 20. 369, ]805 ; J. G. Tralles, ib.,
27. 263, 1807 ; G. G. Hallstrom, Pogg. Ann., 9. 530, 1827 ; P. Chappius, Wied. Ann., 63. 202,
1897 ; L. C. de Coppet and W. Muller, Compt. Rend., 134. 1208, 1902 ; E. H. Amagat, ib., 116.
779, 946, 1893; R. Wright, Journ. Chem. Soc, 115. 119, 1919; S. Lussana, Nuovo Cimento,
(4), 2. 233, 1895 ; J. D. van der Waals, Die Kontinuitdt des gasformigen und flussigen Zustandes,
Leipzig, 151, 1881.
' K. Scheel, Wied. Ann., 47. 440, 1892 ; P. Chappius, ib., 63. 202, 1897 ; W. KreitUng, Aus-
dehnung von Wasser und Alkohol-Wasser Mischungen, Erlangen, 1892; M. Thiesen, K. Scheel,
and H. Dieselhorst, Wiss. Abh. phys. tech. Reichsanst., 3. 1, 1900 ; M. Thiesen, ib., 4. 1, 1904 ;
J. J. Waterston, PhU. Mag., (4), 21. 401, 1861 ; (4), 26. 116, 1863 ; W. Ramsay and S. Young,
Phil. Trans., 103. 108, 1893 ; H. Weidner, Pogg. Ann., 129. 300, 1866 ; F. Rosetti, Pogg. Ann. Erg.,
5. 258, 1871 ; J. J. Pierre, Ann. Chim. Phys., (3), 15. 325, 1845 ; (3), 37. 74, 1853 ; G. A. Hirn,
ib., (4), 10. 32, 1867 ; G. Landesen, Schr. Nat. Ges. Dorpat, 1, 1902 ; 14, 1904; D. A. Goldhammer,
Zeit. phys. Chem., 71. 577, 1910 ; G. Tammann and K. Zepemick, ib., 16. 659, 1895 ; P. H.
Hofbauer, ib., 84. 762, 1913 ; H. Panebianco, Zeit. Kryst., 50. 496, 1912 ; A. Hess, Ber. deut.
phys. Ges., 3. 403, 1901 ; J. Dewar, Chem. News, 85. 277, 288, 1902 ; G. Tammann, Ueber die
Beziehungen zwischen den inneren Krdften und Eigenschaften der Losungen, Hamburg, 1907 ;
J. Ueyev, Nernst's Festschrift, Halle a. S., 278, 1912; W. Marek, Wied. Ann., 44. 171, 1891;
C. Bmnner, Pogg. Ann., 64. 116, 1845 ; W. Struve, ib., 66. 298, 1845 ; R. F. Marchand, Journ.
prakt. Chem., (1), 35. 254, 1845; J. Duclaux, Compt. Rend., 152. 1387, 1911; Journ. Phys.,
WATER 423
(5), 1. 105, 1911 ; D. 1. MendeMeff, 8ur la variation de densiti de I'eau pour echauffement, Paris,
1870 ; Pogg. Ann., 141. 618, 1870.
8 H. Kopp, Liebig's Ann., 96. 153, 303, 1855 ; I. Traube, ib., 290. 43, 89, 410, 1896 ; Zeit.
anorg. Chem., 8. 323, 338, 1895 ; Ber., 28. 410, 2722, 2728, 2924, 3292, 3924, 1895 ; 29. 1203,
2732, 1896 ; D. Berthelot, Compt. Rend., 128. 606, 1899.
» C. M. Despretz, Ann. Chim.. Phys., (2), 62. 5, 1836 ; (2), 70. 49, 1839 ; (2), 73. 296, 1840 ;
S. Lussana, Nuovo Cimento, (4), 2. 233, 1895 ; S. Lussana and G. Bozzola, Atti 1st. Veneto,
(7), 4. 785, 1893; L. C. de Coppet, Compt. Bend., 125. 533, 1897; 128. 1559, 1899; 131. 178,
1900 ; 132. 1218, 1901 ; Ann. Chim. Phys., (7), 3. 268, 1894 ; (7), 28. 145, 1903.
^'^ R. Boyle, An experimental history of cold, London, 1665 ; J. Pliicker, Pogg.
Ann., 113. 382, 1861; C. Brunner, ib., 64. 113, 1845; G. Duvemoy, ib., 117. 454, 1863;
R. Bunsen,»6.,141. 3, 1870; H. Kopp, Liebig's Ann., 93. 129, 1855; L. Dufour, Arch. Sciences
Geneve, 8. 89, 1860 ; Phil. Mag., (4), 24. 167, 1862 ; Compt. Bend., 34. 1079, 1862 ; A. Leduc,
ib., 142. 149, 1906 ; J. H. Vincent, Proc. Boy. Soc, 69. 422, 1902 ; L. Playfair and J. P. Joule,
Mem. Chem. Soc, 2. 401, 1845 ; E. L. Nichols, Phys. Bev., 8. 21, 1889 ; H. T. Barnes, Phys.
Zeit., 3. 81, 1901 ; H. L. Cooke, Trans. Boy. Soc. Canada, 8. 127, 143, 1902 ; J. von Zakrzewsky,
Wied. Ann., 47. 155, 1892 ; J. Dewar, Chem. News, 85. 277, 288, 1902.
^^ P. de Heen, Physique comparee, Bruxelles, 76, 1883.
12 J. Canton, Phil. Trans., 52. 640, 1762 ; 54. 261, 1764 ; J. Perkins, ib., 116. 541, 1826 ;
H. C. Oersted, Ann. Chim. Phys., (2), 21. 99, 1822 ; D. Colladon, ib., (2), 36. 113, 225, 1827 ; H.
V. Regnault, Mem. Acad., 21. 1, 1847 ; W. J. M. Rankine, Phil. Mag., (4), 1. 548, 1851 ; P. G.
Tait, Physics and Chemistry of the Voyage of the H.M.8. Challenger, London, 1888.
13 P. W. Bridgman, Proc. Amer. Acad., 47. 44], 1912; Zeit. anorg. Chem., 77. 387, 1912;
C. A. Parsons and S. S. Cook, Proc. Boy. Soc, 85. A, 332, 1911 ; E. H. Amagat, Ann. Chim.
Phys., (5), 11. 520, 1877 ; (6), 22. 137, 1891 ; (6), 29. 68, 505, 1893 ; C. PagUani and G. Vicentini,
Nuow Cimento, (3), 16. 27, 1884 ; G. A. Hulett, Zeit. phys. Chem., 33. 237, 1900 ; T. W. Richards,
W. N. Stull, J. H. Mathews, and C. L. Speyers, Journ. Amer. Chem. Soc, 34. 971, 1912 ; T. W.
Richards and C. L. Speyers, ib., 36. 491, 1914 ; D. Tyrer, Journ. Chem. Soc, 103. 1675, 1913 ;
T. Peczalsky, Compt. Bend., 157. 584, 1913 ; K. R. Koch, Ann. Physik, (4), 41. 709, 1913 ; P. G.
Tait, Proc Boy. Soc Edin., 12. 46, 1884 ; 20. 63, 141, 1892.
14 0. Tumlirz, Sitzber. Akad. Wien, 168. ii, 1, 1909.
16 M. Schumann, Wied. Ann., 31. 14, 1887 ; W. C. Rontgen and J. Schneider, ib., 29. 165,
1886 ; 31. 1000, 1887 ; J. Drecker, ib., 20. 870, 1883 ; 34. 954, 1888 ; G. de Metz, ib., 31. 14,
1887 ; K. Brucker, Zeit. phys. Chem., 52. 641, 1905 ; S. PagUani, Nuovo Cimento, (3), 27. 209,
1890 ; T. W. Richards and S. Palitzsch, Journ. Amer. Chem. Soc, 41. 59, 1919 ; H. Gilbaut,
Compressibiliti des dissolutions, Paris, 1897 ; Zeit. phys. Chem., 24. 385, 1897 ; G. Aime, Ann.
Chim. Phys., (3), 8. 257, 1843 ; L. Grassi, ib., (3), 31. 437, 1861 ; J. Guinchant, Compt. Bend.,
132. 469, 1901.
16 M. Berthelot, Ann. Chim. Phys.,{S), 30. 232, 1850 ; H. M. Budgett, Proc Boy. Soc, 86. A, 25,
1912 ; R. H. Worthington, Phil. Trans., 183. A, 355, 1892 ; J. Stefan, Wied. Ann., 29. 685, 1896 ;
0. Tumlirz, Sitzber. Akad. Wien, 100. 837, 1900 ; 101. 437, 1901 ; G. A. Hulett, Zeit. phys.
Chem., 42. 353, 1903 ; S. Skinner and F. Entwistle, Proc Boy. Soc, 91. A, 481, 1915 ; S. Skinner
and R. W. Burfitt, Proc Phys. Soc, 31. 131, 1919 ; J. Larmor, Proc London Math. Soc, (2), 15.
191, 1916 ; 0. Reynolds, B. A. Bep., 564, 1894.
§ 4. The Vapour Pressure of Water— Fusion and Boiling
So long as a body retains its normal state of aggregation and properties we can observe
an increase of temperature corresponding with an increase in molecular energy, but as
soon as the destruction of form begins to take place, the increase of temperature no longer
becomes sensible and the energy is directed to breaking up the structure of the body and
to keeping its molecules apart. When this has been accomplished, and not till then, the
additional energy imparted again produces accelerated motion, and the substance gets
hotter and hotter.' — W. Anderson (1887).
When a liquid evaporates in an open space, there is apparently a continual flow
of the molecules from the surface of the liquid into the space outside, and evaporation
proceeds at a steady rate, as long as the significant conditions remain constant.
The vapour pressure of water in the atmosphere varies because a state of equilibrium
has not been attained between the water and the atmosphere. Water is therefore
evaporating slowly, and K. Jablezynsky and S. Przemysky ^ have emphasized that
the rate of evaporation is a slow process of diffusion from the layers in the immediate
vicinity of the water which are saturated with vapour to the surrounding atmosphere ;
but, in consequence of secondary disturbances, the water never attains a state
of equilibrium with the atmosphere.
424 INORGANIC AND THEORETICAL CHEMISTRY
The conditions which favour rapid evaporation are important to the chemical
engineer because so many operations on a large and on a small scale are dependent
on this process. John Dalton made a first approximation to the laws of evapora-
tion in 1803 ; he said :
Some liquids evaporate much more quickly than others. The quantity evaporated is
in direct proportion to the surface exposed when all other circumstances are alike. An
increase in the temperature of the liquid is attended with an increase of evaporation, not
directly proportional. Evaporation is greater where there is a stream of air than where
the air is stagnant. Evaporation from water is greater the less the humidity previously
existing in the atmosphere when all other circumstances are the same.
The speed of evaporation depends on : (i) The nature of the liquid. B. G. Babington
showed that the speed of evaporation of solutions of many salts is less than for water,
and this the more the greater the concentration. Sea-water, lor example, evaporates
approximately 5 per cent, slower than fresh water under similar conditions,
(ii) The temperature of the liquid and that of the surrounding air ; the rate of
evaporation of water increases roughly with temperature, and it is roughly pro-
portional to the saturation pressure at that temperature whien the general humidity
of the air is low. If the water surface be colder than the dew-point temperature,
evaporation is negative and condensation begins, (iii) The pressure of the atmosphere.
The presence of one gas retards the diffusion of other gas molecules of hke or
different nature ; consequently, when the vapour pressure is comparatively small, the
rate of evaporation varies nearly inversely as the total barometric pressure, (iv) The
hygrometric state of the air. As a first approximation, the rate of evaporation is
directly proportional to the difference of the temperature of the wet and dry bulb
hygrometers, (v) The rate at which the vapour is removed from the surface of the
liquid. The speed of evaporation increases as the velocity of the wind, but the exact
rate is not certain. D. J. Fitzgerald finds the rate of evaporation is approximately
represented by jlTi(pi—po)0--{-ilw) inches per hour, where w represents the velocity
of the wind in miles per hour, (vi) The area of the evaporating surface. The rate
of evaporation increases as the area of the evaporating surface, but not necessarily
at the same rate. John Dalton supposed the rates were proportional, but with a
circular area in still air, the speed of evaporation increases approximately as the
square root of the surface area.
The relation between the rate of evaporation of a liquid from a circular area of
radius r and the pressure of the gas when P denotes the barometric pressure, that is,
the total pressure of gas and vapour ; and pi, the vapour pressure of the liquid at
the evaporating surface or the saturation pressure of the liquid, and Pq is the vapour
pressure in free air at a great distance from the evaporating surface, is, according to
J. Stefan,
P—Po
Rate of evaporation =4r^ log ^— ^^-^
where A; is a constant. This expression represents the rate of evaporation into still
air from a circular tank or pond filled flush with a relatively extensive plane which
neither absorbs nor gives off any vapour. When Pq and pi are small in comparison
with P, this expression reduces to : Rate of evaporation=4rA;(^i— ^o)/-^- The
evaporation from an elliptical surface under similar conditions when the major
and minor axes, a and b, do not differ greatly, is obtained by substituting V«6 in
place of r. When the major axis is several times larger than the minor axis, the
ratio of evaporation from elliptical surfaces is much greater than from circular
ones. Curiously enough, the mass of vapour which evaporates from the surface
of a liquid in a given time is not proportional to the surface area, as was once
supposed, for the rate of evaporation is not the same on all parts of the surface,
being fastest near the edges, and slowest near the centre. Rather is the rate of
evaporation more nearly proportional to the linear dimensions — thus with a
circular vessel, the rate of evaporation is more nearly proportional to the square
WATER 425
root of the area, i.e. to the radius than it is to the area of the circle. If the evapora-
tion takes place into still air from a vertical tube of fixed length and constant cross
section, when h represents the distance of the surface of the liquid from the top of
the tube ; a, the cross-sectional area of the tube : pi and ^2? *^® partial pressures of
the vapour at the free end of the tube and at the evaporating surface respectively :
ka - P—P2
Rate of evaporation —- ^ log ^ — ^—
h ^ P-pi
where ^ is a constant whose value for any given liquid can be determined since all
the remaining terms can be measured. Given Jc, therefore, the rate of evaporation
of the given liquid from a circular tube or well can be computed from the total gas
pressure and the vapour pressures at the surface of the liquid and at the top of the
tube. P. Vaillant found the rate of evaporation of a liquid, i.e. the quantity of
liquid evaporated per second, is proportional to its molecular weight, M, and the
four-thirds power of the maximum vapour pressure p ; so that the speed of evapora-
tion=KMp^, where Z is a constant=0"43 for normal liquids. For water,
jfiL=3 xO"4:3, which means that this liquid is polymerized.
Kinetic theory o£ evaporation. — The molecules of a liquid are probably much
closer together than is the case with gases, and they are accordingly subjected
to the action of comparatively powerful intermolecular forces. Diffusion also
shows that the molecules of a liquid are in motion, but, in consequence of
great overcrowding, the number of collisions must be comparatively great. The
molecules in the body of the Hquid are attracted by the other molecules, equally
in all directions, but at the surface the molecules can be attracted inwards alone.
What will happen to a molecule which, in the course of its wanderings, reaches the
surface ? If its velocity be great enough, the molecule will rush upwards beyond the
range of attraction of the other molecules in the liquid, and thus, passing into the
space above, become an integral part of the surrounding gas or atmosphere. On
the other hand, if the velocity of the escaping molecule be not great enough to
carry the molecule so far, the upward velocity of the molecule will become less and
less, and finally the molecule will fall back and plunge into the liquid again. The
case is somewhat analogous with the behaviour of a stone thrown up into the air.
If the stone were projected upwards with a sufficient velocity, say 50,000 feet
per second, it would leave the earth never to return. Many of the molecules
which leave the surface of the water fall back again ; those which leave and do not
return reduce the volume of the liquid, and finally lead to complete evaporation.
Just as the kinetic energy of some of the molecules of a liquid carries them into
the space above, so does the kinetic energy of the molecules of the gas phase cause
them to penetrate into and become an integral part of the liquid.
Evaporation from and condensation on the surface of a liquid are thus continuous
processes whose ratio may be any value whatever. When both values appear to
be zero, condensation and evaporation are really progressing at equal rates. As
usually understood, the term evaporation refers to the net loss, and condensation
the net gain in a given time. Raising the temperature of the liquid accelerates
the motions of the molecules and so hastens the process of evaporation. A draught
of air across the surface also favours the passage of the molecules away from the
atmosphere above the evaporating liquid and reduces the chance of return.
Steam may be wet or dry. These qualities are of great importance in boiler
and engine trials. Wet steam is water vapour which has minute globules of liquid
water mechanically entrained with the vapour. This may arise (i) by ebullition ;
water is projected into the steam space, part falls back, but part is carried along with
the current of steam, (ii) The steam may be subject to variations of pressure, and
some water is condensed as a mist during the adiabatic expansion of steam ; and
(iii) some water may also be condensed to mist as the steam passes through pipes
which are losing heat by radiation, etc.2
426 INORGANIC AND THEORETICAL CHEMISTRY
The cooling effect during evaporation.— In J. D. van der Waals' theory of
liquids, the mutual attraction of the particles of the liquid is the restraining force
which keeps them more or less together ; this force has been estimated to be very great
— some hundreds of atmospheres. During vaporization, the particles break away
from the surface in spite of this attraction ; this cannot be done without a supply
of energy, and the curious fact is that the escaping molecules attract the required
energy from the rest of the liquid so that a liquid becomes cooler during evaporation.
In 1755, W. Cullen placed water under a bell jar from which the air was rapidly
withdrawn,3 evaporation was so rapid that the water was cooled until it froze ;
similarly, by placing some liquid sulphur dioxide and water in a red-hot platinum
crucible the water therein has been frozen by the rapid evaporation of the sulphur
dioxide. The kinetic theory shows how this can occur. During evaporation, the
fleetest molecules can alone escape from the liquid ; the more sluggish molecules
cannot get beyond the range of attraction of the molecules remaining in the liquid.
The surface of the liquid acts as a kind of grid separating the faster from the slower
moving molecules. The fleetest molecules have the greatest kinetic energy, and
the temperature of a mass of molecules is proportional to the average kinetic
energy of the molecules. If, therefore, the fastest molecules escape, the more
sluggish molecules will remain behind, and the average velocity of the molecules of
the liquid must be reduced. Hence a Uquid which is evaporating is cooling rapidly.
Observations show that the temperature of a vapour is never very far from that of
the liquid which produces it, and it is therefore assumed that the attraction of the
liquid reduces the mean kinetic energy of the escaping particles down to near the
mean kinetic energy of the liquid, so that the kinetic energy of the molecules of the
vapour like that of the molecules of the liquid varies from zero upwards.
To illustrate the cooling effect of evaporation, a little ether is placed in a small beaker
with a few drops of water on the underside, the water will freeze if the ether be evaporated
quickly by blowing a jet of air across the surface. Advantage is taken of this fact to
solidify carbon dioxide by the rapid evaporation of liquid carbon dioxide ; and to solidify
hydrogen by the rapid evaporation of liquid hydrogen. If a large test tube containing
liquid air be fitted with a one-hole rubber stopper fitted with a tube connected with an
air pump, the tube becomes so cold that the outside air, in contact with the test tube, is
liquefied. The rapid evaporation of the liquid air inside the tube may even produce cold
enough to freeze the contents solid. The principle is also utilized in cold storage, etc.
If water be placed in a flask dipping in boiling water, its temperature remains many
degrees below the boiling point owing to the cooling effect of the evaporating
water ; but if the water in the flask be covered with a layer of oil, its temperature
rises to that of the bath, and bubbles of vapour pass through the oil.*
The heat o£ vaporization and fusion. — A relatively large amount of energy is
needed to transform a gram of water into steam. The thermal energy, or the work
done in accelerating the motion of the individual molecules and at the same time
imparting to the molecules sufficient momentum to tear them apart against the
attraction of those molecules remaining in the liquid is measured by the so-called
latent heat of vaporization. The amount of heat required to turn one gram of
water at 100° into steam at 100° reported by different observers ^ ranges from the
5320 to the 538'9 cals. per gram. T. W. Richards and J. H. Mathews found at
100°, 5381 cals., and A. W. Smith, 540*7 cals. per gram. Few methods of measure-
ment have been so unsatisfactory as those employed for the latent heat of vapori-
zation, and the published data are very discordant, due partly to impure materials
but mainly to faults in the method of measurement. The best representative
value may be taken as 540 cals., or 973 Cals. per gram-molecule at 100°. This is
the latent heat of vaporization of water at 100°. The number means that steam
at 100° has the equivalent of 537 cals. of energy — internal or potential — more than
liquid water at 100°, or that 537 cals. of thermal energy are needed to convert a
gram of liquid water at 100° into steam at 100°. In symbols, for a gram-molecule
(that is, 18 grams of water) : H20iiq.->H20gas~-9*7 Cals., meaning that during
the passage from the liquid to the gaseous state, energy equivalent to 9'7 Cals.
WATER
427
in becoming latent or potential, so to speak, is charged on to the molecules —
probably as kinetic energy of translatory motion. This energy is degraded as heat
when steam at 100° is cooled to liquid water at 100°. The values ^ for 0° range
from 599'92 to 587*5 cals. per gram. The best representative value may be taken
as 587*7 cals. per gram, or 10*69 Cals. per gram-molecule. The latent heat of
vaporization of water is the highest known, and this also helps to moderate the
earth's temperature, for it absorbs heat during its evaporation in torrid climates,
and gives it up during condensation in cooler climes. These properties of water
also help to regulate the temperature of living organisms.
During the change from liquid to vapour, a volume Vi of liquid expands to a
volume V2 of vapour. The total change in volume is therefore ^2— '^i, and this
expansion occurs in spite of the fact that the atmosphere continually acts as if it
were a weighted piston which has to be pushed back as the liquid expands into
vapour. If p denotes the pressure of the atmosphere, the work done during the
expansion can be represented by the product p(v2—Vi). If the volumes be ex-
pressed in c.c, the thermal energy equivalent to the work :p(v2— "^i) wiU ^^ 0*00003183
^(V2— "^i) cals. If then I represents the observed latent heat of expansion, and E
that portion which is spent in overcoming external work, the so-called external
heat vaporization, the residue, l—E, will represent the heat actually spent in doing
internal work as the substance expands ; this is the so-called internal heat o£ vapori-
zation A, where l—E=X. According to R. Clausius, the relation between the three
latent heats of vaporization of water and temperature, 6°, is
Z=607 -0*708^ ; ^=31*6+0*083(9 ; A=575*4-0*791^.
Many other formulse have been published. According to F. Henning,^ the effect
of temperature on the latent heat of water is given by the expression ?=93*706
(366*25—^)0*31312^ which deviates from the observed values by 0*3 per cent, at
140°, and by 0*2 per cent, at 180°. At the critical temperature, 366*25°, the latent
heat I is zero. Better agreement between the observed and calculated results at
d° between 100° and 140° is given by the formula : ^=538*46— 0*6422(0-100)
—0*000833(0-100)2, which in many cases can be simpUfied to /=539*66
—0*718(0—100). W. Nernst represents the molecular heat of the vaporization
of ice at T° by 11938+3*5^—0*0096^2 cals. F. Henning's observations of the
latent heat of vaporization of water I in Cals. (15°), at different temperatures
0, and L. Holborn and F. Henning's values for the saturation pressure p and the
temperature coefficient dp/dd in mm. of mercury per degree, are indicated in
Table VI.
Table VI. — Heats of Vaporization of Water at Different Temperatures.
0°C.
I cal3.
dl
de
p mm. Hg.
dp
de
t>2 C.C.
30
579-3
31-71
1-819
33010
40
5740
0-54
55-13
2-939
19600
50
568-5
0-56
92-30
4-588
12050
60
562-9
0-57
14919
6-916
7677
70
557-1
0-59
233-53
10-11
5046
80
551-1
0-61
355-1
14-40
3406
90
545-0
0-62
525-8
19-99
2360
100
538-7
0-64
760-0
27-12
1673
110
5321
0-67
1074-5
36-10
1210
120
525-3
0-70
1488-9
47-16
891-3
130
518-2
0-72
2025-6
60-60
667-5
140
510-9
0-72
2709-5
76-67
507-8
150
503-8
0-72
3568-7
95-66
392-1
160
496-6
0-72
4633
117-7
307-1
170
489-4
0-72
5937
143-4
243-0
180
482-2
—
7514
172-7
194-7
428 INORGANIC AND THEORETICAL CHEMISTRY
Similar remarks might be applied to energy changes during the melting of ice
as to the vaporization of liquid water ; and similarly with the freezing of liquid
water into ice, and the condensation of steam to Uquid water. In the case of melting
ice, one gram of ice at 0° in melting to liquid water at 0° requires about 80 cals. —
this is the so-called latent heat of fusion — in symbols, for one gram-molecule
(that is, 18 grms.), H206oiid->H20iiqi,id— 1'44: Cals. When compared with other
liquids, the freezing point of water is high ; the latent heat of fusion also, if liquid
ammonia be excluded, is the largest which has yet been discovered. All these
properties of water play an important part in regulating the temperature of the
earth, for a comparatively large amount of heat must be abstracted from a large
body of water before it can freeze, and this helps to prevent an excessive fall of
temperature in lakes and seas. The latent heat of fusion of ice was first investi-
gated by J. Black ^ near the beginning of the nineteenth century, and culminated
in the important work of H. V. Regnault (1847), in which 79'25 cals. was obtained
for the latent of fusion of ice. Other determinations at 0° range from the 79'20 cals.
per gram of E. Leduc (1906) to the 80-025 of R. Bunsen (1870). L. F. Guttmann
(1907) introduced certain corrections in A. W. Smith's data, and then obtained
79-67, the identical result obtained by W. A. Roth (1907). The best representative
value may be taken as 79-7 cals. per gram or 1-436 Cals. per gram-molecule — it is
interesting to note that this result is identical with the 79-7 cals. obtained by
J. Black in 1762. According to 0. Petterson, the value drops to 77-71 at —2-8° and
to 75-99 at —6-62°. According to P. W. Bridgman (1912), the latent heat of the
transformation of ice into water at different temperatures is :
Temperature . . . -20° —15° —10° —50° 0°
Latent heat .... 57-7 62*5 68*0 73-7 79-8 cals.
0. Petterson has studied the appli(;ation of Kirchhoff's equation dXIdT^C^—Ci
to water when C^ denotes the specific heat of the liquid, and Cj the specific heat of
the solid. The results were satisfactory When the fusion temperature T is
lowered 1°, the latent heat of fusion A diminishes C2— C^ calories. Similar
relations hold for other substances. The latent heats of fusion of some common
metals are :
Iron.
Copper.
Silver.
Zinc.
Gold.
Tin.
Lead.
69-0
430
24-3
22-6
16-3
13-82
4-0 cals.
If secondary changes — e.g. decomposition during fusion— do not occur, all sub-
stances exhibit characteristic latent heats of fusion and vaporization. In virtue
of these facts, it follows that weight for weight a liquid contains a greater amount
of energy than a solid, and a gas contains a greater amount than a hquid. In
order to change a solid to a liquid, or a liquid into a gas, energy must be added to
the substance, and for the converse changes, gas to liquid, or liquid to solid, energy
must be withdrawn from a substance. In general, when a substance passes from one
physical state to another, a definite amount of energy is simultaneously added to
or withdrawn from the substance. The energy needed for the evaporation of natural
waters is mainly derived from the " heat paid out by the sun."
Most solids expand in passing from the solid to the liquid state ; according
to G. Vicentini and D. Omodei (1886),^ one volume of the following elements
changes hv volumes on melting :
Cadmium. Mercury. Phosphorus. Lead. Tin. Potassium. Sodium.
hv . . 0-047 0-037 0-035 0034 0-028 0-026 0025
corresponding with a 3-3 per cent, average expansion, and the melting point of all
these soHds is raised by pressure. Water, on the contrary, contracts nearly 9 1 per
cent, on melting. As P. W. Bridgman has shown, the freezing temperature of
water is lowered approximately 1° for each 100 atmospheres in accord with the
fact that the specific volume of ice is greater than that of liquid water ; but when
the pressure has reached 2115 atmospheres, the freezing temperature begins to
WATER 429
rise again, because a form of ice denser than the liquid appears. Hence, the
maximum pressure obtained by cooling water in a closed vessel is 2100
atmospheres, and pipes capable of withstanding this pressure would never
burst by the freezing of water. The inolecular theory of the raising of the
melting foint by pressure is somewhat as follows : The mean distance between
the molecules of a solid vibrating about their centres of oscillation, is increased
when the temperature is raised ; this causes the solid to expand. When the
mean distance increases to such an extent that the motion of one molecule does
not retard those of the others, the molecules start migratory motions, and the
solid is said to melt. If, at the melting point, pressure is applied, the molecules
are forced nearer together, and a further rise of temperature is needed for the solid
to melt. An increase of temperature is needed to counterbalance the increased
pressure. The reason the melting point of ice is lowered by pressure depends on
the fact that an increase of pressure facilitates the conversion of the complex ice
molecules to simpler water molecules. A somewhat similar phenomenon probably
occurs with sulphur trioxide which also expands on solidifying. Mutual attractions
draw the molecules of water together during solidification; this requires an
expenditure of energy, for there is an evolution of heat during the passage of water
into ice. The latent heats of fusion and vaporization of water are abnormally high :
Water.
Mercury.
Sulphuric acid.
Acetic acid.
Benzene.
Latent heat of fusion
. 79-89
2-8
22-82
46-4
30-39
Latent heat of vaporization
. 536-4
62
122-1
79-8
94-4
The approximately 80 cals. required to fuse a gram of ice at 0° do not solely represent
the purely physical change, but probably include the latent heat of dissociation
corresponding with the change of some ice to water molecules, and the solution of
the remaining ice molecules in the water molecules. Similarly with the heat of
vaporization, there is here allowance to be made for the passage of some complex
water molecules into the simpler steam molecules.
The effect o! pressure on transition points. — Let unit mass of a hquid pass
into vapour at each of the temperatures T and T-\-dT, and let the respective vapour
pressures be p and j)-\-dp ; further, let the volume of the liquid be Vi and the volume
of the vapour V2 when the pressure is p, the external work done during the
vaporization of unit mass be p{'V2—'^h), and if A denotes the internal latent heat of
vaporization per gram, we have E-^L=X, and for a small change of pressure dp,
the work dE will be {V2—Vi)dp. Substituting these values of E—L and of dE,
i,e. respectively W—Q and dW, in H. von Helmholtz's equation (indicated later),
there remains :
a relation often called E. Clapeyron's equation,!^ because an equation similar in
meaning was deduced by E. Clapeyron (1834). The same result follows another
way. For equilibrium between heat energy, (A/r)MT, and volume energy, {V2—Vi)dp,
in any given change, [XIT)ldT must be equal to (vg— i'l)^^?, where V2—i\ represents
the increase in molecular volume by the change under consideration. Obviously,
if any five of the six terms are known, the sixth can be calculated.
The fraction dpjdT is sometimes called the pressure coefficient because it repre-
sents the change of pressure dp which occurs with unit change of temperature.
In words, the molecular heat of vaporization of a substance is equal to the product
of the absolute temperature into the pressure coefficient and the change of volume
which occurs when the substance changes its state. It is easy to see that an
analogous expression will be obtained if a similar argument be applied to any change
of state — liquefaction, vaporization, sublimation, allotropic and other physical and
chemical changes — and Q can be used in place of A, where Q denotes the heat of trans-
formation. Hence, said P. Duhem (1902) : Tons les changements d'etat physique ou
de constitution chimique dependent des memes his generates.
430 INOKGANIC AND THEORETICAL CHEMISTRY
If the latent heat of a change of state be positive, then the raising or lowering
of the transition point with unit change of pressure {dT/dP) is dependent on whether
the volume V2 is greater or less than the volume Vi. If the volume of the material
decreases during a change of state so that V2 is less than Vi, the transition point will
be lowered by pressure, and if the volume increases during the change of state so
that V2 is greater than Vj, the transition point will be raised by pressure. This
furnishes a general rule for the effect of pressure on transition points : an increase
of pressure favours that state which has the smaller volume. The latent heat of
fusion of ice is always positive, and consequently, as a corollary from Clapeyron's
equation, it follows that an increase of pressure lowers the temperature of trans-
formation of substances which contract on passing fro7n one state to another, for if
V2 be less than Vi, dp and dT must have opposite signs in order that A and T may
remain positive ; but if the pressure be increased, dp must be positive, and dT will
therefore be negative, and an increase of pressure will lower the temperature of the
transformation — e.g. ice, bismuth, bismuth sulphide, cast iron, nitre, and sulphur
trioxide contract on melting, hence, their melting points are lowered by increase of
pressure. Another corollary : an increase of pressure raises the transition temperature
of substances which expand on passing from one state to another ; for if A and T
are positive, and Vi be greater than V2, dp and dT must have the same sign, and if
dp be positive, dT will also be positive — e.g. the boiling points of liquids are raised
by an increase of pressure ; so are the melting points of solids like phosphorus,
sulphur, lead, tin, and many other metals which expand on melting.
In 1850, M. Faraday ^^ drew attention to the fact that when two moist pieces of ice are
in contact, pressure is not essential for the solidification of two moistened surfaces of ice.
J. Thomson (1860) tried to explain Faraday's experiment by capillary action, but M.
Faraday showed that this could not be the case because the blocks froze together when
under water as well as when in air. From a suggestion made by J. Hooker, J. Tyndall
applied the term regelation to the phenomenon, and both he and J. D. Forbes explained
regelation by assuming that the melting point of ice in the interior of a mass is lower than
the normal freezing point at the surface. Hence, it was assumed that when the two surfaces
come into contact, they become interior parts of the enlarged block and thus the water
film is at once frozen. The explanation now generally accepted is that water at 0° is a
saturated solution of ice molecules ; and a film of water at 0° with ice on both sides would
grow together by natural crystallization, without the need for introducing the effects of
pressure. This phenomenon is not to be confused with the lowering of the freezing point
of water by pressure. This latter is probably more truly a regelation because the ice
can be squeezed into water by pressure, and the ice re-forms when the pressure is removed.
W. Spring has also shown that many metals liquefy under great pressures, and
an increase in volume (decrease in specific gravity) has been observed after many
metals have been subjected to external mechanical forces — pressure, twisting,
rolling, etc. ; thus the specific gravity of bismuth changed as follows :
Pressure in atmospheres . . 1 18,000 27,000 36,000
Specific gravity . . . 9-783 9-779 9-655 9*586
The increase in volume which persists after the pressure has been withdrawn is
supposed to be due to the dislocation of the molecules of the solid as it was passing
to the liquid condition under the influence of a gradually increasing pressure, and
that pressure was relieved too quickly to give the molecules sufficient time to take
up the condition characteristic of the solid state.
Examples.' — (1) If the melting point of ice under a pressure of one atmosphere is 0°,
what will be the melting point of ice under a pressure of n atmospheres, and also in vacuo
when the latent heat of ice is 80 cals. and one c.c. of Hquid at 0° furnishes 1-09 c.c. of ice
at 0° ? One atmosphere pressure is equivalent to 1033-3 grms. of mercury per sq. cm.,
and since 1 cal. is equivalent to 47,600 dynes, or 80 cals. are equivalent to 80 X 47,600 dynes,
T = 273, dp = 1033'3, dp^n — 1 ; v^— v. = 0-09; and dT, the melting point 01 ice imder
a pressure of n atmospheres, is — 0-0074(n — 1) ; the melting point of ice in vacuo when
w=0 is 0-0074°. Lord Kelvin found a change of 0-0072° per atmosphere pressure.
According to E. Riecke (1912), if the pressure be p atm., the melting point of ice is lowered
^=0-00036p.
WATER 431
(2) Water boils at 100° under normal atmospheric pressure and at 100-1° under a
pressure of 1-00355 atm. The volume of one gram of steam at 100° is 1645*55 c.c., and
of liquid water at the same temperature 1*04 c.c. Show that the latent heat of steam
at 100° is nearly 532-6 cals. The deviation from the observed 537 cals. rests on the in-
accuracy in the values assigned to dp and dT.
(3) L.T. Reicher (1883) ^^ foxmd that when sulphur changes from a- to /S-sulphur at the
transition temperature 95-6°, there is an expansion of 0-0000126 c.c, the latent heat involved
in the transformation being 2*52 cals. Hence compare the computed change in the transition
temperature with change of pressure with the value 0*05° per atmosphere observed by
L. T. Reicher. dT/cZp =368-6 x 0-0000126 -^ 2-52 -0-045° per atmosphere.
In the special case of vaporization when the volume v^ of the liquid is very
small in comparison with v^ the volume of the vapour, Vi can be neglected without
sensible error ; and if the ordinary gas law, pv—RT, describes the behaviour of the
gas, Clapeyron's equation becomes pX=RT^dpldT), which can be written in the
equivalent forms :
1 dp _ X dlogp_ X
pdl^RT^' ~dT Rf^
This is called Clapeyron and Clausius' eauation— after R. Clausius (1851)— and it
is supposed to represent the observed data more accurately than Clapeyron's at
low temperatures, when the vapour pressure is small.
Example.^— Water at 760 mm. pressure boils at 373° absolute, and dp/dT, the variation
of the vapour pressure, is 27-12 mm. per degree. Hence, if i? = 1-985 cals., the molecular
heat of vaporization is (1-985 x (373)^ x 27-12)^(760 x 18) =547-5 cals., a number about
2 per cent, too large.
Vapour pressure. — Suppose that a liquid is evaporating in a closed vacuous
space. The fleetest molecules cannot escape into boundless space, and consequently
they accumulate as a gas or vapour in the space above the liquid. The concentration
of the vapour in the space above the liquid will go on increasing. The molecules
of the vapour behave like the molecules of an ordinary gas, and consequently a
certain percentage will plunge back into the liquid. The number of molecules
which return to the liquid from the space above per second of course increases as
the concentration of the vapour increases, although the rate at which the molecules
leave the liquid probably decreases as the concentration of the vapour increases.
When the number of molecules which return to the liquid in a given time is
equal to the number of molecules which leave the liquid in the same time, the
vapour must be saturated, and the system in equiUbrium. With the notation
previously used :
100°
Wateriiquid^Watersteam
The equilibrium, it will be observed, is not a passive (static) condition — that is,
a state of rest — for both processes are active (kinetic). There is a shower of mole-
cules streaming into the liquid, and an efflux of molecules away from the liquid.
The effect of one is neutralized by the other ; neither can produce any visible
result. Anything which disturbs this equality — e.g. a desiccating agent or a
condenser in the space above (as in distillation), etc. — will alter the conditions.
Experiments show that at a given temperature the vapour pressure of a Uquid
in contact with its own hquid is a constant quantity, but it increases as the
temperature rises, and is independent of the absolute amount of vapour and of
liquid present in the system. It is easy to see this. If the surface of the liquid be
doubled, it is true that twice as many molecules will leave the surface in a given
time, but twice as many molecules will return.
The higher the temperature of the liquid the swifter will be the movements
of the molecules ; the greater the relative number of molecules escaping from the
liquid per second into the supernatant atmosphere ; and the greater the resulting
pressure. The vapour pressure of water at 0° is just equal to 4*60 mm. of mercury.
432
INORGANIC AND THEORETICAL CHEMISTRY
This means that if a little water be introduced into the Torricellian vacuum of a
barometer, at 0°, the mercury will be depressed from 760 mm. to 755*4 mm. If
the mercury barometer be 760 mm. high, and a drop of water be introduced so
that there is a film of liquid water on the surface of the mercury, the height of the
mercury column will be reduced one-half if the temperature be raised to 81°, because
the vapour pressure of water at that temperature is nearly 380 mm. of mercury. The
higher the temperature, the greater the vapour pressure, provided all the water
is not vaporized ; but for any assigned temperature, the vapour pressure of
water has one fixed and definite value. The effect of temperature on the
vapour pressure of liquid water is indicated in Tables VII and VIII, and on the
corresponding vapour pressure of ice in Table VIII. These tables of physical
constants are very useful. E. W. Morley (1912) has said :
The importance of physical constants is that each one holds condensed in a small volume
the essence of many observations. Some constants are like the words in a dictionary or
the figures in a mathematical table. Such constants must be determined and tabulated
in order that the call for them may be answered without delay or waste of time.
The equilibrium pressures of water vapour in contact with the liquid — i.e. the
vapour pressure of water at different temperatures — have been measured by H. V.
Regnault, G. Magnus, L. P. Cailletet and E. Colardeau, and others. The results in
Table VII up to 70° are based on the measurements of K. Scheel and W. Heuse ;
from 70° to 100°, on the measurements of H. F. Wiebe ; and from 100° to
A. Baumann.
, and at 150°,
370° on the
In illustration.
3568-7 mm.
measurements of L. Holborn, F. Henning, and
the vapour pressure of water at 95° is 634*01
mm
Table VII.-
—The Vapour
Pressures or
Water
OVER THE Liquid between
-16° AND 370°.
Tempe-
Vapoiir pressures of liquid water in mm. of mercury.
rature.
°C.
0
1
2
3
4
5
6
7
8
^•
-1
2-144
1-979
1-826
1-684
1-551
1-429
1-315
_
_
-0
4-579
4-255
3-952
3-669
3-404
3-158
2-928
2-712
2-509
2-321
+ 0
4-579
4-926
5-294
5-685
6101
6-543
7-014
7-514
8-046
8-610
1
9-210
9-845
10-519
11-233
11-989
12-790
13-637
14-533
15-480
16-481
2
17-539
18-655
19-832
21-383
22-383
23-763
25-217
26-747
28-358
30-052
3
31-834
33-706
35-674
37-741
39-911
42-188
42-19
44-58
47-08
49-71
4
55-34
58-36
61-52
64-82
68-28
71-90
75-67
79-62
83-74
88-05
5
92-54
97-24
102-13
107-24
112-56
118-11
123-89
129-90
136-16
142-68
6
149-46
156-52
163-85
171-47
179-40
187-64
19619
205-07
214-29
223-86
7
233-79
244-11
254-82
265-91
277-41
289-32
301-65
314-42
327-64
341-32
8
355-47
370-11
385-25
400-90
417-08
433-79
451-07
468-91
487-33
506-36
9
526-00
546-27
567-19
588-77
611-04
63401
657-69
682-11
707-29
733-24
10X1
76000
1074-5
1488-9
2025-6
2709-5
3568-7
4633-0
5937-0
7514-0
94040
10x2
11647
14291
17376
20950
25064
29771
35127
41186
48011
55680
10x3
64290
73860
84480
96270
109300
123660
139480
157200
—
The vapour pressure of ice is less than that of water, and is quite appreciable.
Determinations have been made by H. V. Regnault, i3 L. Rolla, etc. The results
in Table VIII are based on the measurements of K. Scheel and W. Heuse. The
vapour pressure of ice explains how ice and snow can evaporate at temperatures
below 0°, without melting to liquid water. The curve ROO, Fig. 9, represents the
vapour pressure of liquid water, and the curve PO, the vapour pressure of ice.
W. Nernst calculates that at —20° the vapour pressure of water is 0*940 ; at —73°,
2-5x10-3; and at -173°, 6-6x10-16.
The exact relation between the temperature and pressure of a vapour in contact
WATER
433
with its own liquid is not known. Quite a number — bet\^een thirty and forty — of
empirical formulae has been proposed. J. Dalton's, the earliest,^* represented the
pressures increasing in geometrical progression while the temperatures increased
in arithmetical progression. In symbols, if Pq be the pressure at 0° and p the
pressure at 6°, p=pQa6, or, what is the same thing, log p=a-\-hd, where a,
a, and b are constants ; but H. V. Regnault's exact measurements on vapour
pressures proved J. Dalton's rule to be inaccurate. A number of formula of this
Table VIII. — The Vapoub Pressure of Water over Ice between 0° and —65°.
Vapour pressure of Ice in mm.
of mercury.
Temperature.
0
1
2
3
4
5
6
7
8
0
-0
4-579
4-215
3-879
3-566
3-277
3-009
2-762
2-533
2-322
2-127
-1
1-947
1-780
1-627
1-486
1-3.56
1-237
1-127
1-026
0-933
0-848
-2
0-770
0-699
0-633
0-574
0-519
0-469
0-424
0-383
0-345
0-105
-3
0-280
0-252
0-226
0-203
0-182
0-163
0-146
0-131
0-117
0-105
-4
0-094
0-083
0-074
0-066
0-059
0-052
0-047
0-042
0-037
0033
-5
0-029
0-026
0-023
0-021
0 021
0-018
0-016
0-012
0-010
0-009
-6
0-008
0-007
0-005
0-004
0 003
0003
type but with more terms have been employed, e.,^. J. B. Gobel (1905) represented
the vapour pressure of water, p, at 6° in the vicinity of 0° by ^=0 '4600 +0*03293^
+O-OO1O502_|_O-OOOO16703 ; and for ice, ^=0-45996+0-03741^-f 0-001895^2
4-0-0000716^3^ }Y. J. M. Rankine's vapour pressure formula (1849) is one of the
favourites ; G. Kirchhoff, in 1858, and A. Dupre, in 1869, employed similar
formulae. W. J. M. Kankine represents the vapour pressure p, at the absolute
temperature T, by the expression :
logp:
■■a+^+c\ogT
(1)
The constants, a, h, c of this formula have a physical meaning in that they are
related to the other properties of matter — specific heat, vaporization, and
molecular weight. For water, between —100° and 365°, P. Juliusburger gives
logio^=9-30027—21113-2T-i— 0-28771 logio^ mm. of mercury. Rankine's formula
has also been deduced by H. Hertz, J. W. Gibbs, M. Planck, and 0. Stern from
reasoning based on the assumption that Boyle's law is valid, meaning that the
formula is strictly applicable only with very small pressures. J. D. van der Waals'
vapour pressure formula
log P^.
<h')
(2)
has also been largely used. Here ^c=164940 mm., and Tc=-374° or 647*09° K.
respectively denote the critical pressure and absolute critical temperature ; / is a
number which H. von Jiiptner and H. Happel found to vary partly with the nature
of the substance and partly on the temperature ; and L. Schames found that there
is a minimum value of /which is the same with all substances at the same reduced
temperature. The value of /falls from 3-3261 at 0° to a minimum 3-1244 at 215°,
and rises to 3*2283 at 360°. I. W. Cederberg found that/=:aj8(r/^c-v)% where
a denotes the minimum value of/; ^ is a constant ; and y the reduced minimum
temperature. For water, a=3-1244 ; ^=1-7887; and y=0-7500. The deviations
calculated from the observed results, between —173° and the critical point, with
log yo/i?=a/3(^/^c-y)'(r/rc-l), are less than O'Ol per cent. J. D. van der Waals'
formula may be written log p—{^og pc+f)—fTc/T, which resembles Rankine's
VOL. I. 2 F
434 INORGANIC AND THEORETICAL CHEMISTRY
formula with the c log ^ term missing. Introducing W. Nernst's values for the
constants,
p. E. Brunelli (1917) claims that log ^=2308647-4-5 log T— 2980-46^-1
— 0"(X)278T4-0'CKXXX)2825T2 represents the vapour pressure of water over a longer
range of temperature than any hitherto proposed. T is taken to be 27309+^.
C. E. Carbonelli represents the vapour pressure jp at the absolute temperature T,
below the critical temperature, T^, by the expression
where a is a constant characteristic of the liquid. It is 2*21503 for carbon disulphide ;
2-58124 for chloroform ; 36765 for alcohol ; 2*92714 for water ; 2*80416 for ethyl
ether ; 2*64664 for benzene ; 2*79064 for sulphur dioxide ; 2*92481 for cyanogen ;
and 2*79 for ammonia. A. March deduces for the vapour pressure p at the absolute
temperature T
from Maxwell and Boltzmann's theorem, J. D. van der Waals' theory and the
quantum theory ; pe and Tc respectively denote the critical pressure and temperature ;
/3 is a constant whose value depends on the nature of the substance.
W. Nernst deduced a vapour pressure formula from Clausius and Clapeyron's
equation X=T{v2—Vi)dp/dT , where the unit of mass is the gram-molecule ; A is
the molecular heat of evaporation or sublimation ; v^ and Vi the specific volumes
of gas and liquid or solid phase respectively ; p is the pressure of the saturated
vapour at the absolute temperature T. If the effect of temperature T and pressure
p upon the specific volume v, and if the heat of evaporation be known, then the
results substituted in the Clausius-Clapeyron equation will give a relation between
p and T. W. Nernst adopts the empirical formulae :
p{v2-Vi)=RT(l-l); ^ndX=(Xo+AT-hBT^)(l-l) . (3)
\ Fc^ \ Pc
where A and B are constants ; R is the gas constant. The result of the substi-
tution is Xo-\-AT-\-BT^=T^Rd log pjdT ; or, after integration log p=-Xo/RT
-\-A log T/R-\-BT/R-{-C, where C is the constant of integration. If the gas con-
stant R be 1*985 and ordinary logarithms are used, W. Nernst's vapour pressure
formula for water becomes
log v= ^ 1 ~ log T+ -^ T+C . . (4)
^ ^ 4-571T^l-985 ^ ^4*571 ^^ * ^ ^
The constants Aq, A, and B can be evaluated through the second of equations (3),
and also indirectly by the relation dXldT=Cp—Cp, where Cp is the molecular
heat of the liquid or solid, and Cp that of the vapour. The numerical values for
any particular liquid can be also obtained by substituting the corresponding values
of p and T for three different temperatures, and solving the resulting equations.
It has been found that the so-called constant C is, for normal substances,
characteristic of the molecule to which it refers, and is independent of the physical
state of the substance, so that C is the same for the vapour pressure formulae of
both liquid and solid carbon dioxide. This has been established for a number of
substances by E. Falck and C. F. Miindel. Consequently C is called the chemical
constant of the molecules of the substance in question, and it enables the chemical
equihbrium of a reacting system to be computed when the thermal value of the
reaction is known.
WATER 435
The second of equations (3) represents the observed relations between the heat
of evaporation A and temperature T fairly well, and W. Nernst found that in a
great number of cases A is 35 ; hence, W. Nernst's vapour pressure formula becomes
'°S^=-4^rT+^'^''°s2'+|fi2'+C . . (5)
It will be observed that Nernst's vapour pressure formula is dependent on the
validity of the first of equations (3), which for small pressures reduces to Boyle's
law, and this law is more nearly in accord with observations the smaller the pressure.
Hence, C is best evaluated at low pressures. The first of equations (3) is also
dependent on J. D. van der Waals' law of corresponding states, and does not apply
to those abnormal substances which deviate from that rule. Water is one of the
abnormal substances, and its vapour pressure does not follow the rule. W. Nernst
(1910) therefore assumed that doubled molecules are present in aqueous vapour
corresponding with (H20)2=f=^2H20 ; and if D be the observed vapour density,
and Dj the theoretical value, the degree of dissociation y is y={D—Di)IDi.
Hence, W. Nernst represents the vapour pressure of water by the formula :
log ^=log^-^?^ -4-94 log J+23-44837 . . (6)
H. Levy also deduced an expression for the vapour pressure of water on the
assumption that the molecules of the liquid are partially associated into dihydrol
molecules. For the vapour pressure of ice, p mm. of mercury, at 6°^ M. Thiesen
gives log (i?/i?o) =9-632(1 —0-00036^)^/T, where T is the absolute temperature,
and pq the vapour pressure at 0°. He also gives log p=^S'S91po6l(262-\-6).
W. Nernst found
log j9=_?^-^_l-75 log T -0-00210r+6-5343
which he afterwards altered to log p=
5896 226 1200
- -^- +^ log T+3 log (e T _i)_^6 log (g t _i)_o-020837 XlO-i5r6_|_o-76876
to accommodate the results with the quantum formula for the specific heat of ice.
The two formulae give equally good results, and therefore, for calculations, the
simpler type is preferred. S. Weber's measurements of the vapour pressure of ice
accord well with the simpler form of W. Nernst's vapour pressure formula :
T°K. .
. 203-96°
199-55°
190-36°
185-61°
177-01°
175-21°
°C. .
. -69-13°
-73-54°
-82-73°
-87-58°
-96-08°
-97-38°
p obs.
2-92
1-50
0-338
0-141
0-029
0-020
p calc.
2-936
1-504
0-338
0-146
0-0293
0-0205
The calculated value for 7979° K., or 193-30°, is 1-3x10-23 mm. ; andfor 157-61° K..
or 115-48°, 000040 mm. K.Scheel and W. Heuse,and M. Thiesen and K. Scheel
found the vapour pressure of liquid water at 0° to be 4*5788 ± 0*0008 mm. of
mercury ; and for ice, 45785 mm. The vapour pressure of ice and water are the
same at the triple point, and therefore, practically speaking, the vapour pressure
of water will be the same for ice and undercooled water at the triple point.
The pressure and volume relations of dry saturated vapours have not been
represented by a satisfactory equation, and the law ^t;**= constant, which is appHcabl©
for the permanent gases, does not give a constant index with the vapours.
B. Leinweber i^ found that for steam the exponent n varies within wide limits for
large pressures, but for low pressures up to 0*35 atm. n is fairly constant.
The distinction between a gas and a vapour. — The distinction between a gas
and a vapour is somewhat vague. If the elastic fluid be very far from its tempera-
ture of liquefaction, or above its critical temperature, it is generally called a gas ;
436 INORGANIC AND THEORETICAL CHEMISTRY
and vapour if it is near its temperature of liquefaction, or below its critical tempera-
ture. Oxygen, nitrogen, etc., at ordinary temperatures are gases ; whereas
water or alcohol on evaporation furnish vapours. Otherwise expressed, a gas is
an elastic fluid at a temperature above its critical temperature, and a vapour is
an elastic fluid below its critical temperature, but in a liquid state. The term
permanent gases was once applied to gases like oxygen, nitrogen, etc., because
they could not be liquefied by any known process. The term has lost its significance
since all known gases which have been tried have been liquefied, and all but helium
solidified. However, the term permanent gas is sometimes even now applied to
gases which approach nearest to the ideal gases, and which deviate least from the
gas laws of Boyle and Charles.
Boiling or ebullition. — Steam or water vapour is an invisible colourless gas
which condenses to a visible cloud of small particles when it comes in contact
with the atmosphere. This is readily shown by boiling water in a flask ; inside
the flask, the vapour is invisible, and a cloud of minute water particles — condensed
steam — ^appears where the steam comes in contact with the cold air. Raising the
temperature of an evaporating liquid increases the average speed of the molecules,
and favours rapid evaporation. When the temperature is high enough, the ex-
posed surface of the liquid is not sufficient to allow the swift-moving molecules to
escape fast enough, bubbles of vapour are accordingly formed within the liquid.
Each bubble as it forms rises to the surface — increases in size as it rises — and
finally escapes into the atmosphere. The process of vaporization by bubble
formation is called boiling ; and the temperature at which boiling commences,
the boiling point of the liquid. When the vapour pressure of the liquid is the same
as the external pressure to which the liquid is subjected, the temperature does not
usually rise any higher. Increasing the supply of heat increases the rate at which
evaporation proceeds, or at which bubbles are formed. Hence it is sometimes con-
venient to define : The boiling point o£ a liquid is the temperature at which the
vapour pressure of the Uquid is equal to the external pressure exerted at any point
on the liquid surface. The external pressure may be exerted by the atmospheric
air, by vapour and air, by other gases, etc. Hence, the vapour pressure curve not
only represents the vapour pressures of a liquid at different temperatures, but it
also shows the boiling points of that- liquid under difl^erent pressures. Water boils
at 100° and 760 mm. pressure. The greater the pressure, the higher the boiling
point ; and conversely, the less the pressure, the lower the boiUng point — roughly,
the boiling point changes about ~° C. per mm. change of pressure for a few degrees
above and below 100°. These phenomena occur with liquids generally, and it is
therefore necessary to state the pressure when giving the boiling point of a liquid —
although if no pressure is stated, 760 mm. is understood. Thus at Quito (9350
feet above sea-level), with the barometer at its average height, 525*4 mm., water
boils at 90'1° ; and on the top of Mount Everest (29,002 feet), barometer at 255*3 mm.,
water would boil at 72°. Table IX represents the boiling points of liquid water
at atmospheric pressures ranging from 680 to 799 mm. of mercury.
From Clapeyron's equation, and Trouton's rule for water, and remembering that
pv=2T cals. when v represents a gram-molecule of saturated vapour, dtldp=Tll3p,
and if ^=760 mm., dT=Tdp/9S^. Consequently, the change dT produced in
the absolute boiling temperature T of a liquid when the pressure changes by the
small amount dp, will be dT=0'00010l2T.dp. Accordingly, a change of dp=l mm.
in the pressure of a liquid boiling at T=373° K, will produce a change, dT=0'0377°,
in the boiling point. The observed result is 0*0370° — approximately ■^j° per mm.
change of pressure. . The formula similarly gives approximate values for liquids
other than water. A comparison of the boihng points of some metals in vacuo
and at ordinary pressures are indicated in the following scheme :
Mercury,
Cadmium.
Zinc.
Potassium.
Sodium.
Silver.
In vacuo
. 155"
450"
560°
365°
418°
1360°
At 760 mm, .
. 357"
749°
920"
667°
742°
2070°
WATER
437
As a rule, if the boiling point has been observed at a pressure p not far removed
from the normal, it can be reduced to the standard by adding or subtracting
c(760—^)(273+6), where 6 denotes- the observed boiling point at a pressure ^, and
c is a constant approximately O'OOOIO for water and alcohols.
Table IX.' — ^Thb Boiling Points of Water at Different Pressures.
Height of
barometer
0° mm.
Boiling points "C.
0
1
2
3
4
5
6
7
8
9
68
96-915.
955
996
*036
♦076
*116
*166
*197
*237
♦277
69
97-317
357
396
436
476
516
555
595
635
674
70
97-714
753
792
832
871
910
949
989
♦028
*067
71
98-106
145
184
223
261
300
339
378
416
455
72
98-493
532
570
609
647
686
724
762
800
838
73
98-877
915
953
991
*029
*067
♦104
♦142
*180
*218
74
99-255
293
331
368
406
443
481
518
555
592
75
99-630
667
704
741
778
815
852
889
926
963
76
100-000
037
074
110
147
184
220
257
293
330
77
100-366
403
439
475
511
548
584
620
656
692
78
100-728
764
800
836
872
908
944
979
*015
*051
79
101-087
122
158
193
229
264
300
335
370
406
The asterisk means that the number in front of the decimal is to be raised one unit.
The boiling point of a liquid is raised either by increasing the external (atmo-
spheric) pressure ; or by the increase of pressure which occurs when a liquid is
heated in a closed vessel. The rise in the boiling point which occurs when a liquid
is heated in a closed vessel fitted with a safety valve was first utilized by Denis
Papin in a digester or autoclave for softening bones, and described in his La maniere
d'amolUr les os (Amsterdam, 1681). In la marmite de Papin, a temperature of 130°
was obtained, and the corresponding pressure — Table VII — was 2025*6 mm. of
mercury, on the assumption that water is alone confined in the boiler. The solvent
action of water at this temperature and pressure is much greater than at 100° —
the maximum temperature attainable under normal atmospheric pressure.
Conversely, observations on the boiling points of a liquid at different pressures
also show the vapour pressures of the liquid at different temperatures. Thus,
the vapour pressure of water at 0° is 4'6 mm. ; and water at a pressure of 4*6 mm.
boils at 0°. Hence, liquids which decompose at their boiling point under ordinary
atmospheric pressure can frequently be distilled without decomposition at the
lower boiling temperature obtained by reducing the pressure. Hydrogen peroxide
can be cited in illustration. Hence, the so-called process of distillation under
reduced pressure, or, as it is sometimes less accurately styled, distillation in vacuo.
S. T. Preston i^ and G. J. Stoney have estimated that there are not less than 2 '7 X lO^^
molecules per cubic centimetre of gas at n.p.t. ; when the gas is reduced to the
lowest obtainable pressure, say, yoooo^li atmosphere, there are still 2*7 X 10^ molecules
per c.c. Hence, in the best of so-called vacua, an enormous number of molecules
is still present. E. Fischer and F. Penzoldt estimate tlj^t one part of musk in
ten million parts of air and one part of mercaptan in fif^ thousand million parts
of air could be detected by the sense of smell.
The critical temperature of water. — There is a continuous diminution in the
additional temperature required to overcome the effect of an increased pressure
on the boiling point of water. This is shown by the following numbers :
Pressure
1 5 10 15 20 25 30 atm
Boiling temperature
100° 152° 180" 199° 213° 213" 235°
Rise per 5 atm. pressure
52° 28° 19° 14° 12° 10°
438 INOKGANIC AND THEOKETICAL CHEMISTRY
Hence, it appears likely that a temperature will ultimately be reached at which an
additional pressure will require no additional rise of temperature to convert the
liquid into vapour. Above that temperature, the temperature of vaporization
will be independent of the pressure. This deduction involves an extrapolation,
and hence there may be a flaw in the reasoning— the curve, for example, may be
asymptotic and approach but never reach the condition just indicated. Experiment
shows, however, that there is a critical temperature, nearly 366°, at and above
which no known pressure can condense water vapour into the Uquid condition.
At this temperature, the vaporous and liquid states merge into one. It is probable
that all other volatile substances have their own particular critical temperatures ;
for instance, the critical temperature of hydrogen is nearly —271°, and of oxygen
—119°. The critical pressure of water is 194*6 atm., the critical volume, 0003864,17
and the critical density, between 0'329 and 0'429.
Measuring the volume of moist gases.— In 1801, John Dalton showed that the
mass of vapour required to saturate a given space at a given temperature, and
consequently also the vapour pressure of a given liquid, is the same whether the
vapour be by itself, or associated with other gases upon which it has no chemical
action. In other words, the total pressure of a mixture of gas and vapour is the
sum of the partial pressures of each constituent ; hence, added J. Dalton (1801) :
If the aerial atmosphere was at once annihilated, leaving only its aqueous portion,
this would be but little augmented, because it already exists in the air nearly at a maximum
of that which the temperature can produce and support. The suppression of the obstacle
will only accelerate the evaporation, without sensibly augmenting the absolute quantity.
When a gas is confined over water, the observed volume of the gas is determined
by the mass of the gas as well as by the temperature and the barometric pressure.
By Ddlton's law the total pressure of the gas is the joint effect of two partial
pressures : (1) the partial pressure of the water vapour at the observed temperature ;
and (2) the partial pressure of the gas under observation. Hence the barometer
reading does not represent the pressure of the gas alone, but rather the pressure
of the gas plus the pressure of the water vapour. To find the latter, note the
temperature, and Table VII furnishes the desired vapour pressure of water expressed
in millimetres of mercury. This must be subtracted from the observed pressure
(barometer) in order to find the pressure of the gas at the temperature in
question.
Example.- — -What is the volume of 4^ litres of a gas at 0°, 760 mm., when it is measured
in contact with water at 15°, and the barometer reads 767*8 mm. ? From Table VII, the
vapour pressure of water at 15° is 12*8 mm. Hence the gas is really under a partial pressure
of 767*8 — 12*8 = 755 mm. of mercury. The problem is now to be solved like the example
previously indicated. The answer is 4*24 litres.
The boiling points o! hquids. — ^As a rule substances formed with a large evolution
of heat, and highly associated substances have a high boiling point. In 1801,
J. Dalton 18 suggested that different liquids, at temperatures equally distant from their
boiling points, have the same vapour pressure ; or, if 6 be the boiling point of the
substance under normal pressure, and ti the boiling point of the same substance
at another pressure ; and if 62 and ^2 ^6 the corresponding boiling points of another
substance, then, ^;^— ^2=^1— ^2- C). G. Schmidt found the rule to be valid for the
series of fatty acids — formic, acetic, propionic, butyric, . . ., acids — but not the
corresponding alcohols, and numerous workers i^ have found the rule to be invalid
when applied to liquids generally. U. Diihrung,20 in 1878, maintained that if $1
and 02 respectively denote the boiling points of two liquids under the same pressure ;
6i=a-{-hd2 when a and h are constants, and W. Kamsay and S. Young found a
similar rule applicable to some organic liquids ; for, if Ti and T2 represent the
absolute boiling temperatures of two liquids under a pressure p, and Ti and T2'
the boiling temperatures under another pressure p\ the ratios TilT2=Ti/T2^
=constant. As a rule, however, the relation is not so simple, and W. Eamsay and
WATER
439
S. Young find that Ti/T2='TilT2+a{Ti—Ti) better expresses the result, or,
when a is very small, TiJT^^T-^IT^+alT^—T^).
According to P. Walden,2i for a great number of non-associated organic liquids,
the boiling temperature Tj, is related with the specific cohesion a^ in sq. mm, by the
empirical expression Ma-ITi,=V\Q, where M denotes the molecular weight of the
substance ; since o-, the surface tension in dynes per cm., is equal to i'da^D,
where D is the density of the liquid, o-v/T^=5"67, where v denotes the molecular
volume.
Many attempts have been made to establish a relationship between the boiling
points and the composition of members of the same homologous series of carbon
compounds. H. Kopp,22 for instance, noticed that every addition of CH2 raised the
boiling point of some series by about 19°, and he laid down the rule : the same
differences in the chemical composition or organic compounds correspond with the
same differences in the boiling points ; but C. Schorlemmer 23 showed that with the
monohalogens of the normal paraffins, the differences were not constant, but de-
creased 2° at each step. It was soon found that Kopp's rule was not at all general,
and many empirical formulae have been proposed, but these are usually apphcable
only to a limited number of homologous series,^* and usually over but a limited
range of a particular series. Many of these formulae are described in W. Nernst
and A. Hesse's Siede und SchmelzpunJct (Braunschweig, 1893), and in W. Marckwald's
Ueber die Beziehicngen zwischen den Siedepunkten und der Zuzammensetzung der
chemischen Verhindungen (Braunscheweig, 1898). J. Walker had some success
with the formula Tf,=aM^, where a and b are constants for particular series, but
vary from series to series ; M represents the molecular weight ; and G. G. Longi- ^
nescu,25 with the formula TilT2=Mini^lM2n2*, where Mi and If 2 respectively
denote the molecular weights of liquids with ni and n2 atoms per molecule, and
boihng points T^ and T2. G. G. Longinescu's formula can also be represented
in the form (Tf)llOOD)^=n ; where Tb represents the boiling point ; D the density
of the liquid at 0° ; and n the number of atoms per molecule. For water, w=12,
while for the normal molecular weight 18, n should be 3. Hence, says G. G.
Longinescu the molecule of water must be more complex than (H20)4. P. Pawlewsky
found about seventeen organic liquids in which the difference between the boiling
points under normal pressure and the critical temperature was the same ; but
further investigations have shown that the rule is not even roughly applicable to
liquids generally. This is illustrated by the penultimate column of Table X,
from S. Young's Stoichiometry (London, 1908). The last column, however, shows
that the ratio of the absolute critical temperature, Tc, and the absolute boihng
temperature, T^, is more nearly constant, as pointed out by C. M. Guldberg in
1890,
n
=1-7
This rule, however, is but a rough approximation, though it may be employed to
get an idea of the critical temperature of a substance.
Table X.-— Relation between the Boiling Point and the Critical Temperature.
Substance.
Critical
temperature T^
Boiling point Tj-
Difference
Tc-Th.
Ratio
Hydrogen
35°
20°
15°
1-75
Oxygen .
154
90-5
63-5
1-70
Methane .
191
108-3
82-7
1-76
Octane . . .
569-2
398-8
170-4
1-43
Benzene . .
561-5
353-2
208-3
1-63
Water .
633
373
260
1-59
440 INORGANIC AND THEORETICAL CHEMISTRY
According to T. E. Thorpe and A. W. Rucker, if Di and D^ denote the densities
of a liquid respectively at the absolute temperatures Tj and T2, the critical tempera-
ture Tc is 0-50125(T2Di— TiZ)2)/(Z)i— Da) ; or, if Vb denote the specific volume
at the boiling point,
nV -273
-'''- 2(F»-1)
According to S. Young, the quotient pc^JTc, where fc^ Vc, and Tc respectively
denote the critical pressure, volume, and temperature, is approximately 22 for
normal liquids ; water gives the value 26*4. W. Herz showed that the ratio
increases in homologous series with an increase in the number of carbon atoms.
Relation between heat of vaporization or fusion and the boihng or freezing
point. — In 1823, C. M. Despretz 26 showed that the heat of vaporization Z of a
liquid divided by the increase in the specific volume which is experienced by the
liquid in passing into the state of vapour, is equal to the result obtained with any
other liquid at a temperature where the vapour pressures are equal ; that is, if
vi and Vg be the specific volumes of a substance in the liquid and gaseous states
at its boiling point, L=k{vg—vi), where A; is a constant, the same for all liquids.
J. D. van der Waals has given a theoretical foundation for C. M. Despretz's rule,
and shown that if the absolute temperatures of any two liquids be the same fraction
of their respective critical temperatures, then the volumes of the liquids and of
their saturated vapours is the same fraction of their critical volumes. C. M.
Despretz's rule agrees well with many liquids. Assume that Q, the latent heat of
vaporization in R. Clausius' equation, ^(log p)ldT=QIRT^, does not vary with
temperature, and that the molecular weight of the substance in the two states of
aggregation is the same, it follows, on integration, that log p=—QIRT-{-a. constant ;
otherwise expressed, Q=T {Riconst&nt— log p)}, where the bracketed term is
constant at the standard pressure p. Hence, the heat Q required to vaporize
one gram-molecule of a substance is equal to the absolute boiling temperature, at
atmospheric pressure, multiplied by a number which is always the same. A rule
analogous to this was proposed by F. Trouton (1884) : The molecular heat of
vaporization of a liquid is proportional to the absolute boiling temperature ; or,
the quotient of the molecular heat of vaporization and the absolute temperature
of the boiling point at one atmospheric pressure is a constant. This is known as
Trouton's rule, in symbols MX/T—a. constant which approximates to 20, when
L represents the molecular heat of vaporization, which is the product of the latent
heat of vaporization (per gram of liquid) into the molecular weight M of the
substance ; and T represents the boiling point of the liquid on the absolute scale.
R. Schiff (1886) found the rule valid for organic liquids, particularly those in related
groups. The following examples illustrate Trouton's rule :
Latent heat of vaporization (A) . 86-7 51-'4 30-5" 124-4 537
Molecular heat of vaporization . 6384 7962 4230 9666
Boiling point .... 46'' 76° 113-9 -61-6° 100°
Absolute boiling point (T) . . 319° 349° 386-9° 211-4° 373°
Trouton's constant (AM/T or L/T) 20-32 20-07 20-49 20-01 25-9
R. de Forcrand has made a simple extension of Trouton's rule : The molecular
heat of solidification of a gas is proportional to its absolute boiling point (760 mm.) ;
so that if L and >S respectively denote the molecular heats of vaporization and
fusion, (Z>+*S')/T=a constant which for a number of substances falls between 28
and 32. If A and s respectively denote the molecular latent heats of vaporization
and solidification at constant pressure, and M the molecular weight of a substance,
Trouton's and Forcrand's rules can be respectively symbolized :
f =20 (^+^=30
Carbon
Phosphorus
Stannic
disulphide,
trichloride,
chloride,
CS2.
PCI,.
SnCl4.
86-7
51-4
30-5
6384
7962
46°
76°
113-9
319°
349°
386-9°
1 20-32
20-07
20-49
WATER 441
It is generally considered that Trouton's rule is near the mark for substances
in the normal condition, or rather, when the molecules of the substance are the same
in the liquid and in the gaseous condition. There are two cases :
(i) Trouton's constant is greater than 20. Illustrating by example, ethyl alcohol,
C2H5OH, furnishes an abnormally high value, viz. 26*9 (=9443-^351•4). This is
attributed to the association of the molecules (C2H50H)„ in the liquid state, while
the molecules of the vapour are normal, C2H5OH. Consequently, the observed
latent heat of vaporization is the sum of two factors : (a) The heat absorbed
during the conversion of the molecules of liquid C2H5OH into vapour C2H5OH ;
and (6) the heat absorbed during the depolymerization, (C2H50H)n=wC2H50H,
in the liquid at its boiling point. Water is another example with Trouton's constant
25 '9 ; nitric oxide, 27'7 ; etc.
(ii) Trouton's constant is less than 20. Acetic acid, CH3COOH, furnishes an
abnormally low value for Trouton's constant, viz. 12'7 (=5094-f-391). If the
molecules of a vapour are associated and the molecules of the liquid are non-
associated, the observed molecular heat of vaporization will be less than would be
the case if the molecules were not associated in passing into the state of vapour
because heat is evolved during the polymerization of the molecules. In the case
of acetic acid, it happens that liquid and vapour molecules are associated to
approximately the same extent, and if the molecular latent heat be taken for
(CH3COOH)2 the quotient is normal.
Trouton's rule thus enables an estimate to be made of the molecular complexity
or the molecular weights of substances at their boiling points. Thus, iodine approxi-
mates I3 ; sulphur, Sg.g ; mercury, Hgi-ig 5 nitric acid, (HN03)x.37 ; etc. Even
among non-associated substances, however, Trouton's constant may increase con-
siderably with temperature if the substances chosen have widely different boiling
points. Thus :
Hydrogen. Oxygen. Carbon disulphide. Aniline.
Absolute boiling point . .20*4 90-6 319 457
Trouton's constant . . .10-4 18*13 20 '4 23-1
W. Nernst (1906) employed the empirical formula A/T=8-5 log T ; or X/T
=9-5logT-0-007T; E. C.Bingham (1906), A/ J=17-f0-011T ; andR.de Forcrand,
A/r=-10-l log T-l-5-0'009T+0'0000026r2 to represent the increase in the
value of Trouton's constant with temperature. W. Hess found that Trouton's
rule gave irregular results with homologous carbon compounds.
H. Crompton 27 introduced the idea of valency in Trouton's formula ; P. W.
Robertson, the cube root of the atomic or molecular volume ; W. Nernst, specific
cohesion ; E. Baud, the change of volume at the melting point ; J, de Guzman,
the coefficient of viscosity ; H. Tsutsumi, the specific resistance ; H. S. Allen and
K. Honda, electronic frequency ; and H. G. Wayling, Moseley's atomic number.
If N be the sum of the atomic numbers of the atoms of the elements forming a com-
pound, and T the absolute temperature of the melting point, H. G. Wayling writes
N/T=a, number ranging from 2 to 5. The salts with water of crystallization have
higher values ; G. N. Lewis represents the relation between the heat of vaporization,
the absolute temperature T, the coefficient of expansion of the liquid a, the density
D, and the coefficient of compressibility j8, by A=— Ta/2)j8, when the liquid
is normal, and not polymerized or associated.
J. H. Hildebrand argues that the quotient of the heat of vaporization by the
absolute boiling point represents the increase in the entropy of a substance during
vaporization, per atmosphere pressure, and hence the term entropy of vaporization
can be employed for this ratio at temperatures other than the boiling point.
Trouton's rule fails for normal substances at extreme temperatures, because the
constant is greater the higher the boiling point. J. H. Hildebrand further showed
that by plotting log p against log T in Clapeyron and Clausius' formula
d log p A
dl6gT~RT
U2 INOKGANIC AND THEORETICAL CHEMISTRY
the tangent to the resulting curve at any point represents the entropy of vaporiza-
tion at that temperature divided by R. If Trouton's rule be valid, the tangents
to the curves for different substances should have the same slope at a value of log p,
corresponding to one atm. ; in reality, the slopes of the curves for equal values of
log p, increase regularly with log T. Consequently, the entropy of vaporization of
different substances cannot be the same at equal pressures, but rather at pressures
which increase in some way with the temperature. J. H. Hildebrand found that
the tangents to the curves at points cut by a line whose equation is log ^=log T-\-K,
where ^ is a constant, have the same slope " with a remarkable degree of precision."
For vapours at a low enough concentration to obey the gas law p=RTC, where
C denotes the molecular concentration, log ^=log T+log RC, and hence log RC—K,
and therefore along such a line C is constant. Hence, the entropy of evaporation
is the same for all normal liquids, not as in Trouton's rule when the vaporization takes
place at the same pressure {one atmosphere), but when it takes place at the same concen-
tration of vapour ; i.e. when the mean distance between the molecules is the same.
J. H. Hildebrand extrapolated the experimental data for A and T for values of G
arbitrarily selected to correspond with 0*00507 gram-molecules per litre. The values
of L/RTc, where Tg refers to equal concentrations of vapour, are between 13"1 and
13 9 for normal liquids and above 16 for associated liquids. There is no systematic
deviation with temperature, and the deviations are much less than with Trouton's
rule.
It will be observed that when a molecule escapes from a liquid to a vapour, it
is relieved of the high internal pressure which exists in the liquid, and it may con-
ceivably expand with an absorption of an amount of energy e so that the entropy of
vaporization becomes {X-\-e)IRT. The value of e is probably greater than for
molecules containing the larger numbers of atoms, but in general the value of e
is probably small in comparison with A, for most of the energy of vaporization is
expanded in overcoming intermolecular attraction, and but little in the expansion
of the molecule itself. With associated liquids, energy is further expanded in the
dissociation of complex molecules into simpler ones, and the total entropy of vapori-
zation is greater than the normal value. For liquids at low temperatures where the
specific heats change rapidly with changes of temperature, deviations from the
rule at low temperatures might be anticipated.
Refeeences.
1 J. Dalton, Proc. Manchester Lit. Phil. Soc, 5. 574, 1802 ; J. Stefan. Sitzber. Akad. Wien
68. 385, 1870 ; 73. 943, 1881 ; J. von PaUich, ih., 106. 384, 1897 ; A. Winkelmann, Wied Ann.
35. 401, 1888 ; B. I. Sresnewsky, Journ. Russian Phys. Chem. Soc, 14. 483. 1882 : 15. 1, 1883
R. E. Horton, Eng. Neics Escord, 78. 191, 1917 ; D. J. Fitzgerald, Trans. Amer. Inst. Civ. Eng.
15. 588, 1886 ; W. J. Humphreys, Journ. Franklin Inst., 185. 517, 1918 ; B. G. Babington
Proc. Roy. Soc, 10. 127, 1859 ; K. Jablezynsky and S. Przemysky, Journ. Chim. Phys., 10. 241
1912 ; P. Vaillant, Compt. Rend., 150. 213, 1910 ; M. le Blanc and G. Wuppermann, Zeit. phys
Chem., 91. 143, 1915.
2 W. C. Unwin, B. A. Rep., 393, 1894.
' W. Cullen, Essays and Observations, Edinburgh, 2. 145, 1755.
* C. Tomlinson, Proc. Roy. Soc, 37. 113, 1884.
5 T. Andrews, Journ. Chem. Soc, 1. 27, 1849 ; J. C. Brown, ib., 83. 987, 1903 ; C. Schall,
Ber., 17. 2199, 1884 ; D. Marshall and W. Ramsay, Phil. Mag., (5), 41. 38, 1896 ; E. H. Griffiths
and D. Marshall, ib., (5), 41. 1, 1896; L. Kahlenberg, Journ. Phys. Chem., 5. 215, 284, 1901 :
J. H. Mathews, ib., 21. 536, 1917 ; T. C. Sutton, Proc Roy. Soc, 93. A, 155, 1917 ; H. V. Reg-
nault, Mem. Acad., 21. 728, 1847 ; W. E. Luginin, Ann. Chim. Phys., (7), 13. 289, 1898 ; T. W.
Richards and J. H. Mathews, Journ. Amer. Chem. Soc, 33. 863, 1911 ; M. Berthelot, Compt.
Rend., 85. 646, 1877 ; A. W. Smith, Phys. Rev., 17. 193, 1903 ; 34. 173, 1911.
« C. Dieterici, Wied. Ann., 37. 504, 1889 ; A. Winkelmann, ib., 9. 208, 358, 1880 ; A. Svensson,
Wied. Ann. Biebl, 20. 356, 1896 ; H. C. Dickinson and N. S. Osborne, Journ. Franklin Inst., 179.
489, 1915.
7 F.Henning,vlww.P%.«l,(4),21.849,1906; (4), 22. 626, 1907 ; (4), 29. 441, 1909; (4), 58. 769,
1919; C. Dieterici, ih., (4), 16. 610, 1905 ; L. Holbom and F. Henning, ib., (4), 26. 833, 1908 ;
(4), 29. 441, 1909 ; C. Dieterici, Wied. Ann., 37. 504, 1889 ; A. Winkelmann, ib., 9. 208, 358, 1880 ;
W. Nemst, Verh.deut.phys. Ges., 11. 313, 1909 ; A. Franc, Zeit. Ver. deut. Ing., 979, 1036, 1069,
WATER 443
1891 ; H. V. Regnault, Relation des expiriencea entreprises pour determiner les principales lots
physiques et les donnies numeriques qui entrent dans le calcul des machines a vapeur, Paris, 635,
1847 ; E. H. Griffiths, Phil. Tram., 186. A, 261, 1895 ; G. P. Starkweather, Amer. Journ.
Science, (4), 17. 13, 1899 ; N. G. Ekholm, Bihang. Handl. Svensk. Akad., 15. 6, 1884 ; H. N.
Davis, Proc. Amer. Acad., 45. 267, 1910 ; A. W. Smith, Phys. Rev., 26. 192, 1908 ; H. V. Regnault,
Mem. Acad., 21. 728, 1877.
^ J. Black, Lectures on the Elements of Chemistry, Edinburgh, 1. 156, 1803 ; Ann. Phil., 5.
326, 1815 ; H. V. Regnault, Mem. Acad., 21. 1, 1847 ; Ann. Chim. Phys., (3), 8. 27, 1843 ; H. C.
Dickinson, D. R. Harper, and N. S. Osborne, Journ. Franklin Inst., 176. 453, 1913 ; 179. 489,
1915 ; A. W. Smith, Phys. Rev., 16. 383, 1903 ; 17. 193, 1903 ; A. D. Bogojawlensky, Mem. Soc.
Dorpat, 13. 1, 1904 ; L. F. Guttmann, Journ. Phys. Chem., 11. 279, 1907 ; P. W. Bridgman, Proc.
Amer. Acad., 47. 441, 1912 ; F. de la Provostaye and P. Desains, Ann. Chim. Phys., (3), 8. 5,
1843 ; P. Desains, Compt. Rend., 16. 981, 1843 ; E. Leduc, ib., 142. 46, 1906 ; G. G. Person,
An7i. Chim. Phys., (3), 21, 312, 1847 ; (3), 24. 265, 1848 ; (3), 30. 73, 1850 ; J. von Zakrzewsky,
Bull. Acad. Cracow, 153, 1892 ; Wied. Ann., 47. 157, 1892 ; 0. Petterson, Journ. prakt. Chem.,
(2), 24. 129, 151, 1881 ; R. Bunsen, Pogg. Ann., 141. 31, 1870.
» G. Vicentini and D. Omodei, Atti 1st. Veneto, (3), 3. 1, 1886.
10 E. Clapeyron, Journ. VJ^cole Polyt., 14. 153, 1834 ; J. Thomson, Trans. Roy. Soc. Edin., 16.
575, 1849; R. Clausius, Pogg. Ann., 81. 168, 1850; A. Horstmann, Liebig's Ann. Suppl., 8.
112, 1872; Lord Kelvin (W. Thomson), Phil. Mag., (3), 37. 123, 1850; G. Tammann, Wied.
Ann., 68. 553, 629, 1899 ; Ann. Physik, (4), 2. 1, 1900 ; E. Riecke, Neues Jahrb. Min., 97, 1912 ;
G. Quincke, Proc. Roy. Soc, 76. A, 431, 1905.
11 M. Faraday, Athenceumy 640, 1850; Experimental Researches in Chemistry and Physics,
London, 377, 1859 ; Proc. Roy. Soc, 10. 440, 1860 ; J. Thomson, ib., 8. 455, 1857 ; 9. 198, 1859 ;
11. 473, 1862 ; Phil. Mag., (4), 23, 407, 1862 ; (4), 24. 395, 1862 ; Collected Papers, Cambridge,
222, 1912 ; J. D. Forbes, Phil. Mag., (4), 16. 544, 1858 ; (4), 17. 197, 1859 ; J. Tyndall, Heat a
Mode of Motion, London, 224, 1880 ; Phil. Trans., 148. 228, 1858 ; H. T. Barnes, Ice Formation,
New York, 83, 1906; J. T. Bottomley, Nature, 5. 185, 1872.
12 L. T. Reicher, Rec Trav. Chim. Pays-Bas, (2), 2. 46, 1883.
1^ H. V. Regnault, Relation des experiences entrepriscs pour determiner les principales lois
physiques et les donness numeriques qui entrent dans le calcul des machines a vapeur, Paris, 1847
K. Scheel and W. Heuse, Ann. Physik, (4), 29. 723, 1909 ; (4), 31. 715, 1910 ; L. Holbom and
F. Henning, ib., (4), 26. 833, 1908 ; F. Henning, ib., (4), 22. 609, 1907 ; L. Holbom and A. Bau
mann, ib., (4), 31, 945, 1910 ; M. Thiesen, ib., (4), 29. 1057, 1909 ; Wied. Ann., 67. 690, 1899
G. Maanus, Pogg. Ann., 61. 222, 1844 ; L. P. Cailletet and E. Colardeau, Journ. Phys., (2)
10. 133, 1891 ; A. Smith and A. W. 0. Menzies, Journ. Amer. Chem. Soc, 32. 897, 1910
W. Ramsay and S. Young, Phil. Trans., 183. 107, 1892 ; C. F. Knip, Phys. Rev., 11. 141, 1900
A. Battelli, Mem. Accad.' Torino, (2), 43. 63, 1892; J. J.Jnhlin,Biheng.Svenksa Akad., 17. 72
1891 ; 0. Knoblauch, R. Linde, and H. Klebe, Mitt. Forsch. Inaen., 21, 1905; J. M. Crafts,
Journ. Chim. Phys., 13. 102, 1915; W. Nemst, Verh. deut. phys. Ges., 11. 313, 1909; 12. 565
1910 ; H. Levy, ib., 11. 328. 1909 ; K. Scheel, ib., 5. 287, 1903 ; O. J. Broch, Trav. Bur. Internat
Poids Mes., 1. 22, 1881 ; P. Chappius and J. A. Harker, ib., 12. 65, 1902 ; H. F. Wiebe, Zeit
Instrumentkunde, 13. 329, 1893 ; Tafeln iiher die Spannkraft des Wasserdampfes, Braunschweig
1903 ; L. Rolla, Atti Accad. Lincei, (5), 18. ii, 465, 1900 ; R. Bomstein and W. A. Roth
Physikalisch-Chemische Tabellen, Berhn, 1912.
1* J. Dalton, Mem. Manchester Lit. Phil. Soc, 5. 5.50, 1801 ; W. J. M. Rankine, Edin. New
Phil. Journ., 94. 235, 1849 ; Phil. Mag., (4), 31. 200 1866 ; E. Bose, Phys. Zeit, 8. 944, 1907 ;
G. Kirchhoff, Pogg. Ann., 104. 612, 1858 ; H. Hertz, Wied. Ann., 13. 198, 1881 ; M. Planck, ib.,
13. 535, 1881 ; K. Scheel, Verh. deut. phys. Ges., 3. 391, 1905 ; M. Jakob, Zeit. Ver. deut. Ing.,
1984, 1912 ; I. W. Cederberg, Phys. Zeit., 15. 697, 1914 ; 0. Stem, ib., 14. 629, 1913 ; E. Falck,
ib., 9. 433, 1908 ; C. F. Mundel, Zeit. phys. Chem., 85. 435, 1913 ; J. B. Gobel, ib., 53. 213, 1905 ;
P. H. Hofbauer, ib., 84. 764, 1913 ; H. von Juptner, ib., 55. 738, 1906 ; 60. 101, 1907 ; 63. 355,
1908 ; J. D. van der Waals, Arch. Nierl., (2), 9. 1, 1904; J. W. Gibbs, Trans. Connecticut Acad.,
3. 108, 343, 1878; Scientific Papers, London, 1. 153, 1906; A. Dupre, Theorie mecanique
de la chaleur, Paris, 96, 1869 ; P. JuUusburger, Ann. Physik, (4), 3. 618, 1900 ; 0. Sackur,
ib., (4), 40. 100, 1913 ; K. Scheel and W. Heuse, ib., (4), 29. 723, 1909 ; H. Happel, ib.,
(4), 13. 345, 1904 ; M. Thiesen, ib., (4), 29. 1052, 1909 ; M. Thiesen and K. Scheel, Wiss. Abh.
phys. tech. Reichsanst., 71, 1900 ; W. Nemst, Gott. Nachr., 1, 1906 ; Theoretische Chemie, Stuttgart,
236, 741, 1913 ; Vortrdge iiber die kinetische Theorie der Materie und der Elektricitdt, Leipzig, 85,
1914 ; Ber. deut. phys. Ges., 12. 313, 568, 1910 ; G. A. Burrell and I. W. Robertson, Joiirn. Amer.
Chem. Soc, 37. 1893, 1915 ; P. E. Brunelli, Nuovo Cimento, (6), 14. ii, 55, 1917 ; A. March,
Phys. Zeit., 17. 299, 1916 ; L. Schames, Verh. deut. phys. Ges., 15. 1017, 1913 ; H. F. Wiebe,
Tafeln iiber die Spannkraft des Wasserdampfes, Braunschweig, 1903 ; J. B. Briot, Sur le develop-
ment des forces ilastiques de la vapour aqueuse, Paris, 1841 ; H. Levy, Thermodynamische Behandlung
einiger Eigenschaften des Wassers tind des Wasserdampfes, Berlin, 1910; C. E. Carbonelli, Gazz.
Chim. Ital, 49. i, 151, 1919
15 B. Leinweber, Zeit. Ver. detit. Ing., 60. 363, 1916.
^^ E. Fischer and F. Penzoldt, Sitzber. phys. med. Soc. Erlangen, 18. 7, 1886; J. Passy, Compt.
Rend. Soc. Biol, (9), 4. 84, 1892; C. B. Bazzoni, Journ. Franklin Inst., 180. 463, 1915; S. T.
Preston, Phil. Mag., (5), 4. 110, 1877; G. J. Stonev, ib., (5), 1. 177, 1876.
444 INORGANIC AND THEORETICAL CHEMISTRY
" D. I. Mendel6eff, Pogg. Ann., 141. 618, 1870; A. Nadejdin, Bull Acad, St. Petersburg, 12.
299, 1885; L. P. Cailletet aud E. Colardeau. Compt. Bend., 106, 1489, 1888; I. Traube and
G. Teichner. Ann. Physik, (4), 13. 620, 1904.
18 J. Dalton, Mem. Manchester Lit. Phil. Soc, 5. 560, 1801 ; C. G. Schmidt, Zeit. phys. Chem.,
7. 433, 1891.
1^ J. T. Mayer, Commentatio de lege vis elasticce vaporum, Gottingen, 1809 ; A. Ure, Phil.
Trans., 108. 338, 1818 ; C. M. Despretz, Ann. Chim. Phys., (2), 16. 106, 1821 ; (2), 21. 143, 1822 ;
A. Avogadro, Mem. Acad. Torino, 36. 215, 1832 ; C. Mangold, Sitzber. Akad. Wien, 102. 1093,
1893 ; H. V. Regnault, Mem. Acad., 21. 465, 1847 ; 26. 375, 1862 ; G. W. A. Kahlbaum, Zeit.
phys. Chem., 26. 596, 1898; B. Woringer, ib., 34. 257, 1900; G. W. A. Kahlbaum and C. G. von
Wirkner, Ber., 27. 1894, 3364, 1894 ; H. Landolt, Liebig's Ann. Suppl, 6. 129, 1868 ; A. Moss,
Phys. Rev., 16. 356, 1903 ; 0. Schumann, Wied. Ann., 12. 34, 219, 1880 ; 0. Masson, Phil. Mag.,
(5), 30. 412, 1890.
*" U. Diihrung, Nev^ Qrundgesetze zur rationellen Physik und Chemie, Leipzig, 20, 1878 ; Ber.,
27. 3028, 1894; Wied. Ann., 11. 163, 1880 ; 51. 223, 1894; 52. 5, 56, 1894; W. Ramsay and
S. Young, Phil. Mag., (5), 20. 515, 1885 ; (5), 21. 33, 1886 ; (5), 22. 37, 1886 ; A. W. Porter, ib.,
(6), 13. 724, 1907 ; J. D. Everett, ib., (6), 4. 335, 1902.
21 P. Walden, Zeit. phys. Chem., 65. 183. 1909.
" H. Kopp, Liebig's Ann., 41. 86, 169, 1842 ; 96. 1, 1855.
2» C. Schorlemmer, Chem. News, 25. 101, 1872.
2* J. Walker, Journ. Chem. Soc, 65. 193, 1894.
2^ G. G. Longinescu, Ann. Scient. Univ. Jassy, 1. 359, 1901 ; Journ. Chim. Phys., 1. 288,
1903 ; 6. 552, 1908 ; P. Pawlewsky, Ber., 15. 460, 2460, 1882 ; 16. 2633, 1883 ; C. M. Guldberg,
Zeit. phys. Chem., 5. 374, 1890 ; T. E. Thorpe and A. W. Riicker, Journ. Chem. Soc., 45. 135,
1884 ; W. Herz, Zeit. anorg. Chem., 95. 253, 1916.
2» C. M. Despretz, Ann. Chim. Phys., (2), 24. 323, 1823 ; W. Ramsay and S. Young, Phil.Mag.,
(5), 20. 516, 1885 ; (5), 21. 135, 1886 ; O. Masson, ib., (5), 30. 412, 1890 ; H. le Chatelier and
R. de Forcrand, Ann. Chim. Phys., (7), 28. 384, 531, 1903 ; J. D. van der Waals, i)i'e Continuitdt
des gasformigen und flussigen Zv^tandes, Leipzig, 138, 1881 ; London, 455, 1891 ; R. SchifF,
Liebig's Ann., 234. 338, 1886 ; F. Trouton, Phil. 'Mag., (6), 18. 54, 1884 ; R. de Forcrand, Compt.
Bend., 156. 1439, 1648, 1809, 1913 ; W. Nemst, Gott. Nachr., 1, 1906 ; E. C. Bingham, Journ.
Amer. Chem. Soc, 28. 723, 1906 ; J. H. Hildebrand, ib., 37. 970, 1915 ; R. Pictet, Archiv. Genkve,
76. 1876.
27 H. Crompton, Journ. Chem. Soc, 67. 316, 1895 ; P. W. Robertson, ib., 81. 1233, 1902 ;
E. Baud, Compt. Bend., 152. 1480, 1911 ; J. de Guzman, Anal. Fis. Quim., 11. 363, 1913 ; H. S.
Allen, Proc Phys. Soc, 2%. 204, 1916; K. Honda, Scient. Bep. Univ. Tokyo, 7. 120, 1918;
H. Tsutsumi, ib., 7. 93, 1918 ; H. G. WayIing,PM. Mag., (6), 37. 495, 1919 ; J. H. Hildebrand,
Journ. Amer. Chem. Soc, 37. 970, 1915 ; G. N. Lewis, Zeit. phys. Chem., 78. 24, 1911 ; W. Hess,
Zeit. anorg. Chem., 95. 253, 1916.
§ 5. Gibbs' Phase Rule
The phase rule is one of the most comprehensive generalizations known to man. It is
of unlimited application, and offers an accurate and ready means of classifying all states
of physical and chemical equilibria. — W. Mayerhofer (1893).
On plotting the vapour pressures of water at different temperatures, a curve OQ,
Fig. 9, is obtained. This gives the vapour pressure of water corresponding with any
given temperature when the liquid and vapour are in contact, and in equilibrium.
Call this the steam line, or vaporization curve. The ordinate of 0 represents the
vapour pressure of water at 0° ; at lower temperatures the water freezes. Plot in
a similar manner the vapour pressures of ice at different temperatures, and the curve
OP, called the hoar-frost line, or the sublimation curve, is obtained. Under these
conditions, there is no intermediate Uquid state, vapour condenses at once to a solid,
and the solid passes directly into vapour. Solid iodine, below its melting point 114°,
also vaporizes without liquefaction ; arsenic can be liquefied only be melting the
element under pressure ; since under ordinary conditions, arsenic subUmes without
fusion. It is found that the effect of pressure on the melting point of ice can be
represented by a curve ON, Fig. 9. The left-to-right downward slope of the
curve shows that the melting point of ice is lowered by increasing the pressure.
Thus the melting point of ice at different pressures, according to G. Tammann (1900),
is approximately :
Pressure .... 260 490 1100 1790 2020 atm.
Melting point . . . -2° -4° -lO'll" -IT-e" -2059^
WATER
445
and in vacuo^ ice melts at +0*0075°. To emphasize these relations the curves in
the diagram are slightly exaggerated. The curve ON is called the ice line or fusion
curve ; it represents the melting point curve of ice under uniform pressures. Before
progressing further, it will be convenient to fix special meanings to three terms :
component, phase, and degree of freedom or variance.
Components.— The components of a system are the individual substances
which are not decomposed in the process. The number of components chosen to
represent a system is the smallest possible. The components may be elements, or
compounds which behave in a system, for the time being, as ^/they were elements.
There is only one component in the system just considered, namely, water — H2O ;
and the components in an aqueous solution of sodium chloride are water (H2O)
and sodium chloride (NaCl). A solution of sodium sulphate in water in a closed
vessel contains four elements — sodium, sulphur, oxygen, and hydrogen — but
neither the sodium sulphate nor the water is liable to decomposition under the con-
ditions of the experiment. Hydrogen cannot be removed without simultaneously
removing oxygen, nor can sulphur be abstracted without taking away sodium and
oxygen at the same time. Accordingly, while the composition of the system can be
expressed in terms of four components, two are necessaty and two are superfluous, for,
if the quantities of any pair of these
four elements are stated, the other two
can be computed. Hence, only two
components are involved, namely, water
(H2O) and sodium sulphate (Na2S04).
Phases. — The components may
group themselves in various ways.
They may pass from one physical state
to another, as when water boils or
freezes ; they may form simple solutions,
as when salt dissolves in water ; they
may combine with one another in
various ways, as when sodium sul-
phate (Na2S04) forms the decahydrate
(Na2S04.10H20), etc. Every homo-
geneous state — soUd, Uquid, or gaseous
— which the components can produce
is called a phase. The phases of a system are the physical states in which the
components can exist. A eutectic or cryohydrate — represented by the solid
which separates from an aqueous solution of sodium chloride in the act of freezing
• — is not a phase because the eutectic contains two phases — NaCl and H2O. With
an aqueous solution of sodium sulphate at the transition point. Fig. 9, there are
four phases — Na2S04 ; Na2SO4.10H2O ; the saturated solution ; and the vapour
arising from the solution. With freezing water, there are three phases — ice, water,
and vapour. In homogeneous systems there can be only one phase, e.g. aqueous
solutions, solid solutions, gaseous systems ; and in heterogeneous systems there
are always two or more phases.
Variance or degrees of freedom of a system. — It will be remembered that the
condition of equilibrium of a gas with respect to temperature, pressure, and volume
is defined by the equation, pv=RT,ioT Risa, numerical constant whose value depends
upon the units of measurement. If only one of these variables be fixed, say the
volume, the state of the system will remain undefined, because the gas can retain
one fixed volume, and yet have very different values for temperature and pressure.
Two of the three variables must be known before the state of the system can be
defined unequivocally, without ambiguity. If any two of the three variables be
fixed, the third variable can assume only one definite value. The two fi:xed variables
are said to be arbitrary or independent variables ; the third variable, which can
be calculated from the condition of equihbrium {pv=RT) when the two independent
0° 10° 20°
Temperature.
Fig. 9. — Vapour Pressure Curves of Water.
446 INORGANIC AND THEORETICAL CHEMISTRY
variables are known, is called the dependent variable. Another term sometimes
used for the independent variable is degree of freedom ; the number of degrees of
freedom is also called the variance of the system (that is, the variableness of the
system, from the Latin variabilis, variable). The gaseous system under consideration
has two degrees of freedom because two physical conditions can be varied indepen-
dently. The degree of freedom or variance of a system is the number of indepen-
dent variables which must be fixed before the state of the system can be defined
without ambiguity. The gaseous system defined by the equation, pv=RTy is
bivariant, because it has two independent variables, or two degrees of freedom. The
system consisting of liquid water and vapour has two variables : vapour pressure
and temperature. So long as liquid water is present, the pressure is determined
solely by its temperature; given either the pressure or the temperature, the other can
be determined from the relation symbolized in the vapour pressure curve — Fig. 8.
Hence the state of the system is defined by two variables — the one is dependent, the
other independent. In other words, the system has one degree of freedom, that is,
the system is univariant.
Triple point. — The three curves PO, OQ, and ON — Fig. 9 — represent the con-
ditions of equilibrium of three two-phase systems : soUd-vapour, vapour-liquid,
and soUd-liquid respectively. These three curves meet at the point
0. Here three phases can coexist in equiUbrium. Hence the
point 0 is called a triple point. The co-ordinates of the triple
point are : pressure, 4'57 mm. ; temperature, 0'0076° C. If
the pressure or temperature be altered ever so little one of the
phases — ice or liquid water — will disappear and a two-phase
univariant system represented by a point on one of the curves
OP, OQ, ON will appear. At the triple point the system is
invariant. Confining our attention, for the moment, to the liquid
and solid, and neglecting the vapour, we can define : The freezing
or melting point is the temperature at which both soUd and
_ ,^ ^ . Uquid can exist side by side in contact with one another
'^^e^ron'^rh; without changing.
Phase Rule. Gibbs' phasc rulc.— J. W. Gibbs (1876-78) discovered an
important relation between the number of components, the
phases^ and the degrees of freedom of a system. According to one setting of
Gibbs' phase rule, a system will be in equihbrium when its variance is eaual to
the number of components in the system less the number of phases increased by 2.
In symbols :
J=C-P+2
where C denotes the number of components, P the number of phases, and F the
variance or degrees of freedom of the system. Otherwise expressed, P must be equal
to or less than C-}-2, that is, a system of C components in a state of equilibrium
cannot have more than C+2 phases. The number of possible variations in the
physical conditions of temperature, pressure, and concentration, without changing
the number of phases, is two more than the difference between the number of
components and the number of phases. Conversely, the number of phases in
a system can be determined from the maximum number of possible variations in
the physical conditions. The phase rule thus serves as a test for stable states of
equilibrium. Suppose the system water and steam. Fig. 10, be in equihbrium ; the
vapour pressure indicated by the manometer M is not altered if the cock S be closed,
and the globe A removed. This experiment emphasizes the fact that unlike chemical
equihbria in homogeneous systems, the equilibrium between diflterent phases —
heterogeneous equilibrium — is independent of the amount of substance in each
phase, a milligram of a solid in a saturated solution will be as truly in equilibrium
as a kilogram. Accordingly, the phase rule is a qualitative, not a quantitative
WATER 447
criterion of eauilibrium, it says nothing definite about the amount of each
phase.
Derivation of the phase rule.— The following argument is based on that of H. W. B.
Roozeboom. ^ A system is in stable equilibrium when its free energy has a minimum value.
If a system has several phases in contact with one another, each phase can be regarded
separately. Consider solid barium peroxide, BaOg, in contact with gaseous oxygen and
solid barium oxide, BaO. If some oxygen can pass from the gaseous to the solid phase,
some of the monoxide will pass into the dioxide. If E^ denotes the change in the free
energy per unit mass of oxygen added to the one phase, and E^ the corresponding change in
the free energy for the same component removed from the other phase, the total change in
the free energy per unit mass of oxygen will be E^ — E^ ; if this magnitude be negative,
oxygen will pass from phase 1 to phase 2, and conversely ; if the system is in stable equili-
brium, E^—Ei must be zero, and E^=E^. More generally, for stable equilibrium, the free
energy of each component in every phase must have the same value, and no other condition
is necessary.
If a heterogeneous system in equilibrium contains P phases and C components, and if
each phase be supposed to contain a certain amoimt of each one of the C components, it
follows that the composition of unit mass of each phase will be fixed when the amoTints of
C — 1 of the components which the phase contains are known, for the amount of the remaining
component is determined by difference. Since the composition of each phase is quantita-
tively defined by C — 1 variables, the composition of P phases, otherwise expressed, the com-
position of the whole system, will be fixed by P{C — \) variables. Besides composition,
however, r other variables — temperature, pressiu-e, etc. — can change independently, and
consequently, the state of the systemwill bedefined by r-[-P{C — \) variables. These variables
can be determined by remembering that the free energy of each component in each phase
can be represented by an equation which is a function of the pressure, temperature, composi-
tion (concentration), etc. ; but since any change in one phase implies a corresponding
change in each of the remaining P — 1 phases, the changes in each component will be de-
scribed by P — 1 separate equations. When the system is in equilibrium, the free energy
of each component in every phase must be equal, and therefore, the free energy of the C
components in the system will be described by C{P—\) equations. Consequently, for equili-
brium, the number of midetermined variables F in excess of the number of equations will
be P=r+P(C — 1)— C(P — 1), an expression which reduces to the phase rule for r
independent variables, and one dependent variable: F = G — P-\-r. When in addition to
composition, the state of the system is defined by pressure or temperature, r=2, and the
rule reduces toP = C — P + 2.
Invariant systems. — ^An invariant system has no degrees of freedom, and the
state of such a system cannot therefore survive a change of temperature or pressure.
In that case F=0, or P=C-\-2. This means that the system will have C-\-2
phases, if it is in equihbrium. If there be one component in the system, as in the
case of water at the triple point, three phases can coexist in equilibrium — ice, liquid
water, and steam. Otherwise expressed, if a system has three phases and one
component, the phase rule tells us that it will be invariant, and therefore the slightest
alteration of pressure or temperature will cause one of the phases to disappear.
Again, in a system with three components — bismuth oxide, nitric anhydride, and
water — and five phases — solution, vapour, and three solids — the system is invariant,
and the three soUd phases can exist at one temperature, one pressure, one concen-
tration of the solution. Three phases of one substance cannot exist in equilibrium
in one system — say, sulphur with two liquid and one vapour phase ; or water with
one solid, one liquid, and one vapour phase — and have an extended range of co-
existence for the two non- vapour forms, because such a system must be invariant,
and therefore cannot exist except at a single temperature and pressure.
Univariant systems. — These systems have one degree of freedom, and when the
system is in equilibrium, F=l, or P=C+1. If one of the variables be known,
the state of the system can be determined as indicated above. If the system bismuth
oxide, nitric anhydride, and water has two solid phases, it will be univariant, and the
system can exist at different temperatures or with different concentrations of the
solution, but at any assigned temperature, the liquid in equilibrium with the two
given solid phases cannot vary in concentration.
Bivariant systems. — These systems have two degrees of freedom, and hence
F=2, or P=C. Two variables must be known before the state of the system can
448 INORGANIC AND THEORETICAL CHEMISTRY
be determined. A saturated solution in the presence of an excess of the solute is
univariant, but bivariant if not saturated. In the former case there are two compo-
nents and three phases — soUd, solution, and vapour ; in the latter case there are two
components and two phases. Hence in the one case, ^'=2+2— 3 ; and in the
other, F=2-\-2—2. Again, in the region PON, Fig. 9, the system will be
bivariant, because there is only one phase and one component. Pressure and tempera-
ture may be altered without interfering with the state of the aggregation of the ice
so long as the variations keep within the boundary lines PO and ON. The same
remarks may be applied to the condition of the water represented by points in the
regions NOQ and POQ. In the system bismuth oxide, nitric anhydride, and water
previously considered, if only one solid phase is present the system will be bivariant,
and the solid can be in equilibrium at a constant temperature with a solution of
varying concentration, or with a liquid of a fixed concentration at different tempera-
tures.
Modification in the phase rule with restricted systems. — One of the chief
difficulties in the application of the phase rule turns on the proper selection of the
components. For example, if the four substances concerned in the system, HgS04
-f-H20^Hg04-H2S04, be considered as components of the system, the variance
will be one more than would be the case if mercuric oxide, water, and sulphur trioxide
be regarded as the components. It will be observed that in the first case there is a
limiting condition, for the concentration of the sulphuric acid is determined by that
of the mercuric sulphate. Each independent relation or fixed condition among
the components of a system reduces the variance of the system by one. Limita-
tions and restrictions may be introduced from chemical necessity or by arbitrary
choice. For example, in the reaction 2H2+02^2H20 at a high temperature,
the number of components may be taken as one since the free hydrogen and oxygen
are always in the fixed proportions characteristic of water vapour ; there is also one
gaseous phase, and the system is accordingly bivariant. On the other hand, if the
number of components be taken as two — hydrogen and oxygen — ^there is one restric-
tion on the ratio of their concentrations, and the system has accordingly two degrees
of freedom.
There are several different but equivalent methods of selecting the components.
For example, what is here called component has also been called an individual (T. W.
Richards, 1916), constituent (W. D. Bancroft, 1906), or system-component (F. Wald,
1906), and the true number of components C in a system is then regarded as equal to
the number n of individuals less the number of restrictions r, so that C=n—r.
In the equilibrium CaC03^Ca04-C02, the system has one degree of freedom if
no restriction be placed on the temperature or pressure, but if the temperature or
pressure be fixed, there is one restriction, and the system is invariant. There is not
a chemical limitation in the quantities of lime or carbon dioxide because adding
more of either constituent without altering the pressure has no effect on the equili-
brium. In a dilute solution containing potassium nitrate, potassium chloride, and
potassium bromide in equiUbrium with its vapour, in addition to the water there
are the four components, K, NO3, CI, Br, subject to the limitation that the gram-
molecular concentration of the potassium must be equal to the sum of the concentra-
tions of the three radicles, NO3, CI, and Br. In a similar solution of potassium
nitrate and sodium chloride, the components may be regarded as water plus the two
salts ; this makes the system tervariant ; but if the five components, water,
potassium, and sodium, and the two radicles, NO3 and CI, be considered as com-
ponents, it is necessary to reduce the corresponding variance of the system by two
owing to the two limitations imposed by the necessity for the concentration of the
potassium and nitrate radicle to be equivalent and likewise for that of the sodium
and chlorine radicle. This makes the variance of the system three the same as
before.
Object of the phase rule. — The phase rule is (1) a method of grouping into one
class, systems which behave in a similar manner. It is essentially a method for the
WATER
449
classification of states of equilibrium. Systems having the same variance behave
in an analogous manner under the influence of variations in temperature, pressure,
and volume or concentration. It makes no difference whether the changes be chemi-
cal or physical. As indicated above, the phase rule also tells us (2) whether the
phases of a heterogeneous system are those necessary for equilibrium ; (3) it is of
assistance in identifying chemical individuals among a series of basic salts or solid
solutions. This it does by indicating the variance of the system which, in turn,
indicates whether or not the existing constituents have such a degree of stability
that they can survive a change of temperature or concentration. A knowledge of
the conditions of equilibrium of a system containing solution and soUd may therefore
show whether one or a mixture of two solid phases is present. The phase rule is
therefore a help and guide in the interpretation of complex phenomena ; a set of facts
may be under investigation and a number of explanatory hypotheses may be devised.
The phase rule will select which hypotheses are worthy of being tested by direct
experiment, and which can be rejected as fundamentally unsound. There are some
differences of opinion as to the utility and value of the phase rule. Those who have
done successful work with its aid are usually enthusiasts, but some say that it is an
" over-ridden hobby " and dub it the " phrase rule."
Modification for systems affected by other than mechanical and thermal
energy. — Other variables (electricity, gravitation, capillary tensions, etc.) besides
concentration, pressure, and temperature may modify the state of equilibrium of
some systems. Thus, light modifies the state of certain chemical equihbria. This
introduces another degree of freedom, and the phase rule must be modified to allow
for the action of light on systems sensitive to this agent. Accordingly, the phase
rule would then read F=C-~P-\-3. W. D. Bancroft (1906) 2 adds : Experience
shows that there are many kinds of active light, and the phase rule would have to be
altered accordingly ; usually, however, a beam of light can be treated as though it
were homogeneous if the intensities of the constituent rays are varied uniformly.
Similar remarks ai^i^lj mutatis mutandis to othei agentfi, e.g. variations in volume, the
silent electric discharge, etc. Usually only mechanical energy (pressure) and thermal
Table XI.- — Classification or Systems by the Phase Rule {F=C—P-]-2).
Degrees of
System.
Components.
C.
Phases.
P.
freedom or
variance.
F.
Freezing water . . .
Water
Liquid ; solid ; vapour
Invariant
Water above 0° .
Water
Liquid ; vapour
Univariant
Unsaturated solution of so-
Water; salt
Solution ; vapour
Bivariant
dium chloride
Saturated solution of sodium
NaaS04;H20
Na2S04 ; Na2SO4l0H2O ;
Invariant
sulphate at transition
point
Freezing euteetic — sodium
solution; vapour
Water; salt
Two solids ; one liquid ;
Invariant
chloride and water
vapour
Solution of oxygen in water
Oxygen ; water
Gaa ; liquid
Bivariant
Steam and metallic iron in
Iron ; oxygen
One gas ; two solids
Univariant
a closed vessel
(hydrogen)
2N02^N204 .
NO 2
One gas
Bivariant
2H2 + 02^2H20
Hydrogen ;
(oxygen)
Mercury ;
One gas
Bivariant
Heated mercuric oxide (over
Gas and solid
Univariant
400°)
(oxygen)
Heated barium peroxide
Barium oxide ;
(oxygen)
Two solids ; one gas
Univariant
nCNOHgas^^„NnO„Hnsolid
CNOH
One gas ; one solid
Univariant
Heated CaC03v=^CaO+C02
CaO ; (COg)
One gas ; two solids
Univariant
VOL. I.
2 Q
450
INORGANIC AND THEORETICAL CHEMISTRY
energy (temperature) come into play, and the rule then reads, F=C—P-\-2. In
the application of the phase rule to alloys, minerals, and solutions when the vapour
pressure is negligibly small, only two variables need be considered — concentration
or volume, and temperature. For such condensed systems, the phase rule reads :
F=C-P-^1
Granite, composed of quartz, Si02 ; felspar, K20.Al203.6Si02 ; and mica, say,
K20.3Al203.6Si02, has three components : Si02, AI2O3, and K2O ; and three
solid phases : mica, quartz, and felspar. The system is univariant. It is also in
equilibrimn, because not being at a transition point, it is able to survive a small
variation of temperature without changing the state of the system.
EiXamples. — Table XI shows the phase rule classification of some typical systems.
A component in brackets is regarded as being restricted by stoichiometrical relations.
References.
1 J. W. Gibbs, Trans, Conn. Acad, 3. 116, 1875 ; The Scientific Papers of J. W. Oibbs, London,
1. 65, 1906 ; W. D. Bancroft, The Phase Pule, London, 1904 ; F. Wald, Journ. Phi/s. Chem., 1.
22, 1896 ; J. E. Trevor, t&., 1. 349, 1897 ; R. Wegscheider, Zeif. ph/s. Chem., 43. 89, 1903 ; 45.
496, 1903; 50. 357, 1904; 52. 171, 1905; H. W. B. Roozeboom, ib., 15. 150, 1894; F. Wald,
ib., 13. 337. 1893 ; T. W. Richards, Journ. Amer. Chem. Soc, 38. 983, 1916.
* W. D. Bancroft, Journ. Phys. Chem., 10. 721, 1906.
§ 6. Undercooling, Supersaturation, and Metastability
A metastable system, though stable, is constantly menaced by a spontaneous transfor-
mation.— G. Urbain (1912).
Undercooling. — Heat a solution of sodium thiosulphate to, say, 70° in a glass
flask ; stir the molten mass with a thermometer as it cools ; read the thermometer
every two minutes ; and finally plot the results on squared paper. A curve — called
a cooling curve — resembhng that illustrated in Fig. 11, ^4, will be obtained. The
terrace in the cooling curve at 56° shows that a change of some kind takes place in
the nature of the cooHng substance at 56°. The terrace corresponds with the
temperature at which solidification or freezing was in active progress. The sudden
slackening in the rate of cooling corresponds with the evolution of the latent heat of
[^
MWf
ll|llllllllllllll'MII||||l||
rt Cool 1 no Curve ittrni
[Viiitnrr
Under
CO
Yf^
H
::::i
[jtHmtH
■j*-
: 1
m
■-%-.::
5±
^4
Fig. 11. — Cooling Curves of Molten Sodium Thiosulphate.
fusion as the liquid solidifies. Kepeat the experiment, but do not agitate the liquid ;
take care that the cooling liquid is quite still and protected from dust by, say, a
loose plusj of cotton wool in the neck of the flask. A cooUng curve Hke that shown
in Fig. 11,5, will be obtained. The liquid does not freeze, and no abnormal behaviour
can be detected in the cooling curve. The liquid '* ought to " crystaUize at 56°,
but it does not. Drop a crystal of sodium thiosulphate into the liquid mass. The
contents of the flask seem to soHdify with almost explosive rapidity, and the ther-
mometer immediately indicates a rise of temperature. The phenomenon is illus-
trated by Fig. 11, C The liquid sodium thiosulphate at a temperature below 56°
WATER 451
is said to be surfused, or, better, undercooled. The liquid may be kept in the sur-
fused or undercooled condition an indefinite time, and the process of solidification
can be started, in general, only by the introduction of a crystal of the same type
as that which is formed during the solidification of the given substance. Often a
fleck of the right kind of dust floating in the air suffices to upset the state of apparent
equilibrium. Clear glasses and pottery glazes are solutions of siUcates which have
congealed to hard masses without crystallizing.
Supersaturation. — Similar phenomena occur if water be saturated with Glauber's
salt — Na2S04.10H20 — at 30°. Make sure that no excess of solid is in contact with
the liquid, and let the solution cool as before — without agitation and without dust.
Probably no salt will separate from the solution. The solubility curve of this salt
tells us that the solid " ought to " separate from the system as the temperature is
reduced. Here is another case of apparent, false, or metastable equilibrium. If a
solution holds more salt than corresponds with the normal solubility curve of the
salt, the solution is said to be supersaturated. Although the solution can be kept
an indefinite time in this condition, the seeding or inoculation of a supersaturated
solution by the introduction of a very minute quantity of a crystal of the dissolved salt
will upset the state of apparent equilibrium. According to W. Ostwald,i as little as
10-8 gram of salol suffices to start the crystallization of undercooled salol, and with
sodium chlorate, lO~io gram is needed. The crystal fragment becomes the centre
or nucleus from which crystals radiate into the solution on all sides. Similar results
can be obtained with aqueous solutions of sodium acetate, sodium chlorate, etc.
The following illustrative experiment is due to G. R. Robertson:
A solution of 5 grins, of benzil in hot alcohol is filtered while hot into a 250 c.c. flask,
heated to boiling, and set aside in a warm place to cool. The flask is fanned so as to cool
the glass walls, condense alcohol on the sides, and wash down any benzil into the body of
the liquid and so prevent marginal crystallization. The liquid can thus be cooled to 15°
without crystallization. A minute fragment of benzil is then dropped into the centre
of the flask, and a complex of lemon yellow crystals spreads radially through the mass of
liquid. The experiment can be adapted to lantern.
According to A. L. Potilitzin (1893), salts forming hydrates which have a consider-
able dissociation pressure in dry air at ordinary temperatures usually form supersatu-
rated solutions readily ; while salts forming hydrates which do not readily dissociate
in dry air or in vacuo do not usually form supersaturated solutions so readily.
Calcium sulphate, CaS04.2H20,is an exception, for it does not lose water at ordinary
temperatures, and it forms supersaturated solutions.
In 1795,i J. T. Lowitz 2 found that any crystal will not do for the inoculation.
Thus, if a crystal of nitre be introduced into a mixed solution of nitre and Glauber's
salt, prepared hot, and subsequently cooled, the nitre alone crystallizes out, while
if the solution be seeded with Glauber's salt, the latter alone crystallizes from the
solution ; and D. Gernez tried the action of 220 different substances on supersaturated
solutions of Glauber's salts, and found 39 to be active stimulants ; 18 of the 39
substances were insoluble, and lost their activity after washing with water, and
drying while protected from dust, and the remaining 11 substances lost their property
when purified by recrystallization. Hence, D. Gernez assumed that the 39 sub-
stances which had inaugurated the crystallization of sodium sulphate all contained
this salt as an impurity. In H. A. Miers' experiments on the crystallization of
the organic compounds, salol and betol, it was found that the substances did not
crystallize at first when allowed to cool in open vessels in the laboratory ; but after
a time, when the air of the laboratory had become impregnated with dust, presumably
containing minute grains of both substances, crystallization readily occurred in open
vessels exposed in the laboratory. The fragment of crystal used for seeding must
be either a fragment of the same salt as that in solution or of an isomorphous salt.
The particular salt which separates is to some extent determined by the nature
of the inoculating salt. For example, J. T. Lowitz found that a crystal of potassium
nitrate introduced into a supersaturated solution of both potassium nitrate and
452 INORGANIC AND THEORETICAL CHEMISTRY
sodium sulphate was followed by the separation of nitre alone, while a crystal of
sodium sulphate in a similar solution gave a crop of crystals of sodium sulphate
alone. If fragments of both salts were added to a similar solution, crystals of both salts
were simultaneously deposited. L. de Boisbaudran 3 found the addition of copper
sulphate to solutions of nickel sulphate gave short thick prisms of nickel sulphate,
while magnesium sulphate or ordinary nickel sulphate gave fine needle-like crystals.
The crystallization of a supersaturated solution is not always induced by the addition
of isomorphous substances, as L. de Boisbaudran, C. Tomlinson, and others have
assumed from the fact that a supersaturated solution of nickel sulphate commenced
to crystallize by contact with zinc sulphate.
It is possible that in some cases of seeding by isomorphous salts, the effect is due
to the presence, as impurity, of the salt to be crystallized. Thus, N. Dhar (1916)
showed that no change is induced in solutions of copper sulphate by the addition of
crystalline sulphate of magnesium, manganese, iron, cobalt, zinc, or cadmium ;
sodium selenate has no effect on supersaturated solutions of sodium sulphate, stron-
tium chloride on solutions of calcium chloride, or sodium nitrate on solutions of silver
nitrate. This shows that the induced crystallization of supersaturated solutions
is not a sufficient test for isomorphous substances. Similarly, a solution may be super-
saturated with respect to the hydrate of one salt and not another. Thus, C. E.
Linebarger (1893) showed that at 10° it is possible to prepare four different solutions
of manganous sulphate saturated respectively with the hexa-, penta-, tetra-, and tri-
hydrates. For example, at 10°, the solubility is
MnS04.6H20 MnS04.5H20 M11SO4.4H2O MnSOi.SHaO
Parts MnS04 per 100 of water 71 68 64 61
If a crystal of one of the three lower hydrates be added to a saturated solution of
the hexahydrate a separation of the crystals of the lower hydrate will occur. Thus,
by adding a pentahydrate crystal, the corresponding crystals will be deposited,
more hexahydrate would pass into solution, and be deposited in turn as pentahydrate.
This will continue until all the hexahydrate has been transformed into the penta-
hydrate. A supersaturated solution of ammonia alum, Al2(S04)3.(NH4)2S04.24H20,
will deposit the same salt if sown with crystal fragments, but according to A. Piccini
and V. Fortini (1902), if sown with fragments of octahydrated ammonium thallic
alum, Tl2(S04)3.(NH4)2S04.8H20, crystals of Al2(S04)2.(NH4)2S04.8H20, not
Al2(S04).(NH4)2S04.24H20, separate.
It is possible to distinguish between a saturated and a supersaturated solution
by bringing each in contact with more of the solid. If the solution is unsaturated,
more soUd will dissolve ; if saturated, none will dissolve ; and if supersaturated,
solid will separate until the solution is saturated. The concentration of an unsatu-
rated solution is less, while the concentration of a supersaturated solution is greater
than that of a saturated solution.
Related phenomena. — Many other examples of related phenomena are known.
In analytical work the slow appearance of precipitates in dilute solutions is very
common. Pure water may be easily cooled to — 3° or — 4° without the appearance
of ice if kept quite still while the temperature is reduced ; and the water can be easily
cooled to —6° or —7° if a layer of oil be placed over the surface of the cooling water.
The undercoohng of water was observed by D. G. Fahrenheit ^ in 1724, and the
undercooUng of freezing mercury by T. Hutchins in 1783. The vapour pressure of
liquid water from about 30° to —10° is represented by the curve QOR, Fig. 9 ; if
the water freezes at 0°, the vapour pressure curve of the sohd from 0° to —10° is
given by the curve OP. In the former case the curve QO does not show a break or
abrupt change of direction at 0, and in the latter case it does. Phosphorus, sulphur,
etc., behave in a similar manner. W. C. Roberts-Austen 5 measured the under-
cooUng of gold, copper, and some other metals, and A, D. van Riemsdyk showed that
the sudden flashing of gold beads during cupellation is due to the crystallization of
an undercooled liquid. The melting points of solids usually appear somewhat
WATER
453
higher than their freezing point. Thus, sodium hydroxide is said to melt at 310°,
and to solidify at 290°. The phenomenon is attributed to supercooling carrying the
observed freezing point below its true value, or to a slight lagging in the speed of
the change.
Ice has not been heated above 0° without melting, but liquid water can be heated
to 105° or 106° without boiling. When the boiling does start, it proceeds with almost
explosive violence. The phenomenon is called bumping. In 1772, J. A. de Luc
noted that the bubbles of air which develop in a liquid while it is being heated
serve as nuclei for the generation of the bubbles of vapour formed when the liquid
boils, and he found a liquid free from dissolved air could be heated to 130° without
boiling. F. Donny heated water to 137° without boiling, and observations in
the same direction were made by D. Gernez, G. Krebs, W. E. Grove, etc. By sus-
pending drops of water in a mixture of olive and linseed oils — which has the same
specific gravity as water, and a high boiling point — L. Dufour (1863) raised water
to 178° without boiling. P. J. Coulier (1875) found that dust-free air saturated with
moisture may be cooled below the normal temperature of condensation ; and John
Aitken ^ (1880) showed that dust is necessary for the formation of fogs and rain-
drops, so that in perfectly clean dust-free air, aqueous vapour does not condense,
and mist does not form. Without solid nuclei cooling vapours may become
supersaturated.
The vapour pressure of small drops o! liquid.— There is an exception to the
general observation that at any given temperature the vapour pressure of a given
liquid is always the same whatever be the mass of the liquid. In 1870, Lord Kelvin
(W. Thomson) 7 showed that the
vapour pressure of a liquid with a
concave surface must be less than
that of the same liquid with a
plane surface. If a capillary tube,
A, Fig. 13, dips in water confined
in a closed vessel, it follows that the
vapour pressure of the liquid at a '^
must be less than that of the Fig 12.
liquid at h ; and that the vapour
pressure j)q at h must be equal to the vapour pressure pata plus a pressure h equiva-
lent to the weight of a column of the vapour of height ah and the same- sectional
area as the bore of the capillary tube. For equilibrium, po=p-{-h, otherwise there
would be a perpetual circulation of the liquid owing to distillation from a to 6 or
conversely, while the height of the liquid in the capillary remained constant. The
converse of the above can be extended to convex surfaces. The vapour pressure
of a minute spherical drop of liquid (convex surface) must be greater than that of
larger masses of liquid with approximately plane surface.
If Pq denotes the vapour pressure of a liquid with a plane surface, and p its vapour pressure
for a convex surface with a radius of curvature r, then, if S denotes the specific gravity of
the liquid, and s that of the vapour, and a- the surface tension (or pressure) of the liquid,
Lord Kelvin showed that with common logarithms, log {pQ/p)=2<T8l2'3rpQS. For example,
with water, <r = 80 ; 8, 0*00081 at 0° and 760 mm.; p^, 1014000 ; *S is unity, log{po/p)
= 0'56xl0~ Jr. When r is large, say 10~Ho 10~*cm., the ratio po/p is nearly unity, and
only when r approximates to the millionth of a centimetre will the vapour pressure of a liquid
be sensibly greater than that of a plane surface. The formula was deduced from thermo-
dynamics by E. Warburg and R. von Helmholtz; and from the molecular theory by G. F.
Fitzgerald, J. Stefan, and B. Galitzine. The vapour pressure of an electrically charged
surface was studied by R. Blondlot, N. Schiller, A. Gouy, and W. Kistjakowsky ; and the
effect of a magnetic field has been studied by P. Duhem and J. Konigsberger.
The kinetic theory interprets the phenomenon by showing that the inter-attrac-
tion of the molecules of a liquid on a molecule partially immersed, as illustrated by
the dotted circle, B, Fig. 12, will be less than on a molecule similarly situated with
454
INORGANIC AND THEORETICAL CHEMISTRY
respect to a concave surface, A, Fig. 12 ; and greater than a molecule similarly
situated in a convex surface, C, Fig. 12. The differences in these magnitudes is
illustrated by the blackened portions of A and C, Fig. 12, and it will be obvious
without a mathematical demonstration. The greater the mole-
cular attraction on the partially immersed molecules the less the
vapour pressure of the liquid, and the less the tendency to evapora-
tion ; the smaller the drop of liquid the more convex the surface,
and the greater the tendency to evaporation. Accordingly, in a
given space the larger drops of liquid will grow at the expense of
the small ones. Hence, a space saturated for a liquid with a plane
surface is not saturated for minute drops, and this explains the
observation that it is difficult for small drops of vapour to form in
a space supersaturated for a plane but not for a convex surface.
j,jQ JO If small drops were momentarily formed they would at once evaporate.
(Diagram- If dust particles be present the water will first condense upon them,
matic). and the liquid spread out on them will have a large radius of
curvature so that re-evaporation will be comparativelv slow, and
the liquid has time to evaporate — assuming the nearer the ratio Pq/j) approaches
unity, the slower the evaporation.
Metastable and labile states of supersaturation. — Inoculation or seeding is usually
necessary to start the process of crystallization of a supersaturated solution ; and
yet the supersaturation may be carried so far that the crystals will grow spontaneously
in the solution without seeding. Indeed, it is possible to draw a supersaturated
solubility curve representing the concentration of a solution at different tempera-
tures where the supersaturation is so great
that crystallization will begin spontaneously
without inoculation. The idea is illustrated
in Fig. 14, where the region between the
normal solubility curve and the supersolu-
bility curve represents what W. Ostwald ^
called the metastable state, where inocula-
tion is necessary to inaugurate the process
of crystallization ; and the region beyond
this, below the freezing point, represents the
so-called labile or unstable state where
crystallization may start spontaneously with-
out inoculation. E.g. a solution of sodium
nitrate at 20° is saturated when it contains
Fig. 14.— Labile and Metastable Equi- f^'S per cent, of the salt, and it is labile when
libria of Saturated Solutions (alter it contains over 48-8 per cent. ; between these
H. A. Miers). two concentrations the solution is metastable.
Similarly, solutions of sodium bromide saturated
with the hydrate, NaBr.2H20, at 30°, are in a labile condition, and crystallize
spontaneously at temperatures between 5°- and 16°; they are in a metastable
condition above 19°.
The existence of the metastable and labile states was predicted by W. Ostwald
in 1897, and demonstrated experimentally by H. A. Miers ^ and his co-workers in
1906. Can a metastable liquid be made to crystallize in any other way than by
introducing a solid crystal ? Ordinary shaking, scratching,"^ and the like fail to
provide the necessary stimuli. According to H. A. Miers,
^5 50 55
Concentration
If the growth of a crystal is really the coming together of vibrating particles which cohere
with one another ... is it not possible that we may be able to communicate these vibrations
to a supersaturated solution, which is so densely crowded that it is ready to crystallize, by
some other means than by inoculating it with an appropriate crystal ? . . . Some knowledge
of these movements may be obtained by studying the sort of shock or movement, if there be
any such, which starts crystallization in supersaturated solutions.
WATER
455
i
S. W. Young did succeed in making water and metastable salt solutions crystallize
by applying mechanical shocks of sufficient intensity-mechanical hammers striking
on metallic anvils. For instance, water was made to freeze at — 0'02° without the
addition of ice, with repeated blows of sufficient intensity. H. A. Miers and
F. Isaac crystallized water at — 1*9° without seeding, and H. Hartley and N. G.
Thomas observed the formation of ice at — 0'5° without seeding.
Supersaturation and the phase rule. — The phase rule, it will be observed,
applies to systems in real equilibrium, not to systems in a state of apparent, false,
or metastable equilibrium. We are repeatedly confronted with those little-under-
stood phenomena which, for convenience, have been grouped under the general
term passive resistance.
The kinetic theory of supersaturation. — The kinetic theory throws a little light
on the phenomenon of supersaturation. A saturated solution in contact with the
solid is supposed to be closely analogous with a closed vessel containing a liquid in
contact with its vapour. When in equilibrium, the same number of molecules pass
from the surface of the solid into the solution and return from the liquid to the
surface of the solid. If the state of equilibrium be disturbed by evaporation or by
lowering the temperature, the equality of the two opposing actions is disturbed and
a new condition of equilibrium is established. In the case of a metastable super-
saturated solution, the exchange of molecules cannot take place because no free
solid is present. Directly a particle of the same substance as the dissolved solid
is added, the dissolved substance is
rapidly deposited about the submerged § ^"^
particle as a nucleus until the concentra-
tion of the solution has reached its
normal value. In the case of a labile
supersaturated solution, the crowding
of the molecules is so great that they
are able to form aggregates large
enough to serve as nuclei about which
the crystals can grow.i^
The speed of crystallization. —
Within certain limits of temperature,
the speed of crystallization is greater
the lower the temperature.^^ This might be expected if it be assumed that the
lower the temperature, the smaller the speed of molecular motion, and the less
the probability of a molecule escaping from the sphere of action of a growing
crystal ; but there is an influence retarding crystallization, for the slower the
molecular motion, the greater the viscosity, and the smaller the number of
molecules . which travel into the sphere of action of a growing crystal. Instead of
the speed of crystallization increasing continuously as the temperature is lowered,
it reaches a maximum value and then progressively diminishes as the temperature
falls, presumably because the viscosity of the groundmass is so great that it is superior
to the vectorial forces which arrange the structural units into crystal forms, and
completely inhibits their action.
20
40
80
60"
Temperature.
Fig. 15. — The Effect of Temperature on
Speed of Crystallization of Piperine.
100
the
Observations on the influence of temperature on the speed of crystallization can be illus-
trated by fusing hippuric acid (melting point 188°) in a dish ; and making a number of
capillary tubes— about 15 cm, long and 1 mm. bore- — by drawing out a test-tube in the
usual manner of making melting-point tubes. The molten acid is sucked into a warm tube,
and immediately cooled under the water-tap. The surfused acid congeals to a glass-like
mass as is demonstrated by breaking one of the tubes. If the tubes are warmed, say, in
the hot air over a Bunsen's flame, crystallization immediately sets in, although the tubes
may be preserved at atmospheric temperatures for some days. The speed of crystallization
can be determined by exposing the surfused compound for, say, four minutes to the tempera-
ture in question, and then counting the number of centres of crystallization- — the greater
the number of crystal nuclei, the faster the crystallization. The maximum speed with
hippuric acid is attained at about 100°.
456 INORGANIC AND THEORETICAL CHEMISTRY
G. Tammaim's observations (1898) on the rate of crystallization of surfused
pipeline, melting at about 129°, show that the speed increases with decreasing
temperatures down to about 40° ; after that, the rate decreases as the temperature
falls. The maximum speed is at about 40°. The speed of crystallization of under-
cooled water has been measured 12 by undercooling water and starting crystallization
by introducing a fragment of the solid. The time necessary for the crystals to fill a
certain length of the tube gives the linear velocity of crystallization. J. H. Walton
and R. C. Judd's values for the linear velocity of crystallization of water in cm. per
minute, for water in a tube (12 mm. outside diameter, and 7 mm. internal diameter),
were
-2-0° -3-61° -4-67'^ -5-86° -6-18° -7-10° -8-19° -907°
Velocity. 31-6 48-4 71-4 1071 114-7 2667 4152 684 0 cm. min.
The internal diameter, and the thickness of the walls of the tube have a marked
influence on the results since they determine the rate of cooling of the crystallizing
liquid. Spontaneous crystallization of the undercooled liquid prevented measure-
ments being conducted at lower temperatures, and the temperature of maximum
velocity of crystallization has not been determined. J. H. Walton and A. Brann
measured the effect of forty-five substances on the velocity of crystallization of water
supercooled to —9°, and found that all retarded the speed. The retardation is a
colligative property. For substances with over eight atoms per molecule, the
greater the number of atoms the slower the rate of crystallization ; for substances
with less than eight atoms per molecule, the power of retardation is a specific property.
In dilute solutions, substances which undergo the greatest hydration in solution have
the greatest effect in reducing the speed of crystallization ; and this is explained by
assuming that if the formation of ice crystals is due to a change of the type, 3(H20)2
^2(H20)3, any part of the solvent would have to be decomposed before crystals
could be produced. H. T. Barnes has measured the rate of growth of the ice
mantle in R. Bunsen's ice calorimeter. Similar phenomena are shown by glasses
and pottery glazes, which are really congealed surfused liquids. There is a special
range of temperature peculiar to each surfused compound which is particularly
favourable to rapid crystallization — zone of rapid crystallization. This is
illustrated in Fig. 15.
References.
^ W. Ostwald, Zeif. phys. Chem., 22. 289, 1897; Lehrbiich der aU^emeinen Chemte, Leipzig, 2.
ii, 740, 1903; G. R. Robertson, School Science, 19. 4S1, 1919; A. L. Potilitzin, Journ. Russian
Phys. Chem. Soc, 25. 73, 1893.
2 C. L. Berthollet, Essai de statique chimiqae, Paris, 1. 32, 1903 ; J. B. Ziz, Schweigger's Journ.y
15. 160, 1815 ; J. S. C. Schweigger, ib., 9. 79, 1913 ; J. L. Gay Lussac, Ann. Chim. Phys., (1), 87.
255, 1813 ; (2), 11. 296, 1819 ; L. Pasteur, ib., (3), 44. 5, 1862 ; H. Lowel, ib., (3), 29. 62, 1850 ; (3),
33. 334, 1851 ; (3), 37. 155, 1853 ; (3), 43. 405, 1855 ; (3), 44. 313, 1855 ; (3), 49. 32, 1857 ; Compt.
Rend., 33. 10, 1851 ; 34. 642, 1852 ; C. Violette, ib., 60. 83 J , 1865 ; 76. 171 , 713, 1873 ; M. Goskynsky,
ib., 32. 717, 1851 ; F. Selmi, ib., 32. 909, 1851 ; Atti Accad. Torino, (2), 11. 325, 1851 ; A. Lieben,
Sitzber. Akad. Wien, 12. 771, 1854 ; L. Pfaundler, ib., 72. 61, 707, 1875 ; H. Schroder and T. von
Dusch, Liebigs Ann., 89. 232, 1853 ; H. Schroder, ib., 109. 35, 1859 ; C. S. Reischauer, ib., 115.
116, 1860 ; F. Zwig and 0. Hecht, ib., 233. 166, 1886 ; D. Gernez, Compt. Rend., 60. 833, 1027, 1865 ;
61. 71, 289, 847, 1865; 63. 843, 1866; 75. 1705, 1872; 76. 566, 1873; 78. 68, 283, 1874; 79.
802, 912, 1074, 1332, 1874 ; 84. 1389, 1877 ; A. Terreil, ib., 51. 506, 1860 ; A. Jeannel, ib., 60,
412, 1865 ; 62. 37, 1866 ; A. P. Dubrunfaut, i6.,68. 916, 1218, 1869 ; L. de Boisbaudran, 68. 1052,
1329, 1869 ; 80. 888, 1007, 1450, 1875 ; E. Lefebore, ib., 70. 684, 1870 ; H. le Chatelier, ib., 96.
715, 1056, 1883 ; F. Margueritte, ib., 68. 1110, 1329, 1869 ; L. G. de Coppet, ib., 73. 1324, 1871 ;
78. 498, 1874; 76. 434, 1873; Bull. Soc. Chim., (1), 17. 146, 1872; Ann. Chim. Phys., (5), 6.
275, 1875; (4), 26. 539, 1872; L. de Boisbaudran, t6., (4), 9. 173,1866; (4), 18. 246, 1869;
A. Recoura, ib., (7), 4. 494, 1895 ; J. 0. G. de Marignac, ib., (5), 1274, 1874 ; C. Tomhnson, Phil.
Tram., 158. 652, 1868 ; 160. 51, 1870 ; M. Faraday, ib., 124. 55, 1834 ; Quart. Journ. Science,
19. 153, 1825 ; A. Liversidge, Phil. Mag., (4), 45. 67, 1873 ; J. G. Greenfell, Proc. Roy. Soc, 25.
124, 1876 ; J. M. Thomson, Zeit. Kryst., 6. 94, 1881 ; W. R. Whitney, Zeit. phys. Chem., 20. 40,
1896 ; T. Graham, Trans. Roy. Soc. Edin., 11. 175, 1831 ; H. Ogden, Edin. New Phil. Journ., 13. 309,
1832; J. T. Lowitz, CrelVs Ann., 1. 3, J795 ; D. Gemez, Compt. Rend., 60. 833, 1865.
» L. de Boisbaudran, Compt. Rend., 63. 95, 1866 ; Ann. Chim. Phys., (4), 9 173, 1866 ; (4),
WATER 457
18. 246, 1869 ; C. Tomlinson and G. van der Mensbruggbe, Phil. Mag., (4), 44. 223, 1872; C. E.
Lineberger, Amer. Chem. Journ,, 15. 225, 1893 ; A. Piccini and V. Fortini, Zeit. anorg, chem.^ 31.
451, 1902; N. Dhar, Proc. Akad. Armterdam-, 18. 1084, 1916.
4 D. G. Fahrenheit, Phil. Trans., 39. 78, 1724 ; T. Hutchins, ih., 73. 303, 1783 ; H. Cavendish,
ih., 73. 30.-], 1783.
5 W. C. Roberts-Austen, Proc. Roy. Soc, 63. 447, 1898 ; A. D. van Riemsdyk, Ann. Chim.
Phys., (5), 20. 66, 1880.
« J. Aitken, Nature, 23. 195, 384, 1881 ; R. von Helmholtz, Wied. Ann., 27. 520, 1806; J. A.
de Luc, Eecherches sur les modifications de Vatinosphere, Geneve, 4. 209, 1772 ; P. J. Coulier, Jowrw.
Pharm. Chim., (4), 22. 165, 254, 1875 ; L. Dufour, Ann. Chim. Phys., (3), 68. 370, 1863 ; F. Donny,
ih., (3), 16. 167, 1844; D. Gernez, ih., (5), 4. 335, 1875 ; G. Krebs, Pogg. Ann., 133. 673, 1868;
136. 144, 1869; 138. 489, 1869; W. R. Grove, Cosmos, 22. 698, 1863.
' Lord Kelvin (W. Thomson), Proc. Roy. Soc. Edin., 7. 63, 1870 ; Phil. Mag., (4), 42. 448, 1871 ;
E. Warburg, Wied. Ann., 28. 394, 1886; M. Cantor, ih., 56. 492, 1895; J. Konigsberger, ih., 66.
709, 1898; R. von Helmholtz, ih., 27. 522, 1886; J. Stefan, ih., 29. 655, 1886; B. Galitzine, i6.,
35. 200, 1888; N. Schiller, ih., 53, 396, 1894; 60. 755, 1897; Jmirn. Russ. Phys. Chem. iSfoc.,29.
7, 1897 ; 30. 79, 175, 1898 ; W. Kistjakowsky, ih., 29. 273, 1897 ; 30. 139, 1S98 ; G. F. Fitzgerald,
Phil. Mag., (5), 8. 382, 1879; Nature, 49. 316, 1894; A. Bacon, Phys. Rev., (1), 20. 1, 1903;
R. Blondlot, Journ. Phys., (2), 3. 442, 1884; A. Gouy, Compt. Rend., 149. 822,1909; P. Duhem,
Sur les dissolutions dhm sel magnetique, Paris, 1890; Des corps diamagnetique, Lille, 1889.
8 W. Ostwald, Zeit. phys. Chem., 22. 289, 1897.
» H. A. Miers and F. Isaac, Journ. Chem. Soc, 89. 413, 1906 ; B. A. Rep., 522, 1906 ; Proc.
Roy. Soc, 79. 322, 1907 ; H. A. Miers and J. Chevalier, Min. Mag., 14. 123, 1906 ; T. V. Barker,
ib., 14. 235, 1907 ; Journ. Chem. Soc, 89. 1120, 1906 ; H. Hartley and N. G. Thomas, ih., 89.
1013, 1906 ; H. A. Miers, Science Progress, 2. 12], 1907 ; The Growth of Crystals, London, 1911 ;
S. W. Young, Journ. Amer. Chem. Soc, 33. 148, 1911 ; S. W. Young and R. J. Cross, ib., 33. 1375
1911 ; S. W. Young and W. van Sicklen, ib., 35. 1067, 1913.
i» L. C. de Coppet, Ann. Chim. Phys., (8), 10. 457, 1907.
^' G. Tammann, Kristallisieren und Schmelzen, Leipzig, 1903 ; R. Marc, Zeit. phys. Chem., 61
385, 1908; 67. 470, 1909; 68. 104, 1909; 73. 685, 1910; 75. 710, 1911; H. Freundlich, ib.,
75. 245, 1911 ; M. Padoa, Atti Accad. Lincei, (5), 27. ii, 59, 1918 ; Gazz. Chim. Ital, 48. ii, 139,
1918.
12 0. Tumlirz, Sitzher. Akad. Wien, 103. 226, 1894; J. H. Walton and R. C. Judd, Journ.
Phys. Chem., 18. 722, 1914 ; H. T. Barnes, Ice Formation, New York, 90, 1906 ; J. H. Walton
and A. Brann, Journ. Amer. Chem. Soc, 38. 317, 1161, 1916; A. Brann, ?6., 40. 1168,1918;
J. J. Czochralsky, Zeit. phys. Chem., 92. 219, 1917 ; R. Nacken, Centr. Min., 191, 1917.
§ 7. The AUotropic Forms of Water
Water is water, not a single substance in the proper acceptation of the term, but a mush
or mixture ; an entirely peculiar material, not to be represented by any one formula, nor
spoken of by any single name.- — H. E. Armstrong (1913).
The vapour pressure curve of a substance comes to an abrupt end at the critical
point, for the liquid ceases to exist. In Clapeyron's equation
at the critical temperature, dv is zero, and consequently also the latent heat of
vaporization becomes zero.
Just as a liquid, at a constant pressure, when subjected to a constantly decreasing
temperature, solidifies at a definite temperature called the freezing temperature,
so will a liquid, at a constant temperature, when subjected to a constantly increasing
pressure, solidify at a definite pressure— the freezing pressure.^ For example, the
freezing temperature of water is 0° under a pressure of one atmosphere, and at -|-1°
the freezing pressure is 7600 atm. The curve of freezing pressure, dT/dp, is convex
towards the pressure axis corresponding with the fact that each successive increment
of pressure produces a smaller and smaller effect. Every substance has its own
peculiar dT/dp-cuTve, but usually the slopes of the curves for different substances
vary within comparatively narrow limits — less than 50° per 1000 atm.
The question has been discussed : What is likely to be the result of extrapolating
the observed results for the effect of very high pressures on the course of the melting-
458
INORGANIC AND THEORETICAL CHEMISTRY
So/ id
Liquid.
P
Fig 16.
point curve 'i J. H. Poynting and W. Ostwald 2 argue that, when the pressure is great
enough, the melting-point curve of a solid will have a critical end-point, analogous
to the critical temperature of a liquid. W. Ostwald further suggests that liquid
crystals are solids near their critical points at ordinary pressures ; and that the
effect of pressure is to diminish the vectorial properties of crystals
and reduce all matter to an isotropic or amorphous condition. If
there be such a critical end-point, dv and A of Clapeyron's equation
should vanish simultaneously, but there is no sign of this in any
substance which has been investigated at high pressures — in some
cases A increases rather than diminishes with increasing pressures.
G. Tammann, M. Planck, and H. W. B. Roozeboom^ have also raised objections to
this hypothesis. G. Tammann assumes that if the pressure be great enough, all solids
can be melted no matter what the temperature. G. Tammann assumes that the
pressure-volume curves of liquid and solid will probably cross when dv=0 as indi-
cated in Fig. 16 ; and under that condition, the latent heat will
change its sign and be no longer positive, but negative. Under
these assumed conditions, the melting point with an increase of
pressure will first rise, and then fall; and at the turning point, the
melting-point curve will be a maximum. G. Tammann, there-
fore, represents the melting-point curve as shown in Fig. 17.
Inside the curve, the crystalUne solid is stable ; outside, the Uquid
is the stable form. It is, however, not necessary to develop the
idea further because P. W. Bridgman's work would probably
have shown some signs in favour of Tammann' s assumptions if
they were valid at the enormous pressures under which he worked. Although the
value of dv becomes smaller and smaller as the pressure increases, there is no sign
of dv approaching zero.
There is, therefore, no satisfactory evidence that the melting temperature of a
solid will pass through a maximum (G. Tammann) or terminate at a critical end-point
(J. H. Poynting) as the pressure increases, but the evidence rather indicates that
the melting temperature will rise indefinitely
with increasing pressure, or else the solid will
change abruptly into another allotropic form.
For example, liquid water under very great
pressures, at a constant temperature, exhibits
some peculiarities which show that the water
then freezes, even though the temperature is
above the normal freezing point, 0°. The melt-
ing point of ice is reduced bypressure, and ice
can be melted by increasing the pressure. At
temperatures below —22°, however, ice cannot
be melted by an augmented pressure, for the ice
passes into another variety more dense than
water. The denser ice changes back into ordi-
^^-"^y ^^®' specifically lighter than water, when
fto'-eo^-AOj-ao" 0° 20° 40° 60° 80° the pressure is removed. G. Tammann (1910) and
/ empera^ure3. p ^ Bridgmau (1912) ^ havc studied the effects
Vn\7L'li^um^'Z^^Co^^^^^^ -I great pressures on the properties of ice, and al-
for the Varieties of Ice. though the two mvestigators do not agree m cer-
tain details, they are agreed on the main points.
There are at least four kinds of ice more dense than water. Suppose that ordi-
nary ice — called ice I — be maintained at —10° and gradually compressed to about
1000 kilograms per sq. cm., the ice melts to water. The relation between the freezing
point and pressure is given by the curve ON, Fig. 18. When the pressure rises to
about 4400 kgrm. the liquid freezes to a form of ice denser than water — called ice V —
and at about 6300 kgrm. pressure, ice V changes to ailother still denser variety —
WATER
459
called ice VI. Again, if the temperature be maintained at —30°, and the experi-
ment repeated, ordinary ice, ice I, changes into another variety — called ice DI — at
a pressure of 2200 kgrm. At —25°, this change is sudden and abrupt, sometimes
it takes place with a kind of click. At about 3000 kgrm. ice III changes into another
variety — called ice 11— and with a further application of pressure, ice II passes into
ice V, and finally into ice VI. The relations between the temperature and pressure
of the five different forms of ice are illustrated in Fig. 18.
In practice, the water is compressed in a suitable cyhnder by means of a piston
worked by a hydraulic press. At any given temperature, the pressure, measured
by a manometer, increases regularly with the displacement of the piston representing
the volume of the substance, as shown by AB, Fig. 19 ; when the pressure has attained
a certain critical value, P, the volume suddenly decreases, BC, without the pressure
rising at all. Afterwards the pressure resumes its regular rise, CD,
with the displacement. The pressure at which the piston drops
abruptly into the cylinder without producing a rise of pressure is
the pressure at which the water freezes at the temperature of the
experiment. The pressure at which water freezes is different at
ever}- different temperature. A series of pressure-temperature
curves, like Fig. 18, is taken to represent the melting curves of
ice to water, or transition curves of one form to another. Given
the displacement of the cylinder, and the density of the water, the density of the
ice can be calculated ; and if the temperature and pressure at which the ice melts,
and the change in volume which simultaneously occurs are known, the latent heat
of fusion of the ice can be computed. All the experiments show is that at certain
temperatures and pressures there is an abrupt change in volume. It is inferred
that the abrupt change of pressure must be due to a change in the molecular
structure of the liquid, either a change of liquid to solid, or from one liquid to
another. The latter possibility is rejected because no substance is known with
two liquid modifications, and in some cases the solid is stable enough to allow
a momentary glimpse to be obtained when the pressure is suddenly released. The
estimated densities of the different varieties of ice (water unity) are :
Ice I. Ice II. Ice III. Ice IV. Ice V. Ice VI.
Density . .0-92 1*03 1-04 — 1-06 1-09
The latent heats of fusion of the different modifications of ice to water are not
very different from that of ordinary ice ; and there is therefore very little heat
involved in the transformation of one variety of ice into another. The latent heats
of transformation of the different forms of ice into one another and to liquid water
at the triple points are indicated in Table XII, along with the volume changes which
occur, and the co-ordinates of the triple points.
Table XII.— Properties of the Different Forms of Ice at the Triple Points.
(L denotes liquid water ; Roman numbers a particular form of ice.)
Co-ordinates of triple point.
The upper line within the brackets represents the change in
Triple point.
Temp. °C.
Press, kgrm.
per sq. cm.
volume dv in c.c. per gram, and the lower line, the latent
heat of transformation in gram calories.
Ill, L, I
II, III, I
V, III, L
V, II, III
VI, V, L
-22-0
-34-7
-17-0
-24-3
+0-16
2115
2170
3530
3510
6380
TTT .T (00466
III->L |_^5o.9
j_^^ 01352
"^ \ — 56-1
TTT_^T /0-0215
TTT^T /00241
III->L l^gj.^
ii-^ni l^'^i^^
v->^ \ + 70-l
III-^I {'1\%'
TT_^T /0'2178
Ar_^T i0'0788
460 INOROAMU ASU TlihUKbiiuAL tllKMlbTUy
It will bo obi^fved tluifc ice VI U Bteblc above 0*, and with tho aj)pliculi()n of
2ii,(iOO Igmn. proMure, water can he frozen under ffre<U prenMurr turn thouffh it he nmrly
fU the bailiruf point of toater under ordinary frmiuree I If tho proMMjrc* n j)on ordinary
ice rineK much above 2000 kgnn. the ioe dianget into ice 1 1 1 wliich han a Ichh volume,
hence, 2000 kgrm. per »q. cm. *' iit the highent preiwurc which can be obtained by
freezing water in a doaed Hpace." Aji soon aM the prcKHurc im removed, th'i different
varieties of ice revert to ordinary ice. If ice II or ice III be at the temperature of
liquid air when the prManre is relieved, the change from the un>}table to the stable
form b slow enougfi to demonstrate that the ice w reall v a solid denser than ordinary
ice. A variety of ice— called ioe IV — was announced ny G. Tammann, but haA not
been confirmed by P. W. Bridgman.
TIm motooaUur formnU of Wtier. — ^Thc vapour d(>nrtity of Ht<!«m \» slightly
greater than is required for the molecuUr form^iia H./), and mur^h too small for
n^Of, It is therefore assumed that steam is a mixture of IImO molecules with a few
H4O2 molecules, and it is found that the equilibrium condition for water vapour
in the vicinity of lOO'', namely, 'il^Ofw^2l\iO, correspomis with about 91 per rent .
of H2O molecules. According to E. Hose (1908),^ measurements of the vapour
density of saturated water vapour under the preHsure of its own vapour at the
corresponding tempemtures, show that :
Temperature
0*
W)"
lOO**
IW)*
200"
PraMuro ....
4-6
92-2
im
:J5HI
DOHH mm.
Per oAnxt. H,0 moleoules
. 93-4
91-8
911
9 10
91 '3
An increase in the proportion of the H^O molcculcH with a riHe of temperature is
eounUtrbalanced by the converse effect of the increaned pre^Hure. According to
Q. Oddo (1910), under normal pressure below ZT^ some water vapour i^ dihHociated
into ions, II^O^H'-f OH^ and the vapour pressure is almost that required for
The estimated diameter of a molecule of water vapour is 4*54 X 10 * cm. ; the
mean free path, 40X10 • cm. at 0' and im mu\. ■ the collision frequency in
7'JH^)xlO^ per c.c. per sec.; the molecular velocity at 0" iH r>(\(\ cm. i)er h<m;. ;
and the values of J. D. vander Wm,W a«00ll7:j, and h -0 (X)ir)l. I.ord Kelvin «
estimated the mean distance between the centrr^n of eontiguoun wnUtr moleeulen
probably lies between lO"*' and 6x10 '^' cui. If the lower limit be uHcd it. J. S
tiand has shown that there are lO^^ molecules per c.c, or IHxlO^^ molecules oi
water ner gram-molecule of the liquid ; thin number in regarded as obsolete, and
U. J. o. Sand obtained S'6xW^ as a better approximation for what \h analogoun
to Avogadro'i constant for liquid water.
Xi<11lid water* —The formuU of water \h ho frequently n;pr<;iS(;nted })y H./), that
it is easy to acquire the belief that this Hymbol correctly repreH<»ntH the mohsculo. of
the liquid. The molecule of liquid water iH irnuh rrjore complex. A compari-son
of the boiling points of the hvdrides of flilorin**, IK.'I, Hulphur, ifjS, nitrogen, NJI3,
and carbon, CH4,
My'li<.;<<i) >»,ii.., llydioK II Ammonia, Mwihnni*,
'\A<>i\<\", iH\ 11,0 tni\it\M'\ M,- NH, CII,
Boiling point . ht i ioo" ~«2 'Mir,'' 104''
Critical temperature ^^ 360" +100" 4 130 90"
has led to the argunwrit th.it if w;it«r were reprenented by the formula H.^O, and
remained in that condition during its eondenifation from the gamtouM (nteam) to the
liquid state, it would probably fall in line with the hydriden junt indieai«*d, and be a
gas at onlinary temperatureH ; that i«, iuHtead of boiling at 1(K)", it would boil at a
much lower temperature. The boiling pointn of li<|ui(U are raised by nHHOcriation,
and a coninarison of the boiling points of water with tho«e f)f Hk* corrcMponding
hydridfts of sulphur, U^Pt, Helenium, lljSe, and tellurium, Il2'r«'
WnijfT, I(yrlr<itf<in llyilrr/K<<ti liy<>r<ii(i<n
il,(> »mI|)IiI'I«, H,H nelKiMtt, H,H« t^clliiridr, U^Ttf
Boiling point . 1100' 02' -42^ o"*
WATER
461
led H. M. Vernon (1891) ^ to infer that the molecule of liquid water is very complex.
J. H. van't HofE (1900) calculated that if water were not associated it would boil at
—207°. P. Walden (1900) also argued that the substitution of oxygen by sulphur
raises the boiling point of the methyl and ethyl compounds by about 60°, and if a
similar difference prevailed with the hydrogen compounds, water should boil at
-120°. Thus,
Methy 1 sulphide, (CH 3)28 37° Ethyl sulphide, (CgH J gS 91° Hydrogen sulphidejHjS -61^
Methyl oxide, (CH 3)20 -23*' Ethyl oxide, (CgH J gO . 36° Water, (H2O) . -120°
Difference
60°
Difference
56=
Difference
59°
H. M. Vernon's deduction is con^med by the lowering of the freezing points of
solutions of water in several other solvents ^ which point to a higher molecular
weight for the liquid than is required by the simple formula H2O ; thus, in ^-toluidine
the molecular weightvaries from 28"7 to 333, although in phenol and related solvents,
acetic acid, liquid hydrogen cyanide, etc., the result corresponds with a molecule
H2O. It is, however, remarkable that so many lines of evidence all converge about
Kontgen's assumption that water contains polymerized molecules, e.g. the thermal
expansion, compressibility, surface tension, viscosity, specific heat, index of refrac-
tion, dielectric constant, magnetization, Trouton's rule, J. D. van der Waals' deduc-
tions, boiling points, molecular volumes, etc. J. Thomsen ^ also assumed that water
has a molecular weight corresponding with (H20)2, be- ^- 5Q
cause the heat developed during the hydration of some -^
salts agrees with the assumption that the water frequently -c" ^^
enters into combination in pairs, (H20)2. >^ 30
W. Sutherland's hypothesis of the constitution o£ Uquid
water.— W. Sutherland (1900) i^^ calls W. C. Kontgen's ice
molecules trihydrol on the assumption that molecular S 10
weight corresponds with (H20)3 or HqO-^ and the water oS
molecules dihydrol, (H20)2 or H4O2 ; the steam molecules
are called hydrol H9O. Alternative terms have been sug-
20
20 40 60 80 100
Temperaf-ures.
gested : trioxylene for trihydrol, and dioxylene for dihydrol. Fig. 20. — Estimated Pro-
Instead of the curve for the variation of the volume of """"
portion of Trihydrol m
, .,, , , ■c.-r • A •• Water at Different
water with temperature — ±ig. 5 — sweeping round a mini- Temperatures.
mum at about 4°, W. Sutherland makes the line continuous
and thus extrapolates a value 1*089 for the density of dihydrol at 0°. This agrees
with values deduced in a similar manner from other physical properties of water.
Similarly he gets the value 0-88 for the density of trihydrol at 0°. Hence, it follows
from the observed density of water at 0°, that 37*5 per cent, of trihydrol and 62*5
per cent, of dihydrol are present. The percentage amounts of hydrol in water at
different temperatures, estimated from the observed densities at pressures of one and
50 atmospheres, are indicated in Fig. 20. It is further estimated that at 2300
atmospheres pressure there will be no trihydrol in water at 0°. W. Sutherland also
estimates that dihydrol and trihydrol have the following physical properties at 0° :
Table XIII. — Comparison of the Supposed Physical Properties of Dihydrol and
Trihydrol.
Dihydrol
Trihydrol
Density
atO°.
1-089
0-88
Compressi-
bility per
atmosphere.
j 0-000016
! 0-000010
Surface
tension.
78-3
73-3
368=
538=
I^atent heat
calories.
Critical i Specific viscositv
temp. heatO". Viscosity. lyaporiza-
1 Fusion. I ^n.
257
250
0-8
0-6
0-0030
0-0381
16
A. Piccard estimated that in liquid water at 0° there is 291 per cent, of ice ; and in
water at 100°, one per cent, of ice. With this assumption, and his observations on
462 INORGANIC AND THEORETICAL CHEMISTRY
the refractory power of water, C. Cheneveau estimates that hydrol at 0° has a
specific refraction by J. H. Gladstone and Dale's formula, of 03228 ; and by Lorenz
and Lorentz's formula, of 02049 ; the corresponding numbers for hydrol at 100^ are
0'3319 and 0'2058. From his experiments on the evaporation and condensation of
small drops, C. Barus concludes that water contains different constituents, and that
one is more volatile than the other. G. Gillet ascribes all the reactions of water to
the hydrol molecules, and he says that a substance is soluble in water because it
forms a soluble compound with the hydrol, and this disturbs the equilibrium (H20)2
^2H20, and that the evolution or absorption of heat in the process of solution is
dependent on this reaction. H. Schade believes that the equilibrium constant
wH20^(H20),» varies with temperature, and that the polymerized molecules are
present in water in a colloidal form.
J. W. Briihl considered that water is an unsaturated compound with quadrivalent
oxygen H— 0— H. W. Sutherland suggests that, on J. W. Briihl's assumption
that oxygen is quadrivalent, the dihydrol and trihydrol molecules are constituted
as indicated by the graphic formulae :
H>0 H>0=0<H H,=OvO=H,
H2
Hj^drol. Dihydrol. Triliydrol.
R. de Forcrand believes that the ice molecules are either (H20)2 or (H20)2.5, and he
assumed that oxygen is quadrivalent, and the hydrogen bivalent. As a result, he
obtained the graphic formula :
In an attempt to account for the properties of dilute solutions, H. E. Armstrong
(1908) further assumed that some of these molecules have the same composition
but a different structure : thus, the dihydrol molecules, H4O2, are assumed to exist
in the ti^o different forms :
jj>0-0<jj H>0<H
Dihydrone. Hydronol.
The former kind are said to be inactive molecules because they are formed by direct
association unaccompanied by rearrangement ; and the latter are said to be active
molecules. When non-electrolytes of the type RX are dissolved in water, active
complexes of the hydronol type HO — RX — H and inactive hydrone complexes of the
type RX=0H2, as well as polymerides, are supposed to be formed ; while with
electrolytes, the compound becomes hjdrolated by the formation of {a) hydronol
complexes HO— RX— H, and (6) complexes of the type X— H2O— R, which are distri-
buted in the solvent. Such solutions, says Armstrong, will be electrolyzed when the
composite molecules a and h act reciprocally on one another under the influence of
an electric strain.
If these views about the constitution of water be well-founded, and if these differ-
ent varieties of ice really exist, it is hardly correct, without some reservation, to say
that the passage of ice to liquid water and to steam, and the converse changes,
are purely physical changes. P. W, Bridgman appears to think that the different
forms of ice arise from a difference in the alignment or arrangement of the molecules
whereby each molecule preserves its individuality so that the changes undergone by
the different forms of ice are purely physical. In physical processes, the molecules
of the substance are supposed to remain intact and unaltered ; in chemical processes,
on the contrary, the molecules are altered. Is the magnetization of iron a physical
or a chemical process ? A chemical process if the molecules are changed ; and
WATER 463
physical, if the molecules are not changed. Selecting one of the many definitions
of chemical action, say H. E. Armstrong's (1885),
Chemical action may be defined as being any action of which the consequence is an altera-
tion in molecular constitution or composition ; the action may concern molecules which are
only of one kind — cases of mere decomposition, or isomeric change, and of polymerization ;
or it may take place bietween dissimilar molecules— changes of combination and inter-
change.
With this definition, the vaporization of ethyl alcohol would be a chemical process
because it is probable that there is a simultaneous depolymerization (C2H50H)^
^nC2H50H, while the vaporization of carbon disulphide would be a physical process,
because, so far as is known, the molecules are not altered in constitution or composition.
In this sense, vaporization would be sometimes a chemical and sometimes a physical
process. Indeed, the niere raising of the temperature of water involves a
change in the composition of the molecules, and is accordingly a chemical change
(depolymerization).
References.
1 J. Johnston, Journ. Franklinjnst., 183. 1, 1918.
2 J. H. Poynting, Phil. Mag., (5), 12. 32, 1881 ; W. Ostwald, Lehrbuch der allgemeinen Chemie,
Leipzig, 2. ii, 373, 391, 1902.
3 G. Tammann, Kristallisieren und Schmehen, Leipzig, 29, 1903 ; M. Planck, Wied. Ann.,
15. 446, 1882 ; H. W. B. Roozeboom, Die Heterogenen Gleichgewichte vom Standpunkte der
PJuisenlehre, Braunschweig, 1. 91, 1901; J. Johnston, Journ. Franklin Inst, ISZ. 1, 1918;
P. Duhem, Archiv. Neerland., (2), 6. 93, 1901.
* G. Tammann, Zeit. phys. Chem., 72. 609, 1910 ; P. W. Bridgman, Proc. Amer. Acad., 47.
441, 1912 ; Zeit. anorg. Chem., 77. 377, 1912 ; Journ. Franklin Inst., 178. 315, 1914.
« E. Bose, Zeit. Elektrochem., 14. 269, 1908 ; G. Oddo, Gazz. Chim. ItaL, 45. i, 319, 395, 1915.
« K. Kundt and E. Warburg, Pogg. Ann., 155. 544, 1875 ; J. H. Jeans, The Dynamical
Theory of Gases, Cambridge 341, 1916 ; 0. E. Meyer, The Kinetic Theory of Gases, London,
193, 1899 ; Lord Kelvin (W. Thompson), Mem. Manchester Lit. Phil. Soc, 9. 136, 1870 ; Nature,
1. 551, 1870 ; H. J. S. Sand, Trans. Faraday Soc, 15. 94, 1919.
7 H. M. Vernon, Phil. Mag., (5), 31. 387, 1891 ; Chem. News, 64. 54, 1891 ; J. H. van't HofF,
Vorlesungen uber theoretische und physikalische Chemie, Braunschweig, 3. 45, 1900 ; P. Walden,
Zeit. phys. Chem., 66. 385, 1909 ; W. Herz, Molekulargrosse der Korper im festen und fiussigen
Zustand, Stuttgart, 1899; R. de Forcrand, Journ. Chim. Phys., 15. 617, 1917.
8 J. Walker, Zeit. pMj.^. Chem., 5. 193, 1890 ; J. F. Eykmann, ib., 4. 497, 1889 ; R. Lespieau,
Compt. Rend., 140. 855, 1905 ; H. C. Jones and G. Murray, Amer. Chem. Journ., 30. 193, 1903.
9 J. Thomsen, Ber., 18. 1088, 1885 ; P. T. Walden, Zeit. phys. Chem., 65. 129, 257, 1908
66. 385, 1909 ; H. Gaudechon, Compt. Rend., 156. 1872, 1913 ; C. Ch6neveau, ib., 156. 1972, 1913
J. Duclaux, Jonrn. Chim. Phys., 10. 73, 1912.
10 W. C. Rontgen, Wied. Ann., 45. 91, 1892 ; W. Sutherland, Phil. Mag., (5), 50. 460, 1900
W. Vaubel, Zeit. angew. Chem., 15. 395, 1902 ; I. Traube, Ann. Phys., (4), 8. 267 1902 ; Dis
cussion in Trans. Faraday Soc, 6. 71, 1910 ; W. R. Bousfield, Zeit. phys. Chem., 53. 302, 1905
J. Duclaux, Monit. Scient., 75. 555, 1911 ; Journ. Chim. Phys., 10. 73, 1912 ; L. Schemes, Ann
Physik, (4), 38. 830, 1912 ; C. Cheneveau, Compt. Rend., 156. 1972, 1913 ; R. de Forcrand, ib., 140
764, 1905 ; C. Barus, Amer. Journ. Science, (4), 25. 409, 1908 ; C. Gillet, Bull. Soc Chim. Belgigue
26. 415, 1893 ; H. Schade, Zeit. Kolloid, 7. 26, 1910; J. W. BrWhl, Ber., 28. 2847, 1895 ; 30. 162,
1897 ; H. E. Armstrong, Proc Roy. Soc, 78. A, 264, 1906 ; Chem. News, 103. 97, 109, 1911.
§ 8. The Physical Properties of Water
Liquid water freezes at 0° into crystalline ice while water vapour freezes directly
into hoar frost, and snow. If water be carefully cooled in a dust-free space it can be
reduced to —12° without freezing,^ and H. C. Sorby cooled water to —13° in a
capillary tube 2*5 mm. diameter, and to —15° in a tube of 0*0127 mm. diameter.
According to J. Thomson, the freezing point of water is lowered 0*00757i° by a pressure
of n atmospheres.
Ice appears to be colourless and transparent when pure, but in reahty it is pale
blue when seen in large masses. Artificially formed ice is often white and more or
less opaque. According to R. Pictet, if artificial ice be slowly formed between 0°
464
INORGANIC AND THEORETICAL CHEMISTRY
and —1-5°, it is as transparent as natural ice ; but if the temperature at which
the freezing occurs be below — 3°, the ice appears to be more or less opaque and white,
and of low specific gravity, and this the more the lower the freezing temperature.
The whiteness and opacity is due to the presence of small air bubbles, 001 to 0-5 mm.
diameter, mechanically entangled among the elementary crystals during freezing.
If precautions be taken to use air-free water, water freezes at very low temperatures
to transparent ice. The bubbles of air are developed owing to the reduction in the
solubiUty of the air dissolved in the water as the water freezes into ice.
The formation of sheet ice on the surface of an expanse of quiet water as cold weather
approaches is interesting. So long as the temperature of water is above 4°, convection
currents help to keep a uniform distribution of the temperature, for the cold and denser water
slowly sinks, and the warmer layers rise to the surface. As the temperature of the surface
falls below 4°, the colder layer remains on the surface and finally reaches the freezing point.
Long needles of ice then shoot out from the borders over the top of the water and the crystals
ramify outwards until the whole surface is covered with a thin sheet of ice. The sheet
of ice then gradually thickens by the conduction of heat through the ice. According to
G. Quincke's hypothesis, a
freezing liquid is regarded as
a liquid jelly which forms in-
visible foam cells containing
water. The lower the tem-
perature, the greater the vis-
cosity of the liquid in the
walls and interior of the foam
cells. In streams which run
too swiftly for the border ice
to meet, the so-called frazil-
ice or slush-ice is produced
on the surface, but it cannot
remain attached and freezes
to a continuous sheet. In
special circumstances where
the bottom of a river can be
cooled by the radiation of
heat, the so-called anchor ice,
or ground ice, bottom ice —
gldce-du-fondf or Grundeis- — ■
may form on the bottom of a
river or stream.
Well-defined crystals of
ice are extremely rare and
difficult to measure. Ac-
cording to A. E. von Nor-
denskiold, F. Rinne, and
A. B. Dobrovolsky, the
bipyramidal crystals of ice belong to the hexagonal system and appear as ditrigonal
prisms or plates ; and P. Groth gives the axial ratio a : c=l : 1617 ; but this, how-
ever, is somewhat doubtful. With hail, combinations of the rhombohedron have been
reported. Sea-ice is usually a complex of crystal particles with the chief axes at
right angles to the surface. The crystals can be often seen when a piece of ice is
examined with a lens while a beam of bright fight is passing through it. Snow
crystals are common ; this is readily demonstrated when snow-flakes are examined
under a low-power microscope. Over a thousand different patterns have been
noted, but all appear in the form of hexagonal (six-sided) nuclei, or six-rayed stars
with the rays developed in bewildering complexity — some are rounded, others
serrated, others reticulated — but all are of inimitable delicacy and beauty. A
rough idea of the form of snow crystals can be obtained from Fig. 21, by
G. Hellmann.2 No two crystals seem to be alike, yet there is no deviation from
the hexagonal style of architecture.
Albertus Magnus, of the thirteenth century, is said to have been the first writer to
mention the Jigura stellce, the hexagonal form of snow crystals, and in 1611, J. Kepler,
Snow Crystals. — G. Hellmann.
WATER 465
the celebrated astronomer, wrote a pamphlet on six- rayed snow. J. Kepler wais
greatly impressed by the beauty and regularity of the shapes of the snow stars ;
but he perforce left unanswered the obvious question : Why are the crystals six-
rayed ? Why does nature unsparingly fashion such strange contrasts, all built
according to one definite type — the six-rayed star — each
Frail, but. a work divine, made so fairily well.
So exquisite, minute, a miracle of design ?
Tennyson
Few indeed would deny what an anonymous writer has said : " The chemist is
assured that if he could wholly understand a drop of water, he would know the origin
and destiny of all things, and hold the key to every happening ; " the same remark
would apply to a drop of any other liquid. According to M. Trautz, Pontus knew
in 1833 that if water be rapidly frozen, it sends out bright flashes of light-— crystallo-
luminescence. The X-ray structure of ice has been investigated by A. St. John.
The viscosity of water at 0° — that is, the resistance which water offers to flow
— is here given along with a few other liquids for comparison :
Water. Mercury. Sulphuric acid. Alcohol. Ether. Benzene. Carbon disulphide.
00178 0-0169 0-3195 00184 00029 0-0089 0-0044
The viscosity of water diminishes with rise of temperature. Thus, G. Zemplen and
B. Pogany found 0*010562 at 18°, and, according to R. Hosking and E. C. Bingham
and G. F. White,
0° 20° 40° 60" 80° 100° 124° 153"
Viscosity 0-017928 0-01002 0-006563 0*004730 0-003570 0-00284 0*00223 0-00181
R. Hosking recommends the interpolation formula, 7)=r]Q(l-\-ad-\-hd^), for the
viscosity r) at 6° when the viscosity at 0° is tjq, and when a and h are constants.
E. C. Bingham defines the fluidity of a liquid as the reciprocal of the viscosity, and
he represents the fluidity j/r as a function of the absolute temperature T such that
T=0-23275^-8676-8/(j^+120)+309-17. G. F. White and R. H. Twing find the
viscosity of undercooled water at — 4'7°, —7*23°, and — 9*30° to be respectively
0-02121, 0-02341, and 0-02549.
Water at high pressures is less viscous than at normal pressures provided the
temperature does not exceed 36°, and this the more the lower the temperature and
the lower the pressure. According to L. Hauser (1900), ^ the percentage changes
in the viscosity coefiicient of water at 400 atmospheres pressure are :
Temperature . . .90°
Percentage change of viscosity +3*4
Most other liquids which have been tried become more viscous under an increasing
pressure. This also is in harmony with the assumption that with water the smaller
the pressure, or the higher the temperature, the less the proportion of ice mole-
cules transformed into less complex molecules, highly viscous ice molecules are
replaced by less viscous water molecules. These facts were predicted by W. C. Ront-
gen (1891) from the hypothesis just outlined, and confirmed by R. Cohen (1891).
The viscosity of some aqueous solutions is less than that of water itself owing to
the fact that the solute converts enough viscous ice molecules into less viscous water
molecules to more than compensate for the increase of viscosity which its own
presence imparts. With aqueous solutions of urethane, there is a steady increase
in the viscosity with rising concentration, probably because the increase in the
viscosity produced by the solute more than compensates the decrease due to the
diminishing concentration of the ice molecules. The coefficient of viscosity of
water vapour ^ at about 20° is 00000975. F. Houdaille found the coefficient of
the viscosity of water vapour to be smaller at low pressures, possibly as a
consequence of dissociation.
The viscosity of ice is enormously greater than that of water, and it depends on the
VOL. I. 2 H
70°
5r
40°
35°
29°
18°
+ 2-5
+ 1-6
+0-7
0
-0-3
-1-6
466
INORGANIC AND THEORETICAL CHEMISTRY
direction of the crystal axes. The viscosity of ice has been studied by B. Weinberg,
J. F. Main, etc. R. M. Deeley & found that the viscosity at 0° is 2 X lO^o in a direction
perpendicular to the optic axes ; while both he and P. H. Parr calculated the viscosity
of glacier ice moving in winter to be between 125 X 10^2 and 147-7 x lO^^. The difier-
ence is taken to mean that in glaciers, the optic axes of the ice crystals are inclined
in different directions. B. Weinberg gives for the relation between the viscosity
rj and the absolute temperature, 77X10-13=1-244— 0-502T+0-0355T2. H. Hess
(1902) found that a bar cut from glacier ice is slowly deflected when loaded in the
middle and supported at both ends. The rate of change varies with the load.
When the bar is relieved from the load it slowly recovers, due to what H. Hess
regards as a kind of residual elasticity. J. C. McConnel (1891) also showed that a bar
of ordinary ice yields continuously to pressure or tension, but if cut from a single
crystal, with its length at right angles to the optic axis, it shows no sign of stretching
under tension, or yielding to pressure. The crystal is brittle. Consequently, the
bending of a bar of ice does not represent a gradual shearing of the ice crj^stals,
but the slipping of a number of layers of finite thickness. This is the probable cause
of the so-called plasticity of ice. The recovery of bent ice after the stress is relieved
is out of all proportion to known effects in other substances, and is attributed to the
slipping back of the forcibly displaced sliding layers. H. Moseley (1871) found Young's
modulus for ice to be 92,700 kgrms. per sq. cm., and B. Weinberg (1905), 5x10^
kgrms. per cm. at 1°. H. Hess cut bars of ice with the crystal axis parallel to the
width, length, and thickness — 1*2 cm. X2"5 cm. and 4 to 16 cm. long — and measured
the modulus of elasticity E ; the bending moment, B grm. cms. ; and the vis-
cosity 7] with different loads as indicated in Table XIV. H. Reusch, R. Koch, and
R. Trowbridge have published values for the elasticity constant. H. Hess concludes
that for moderate loads the coefficient of viscosity r] increases with the duration of
Table XIV.-
—Elasticity and Viscosity
OF Ice
Axis parallel to
Axis parallel to
Axis parallel to
length.
width.
thickness.
Load In gnns.->
2000
5000
6000
1000
1500
2000
3000
1000
1500
2000
B
1350
3400
4000
1500
2250
3000
4450
1600
2350
3100
^xio-i° .
0-54
0-70
0-75
3-5
3-0
2-9
4-0
1-6
2-0
20
iyl5xl0-i» .
6-5
10-5
0-55
3-7
3-7
2-4
11-0
7-5
10-0
8-0
7y60xl0-i» .
17-5
11-5
3-6
8-0
11-0
6-0
90
7-5
11-0
7-0
7^120 X 10-10
10-0
13-5
3-65
120
10-0
10-0
^_.
7-5
9-0
11-0
lySOO X 10-10 ,
110
16-6
3-5
21-0
19-0
17-6
• — ■
8-0
12-0
120
the experiment from 15 to 60 to 120 to 300 seconds, and even after only 300 sees.
the increase is nearly proportional to the time. For large loads near the point of
rupture, the coefficient of viscosity decreases with the duration of the experiment.
H. Hess also measured the relation of pressure to the speed of the flow of ice,
and found that with ice confined in cylinders, the flow increased rapidly with increas-
ing pressure, and when the flow was once started, comparatively small pressures
were required to maintain the flow. G. Tammann (1902) and N. Slatowratsky
(1905) also studied the velocity of flow of ice, and showed that the plasticity of ice
is relatively small, but increases rapidly near the melting point. The results with
the pressures expressed in kgrm. per sq. cm. are shown in Table XV. J. Dewar
(1905) pressed ice into wire-like threads at —80° and 50 tons per sq. in. pressure,
but at lower temperatures he did not succeed in doing so.
T. Andrews (1886) measured the hardness of ice in terms of the depth of penetra-
tion of a steel rod into a cylinder of ice at different temperatures. The results showed
that ice remains " almost impenetrable " from about —37° to about —12° ; its
power of resistance then decreases rapidly to about —7°, and still more rapidly at
WATER
467
higher temperatures, until, at the melting point, the ice gives way " almost entirely,"
when it becomes very soft indeed. Ice on Mohs' scale (resistance to scratching) is
said to have a hardness of 1'5. H. Morphy says the coefficient of friction of ice for
small pressures — up to 14'3 grms. — between —5° and —6° is nearly constant, being
0"36±001 ; and for large pressures— above 15 grms. — 0*17 ±0*01.
Table XV.— Plasticity of Ice.
Temperature.
Highest pressme of
steady flow.
Pressure when rapid
increase occurs.
Melting pressure.
- 5-7°
-10-7"
-15-7°
-21-7°
-27-6°
642
1116
1611
2000
2220
665
1130
1729
2100
2240
678
1225
1681
2070
The surface tension of water is higher than that of all the common liquids,
excepting mercury. For example :
Mercury.
547
Water.
75-0
Ammonia.
64-7
Benzene.
29-2
Acetone.
23-3
Alcohol.
22-0
Ether.
16'5 dynes per cm.
Surface tension and specific gravity determine the height to which a liquid will rise
in a capillary tube. The high surface tension of water plays an important role in
determining the ascent of this fluid in the capillary pores of the soil. Under ordinary
conditions it is estimated that water can rise four or five feet under the influence of
its high surface tension ; if the surface tension of water were like that of most liquids
the liquid would rise but two or three feet. ^ The high surf ace tension of water thus
becomes an important factor in bringing water within reach of plants.
The reported values ^ for the surface tension, or, of water in moist air at ordinary
temperatures range from 7*13 to 7*945 mgrm. per mm. N. Bohr found the surface
tension of water at 12° to be o-=73"23 dynes per cm. or cr=7'465 mgrm. per mm.,
and the specific cohesion to be 14*96 sq. mm. Many other determinations of these
constants have been made by W. Kamsay and J. Shields. The following values
below 40° are by P. Volkmann, and above 40° by C. Brunner :
0^
10"
20°
30°
40°
50°
60°
70°
80°
7-692
7-541
7-389
7-237
7-086
6-91
6-73
6-54
6-35 mgrm. per mm
75-49
74-01
72-53
71-03
69-54
67-8
66-0
64-2
62*5 dynes per cm.
15-406
15105
14-821
14-556
14-295
13-99
13-70
13-39
13-08 sq. mm.
According to H. Sentis, the surface tension of water at d° is (r=76*09(l— 0*002026^) ;
and according to C. Forch o-=o-o(l— 0*00190179^— 0*0000024991^2). The surface
tension of water is extremely sensitive towards traces of impurities, and it is
very difficult to get quite clean surfaces. G. Quincke (1870), F. von Lerch (1902),
and F. Pockels (1899) have measured the surface tension in dynes per cm. at 20°
at the interface of water in contact with different immiscible or partially misciblc
liquids, and their results include
Mercury.
Chloroform.
Carbon
disulphide.
Alcohol.
Ethyl
ether.
Benzene.
Petroleum.
Turpen-
tine.
Olive
oil.
<r = 37-47
2-68
4-122
0-206
1-23
3-365
3-81
1-254
1-81
G. Hagen noted in 1845 that the surface tension of water is gradually reduced by
exposure to air; G. Quincke, P. Volkmann, E. Bonicke, A. Kundt, and C. Forch found
that from 17° to 18°, for solutions with m gram-molecules of gas per litre, the corre-
sponding change in the surface tension of water per gram-molecule of dissolved gas
is Ao-/o-m :
COa NgO HaS Oj Ng Air
m . . . . 0-041 0-029 0-127 0-00146 000073 000087
A(r/«rm . . . —0-310 -0-307 -0-214 -2-225 -088 -1-30
468 INORGANIC AND THEORETICAL CHEMISTRY
At 17°, in vacuo, that is in saturated vapour, the surface tension of water is
about Oil per cent, or 0008 mgrm. per mm. greater than in air at atmospheric
pressure.
T. W. Richards and J. H. Mathews found that the compressibiUty )3 and the
surface tension a of pure liquids are related approximately as ^o-S=a constant,
while A. Ritzel could detect no simple relation with mixtures of water with other
liquids. W. C. Rontgen and J. Schneider found that aqueous solutions of
inorganic salts had a higher surface tension and a lower compressibility than pure
water ; while G. de Metz showed that cane sugar as solute always reduces the com-
pressibility of water, the surface tension of these solutions is sometimes greater
and sometimes less than water. K. Drucker found aqueous solutions of some
organic acids have smaller surface tensions than water, and that the compressibiUty
at first decreases and then increases with rising concentration. The surface tension
of aqueous solutions shows that there is a possible action of the solute in dissociating
some of the complex ice molecules in the surface film. T. W. Richards and
S. Pahtzsch observed that with aqueous solutions of urethane, there is a rapid
decrease in surface tension with concentrations up to 40 per cent, of urethane,
presumably owing to a decrease in the proportion of ice molecules with greater
concentrations, the surface tension decreases in accord with the rule j8o-*=a
constant.
G. Tammann denied the existence of any relation between the internal pressure
of a solution — calculated from the thermal expansion — and the surface tension,
but W. C. McC. Lewis found the internal pressure and surface tension of all but
volatile and colloidal solutions change in the same direction. I. Traube based
an explanation of some properties of solutions on the relationship between internal
pressure and surface tension ; T. W. Richards also interprets the compressibility
of a pure substance as being in part contingent on the internal pressure, for the
external pressure required to compress a substance to given extent is greater the
more the molecules are previously compressed by molecular pressure.
The specific cohesion a^ sq. mm. of water at different temperatures in moist
air, is
0° 10" 20" 30'-' 40° 50^ 60° 70° 80"
o2 . 15-406 15-105 14-821 14-556 14-295 13-99 13-70 1339 13-08
It will be remembered that the specific cohesion a^, is related with the surface tension
so that (T=ia^D—D'), where D and D' respectively denote the specific gravities
of water and moist air. According to L. Weinstein, the value a^ at a temperature
d between 0° and 95° is a2=i4.987(i_o-001458^).
R. Eotvos (1886) found that the variations of the molecular surface tension
(T=a{Mv)^, with changes of temperature 6, namely, da-ldO, are nearly the same, 2*12,
for all normal liquids, but not so with liquids whose molecular complexity changes
with temperature. W. Ramsay and J. Shields (1893) obtained the following values
for water when the molecular weight is assumed to be 18 :
0 . 0° to 20° 20° to 40° 40° to 60° 60° to 80° 80° to 100° 100° to 120° 120° to 140°
S<T/h9 0-88 0-95 1-00 105 1-09 1*14 MS
If the molecular weight of water at 0° be taken to be 3 X 18, and the density of ice be
assumed to represent the density of ice moleculesy the constant becomes 2*08, very
close to that for a normal liquid. Hence, the surface tension of water at 0° is
sufficient to change practically the whole of the water in the surface film to ice mole-
cules with a molecular weight 3x18, corresponding with (H20)3. It also follows
that the surface film of water is not changed very much with temperature up to about
40°, but, at higher temperatures, the surface tension is sufficiently reduced to form
appreciable amounts of water molecules less complex than ice molecules. By
applying the law for the surface tension of mixtures, and assuming that the surface
film contains a mixture of the molecules with a molecular weight 3 X 18, and water
molecules with a molecular weight 2x18, W. Sutherland calculated values for- the
p
WATER
469
variation of the surface tension of water in dynes per cm. with temperature
which were in close agreement with the observed results, and with R. Eotvos'
theory :
Surface tension .
0°
73-32
20°
70-56
40°
67-55
60°
64-27
80°
60-77
100°
57-11
120°
53-30
140°
49-38
The coefficients of diffusion of water vapour into hydrogen, air, and oxygen,
at 0°, are respectively O'BST, 0-193, and O'lSl per second.^ The velocity of sound
in water vapour is, according to A. Masson, 401 metres per second at 0° ; according
to W. Jager, 402 4 and 4100 metres per second respectively at 93° and 96° ; and
according to W. Treitz, 413 metres per second at 110° ; 417*5 at 120° ; and 424-4
at 130°.
Specific heat. — The specific heat or the amount of heat required to raise the
temperature of one gram of liquid water per degree under certain assigned con-
ditions, is taken unity as a standard — e.g. the zero calorie is the quantity of heat
referred to water between 0° and 1° ; the mean calorie, to water between 0° and
100° divided by 100 ; and the 15° -calorie, to water between 14-5° and 15*5°.
The specific heat of water, as unity, is here given along with the specific heats of a
few other liquids for comparison :
Water.
1
Mercury.
0-0334
Sulphuric acid.
0-317
Alcohol.
0-547
Ether.
0-529
Benzene.
0-397
Carbon disulphide.
0-235
The specific heat of water is abnormally high ; and it is remarkably nearly constant
over a comparatively large range of temperature. From the time of H. V. Regnault
(1847) ^ up to the present the specific heat of water, at constant pressure, has been
the subject of investigation with more and more refined attempts to increase the
degree of accuracy. Some of the later determinations almost agree up to the
third significant figure. C. Dieterici's results (with the specific heat at 15°
unity) are :
0°
1-0088
10°
1-0021
20°
0-9987
30°
0-9984
40°
0-9987
50°
0-9996
1-0008
80°
1-0045
100°
1-0099
25° of
to 35°.
0-9983. The
C. Dieterici's
rising to 1*1543 at 300°. There is a minimum near
results of other observers give minima ranging from 12*
formula, referred to water unity at 0° for the
specific heat Cp of liquid water at 0°, between
35° and 300° is 0^=0-99827—0-00016368^
+0-0000020736^2, h. L. Callendar gives for the
specific heat C of water between 0° and 20°,
0=0-9982+0-0000045(^-40)2+0-0000005(20-^)3,
and between 20° and 60° the last term is omitted.
Accurate measurements have also been made by
C. E. Guillaume, A. Cotty, W. R. and W. E. Bous-
field, and W. Jager and H. von Steinwehr. J.
Narbutt claims that for 6° between 0° and 100°
the best observations are represented by the
formula 0=1-00733-0-0007416^+0-000016845^2
—0-000000095520^ when 0 for 15° is unity ; this
gives a minimum between 20° and 30°. According
to H. T. Barnes, the specific heat of undercooled water rises to 1-0155 at —5° (water
at 15° unity). This is illustrated in Fig. 22. The specific heat of water at constant
volume is obtained by computation from the expression Cv=Cp—9a^TvQlP,
where ^ represents the coefficient of compressibility, and a the coefficient of thermal
expansion ; T the absolute temperature ; and Vq the volume at 0°. When Cp
at 0° is 1-0000, Cv is 0-9995. The specific heat of water at 0° is altered —0-0001025
per atmosphere increase of pressure. The specific heat of a solid is usually less, but
|-U£
'
~
1
101
\
^
\
y
\
.-^
/
100
\
y
\
y"
y
cr
.^
—
_
_
_
■ID 0 10 20 30 40 50 60 70 80 90100 110
Fig.
22.— The Specific Heat of
Liquid Water.
470 INORGANIC AND THEORETICAL CHEMISTRY
sometimes greater, than that of a liquid, with water the difEerence is abnormally
great. For instance,
Water.
Lead.
Mercury.
Sulphur.
Specific heat, solid
. 0-502 (0°)
0-314 (0°)
00319 (-40°)
0-2026 (100°)
Specific heat, liquid
. 1-000 (0°)
00402 (356°)
0-0333 (0°)
0-234 (120°)
J. Dewar found that the specific heat of ice falls from 0502 at 0° to 0463 at —78°,
and to 0'146 at —252 "5°. W. A. Smith found the specific heat of highly purified ice to
be almost constant up to a temperature close to zero, but there is a sensible increase
in the specific heat of ordinary ice, owing, it is supposed, to incipient fusion caused
by the lowering of the melting point by dissolved impurities. H. C. Dickinson and
N. S. Osborne found the specific heat of ice for a temperature 6 between —40° and
-0-05° to be O=0-5057+0-001863^-79-75^-2 cals. (20°), when the constant
was found to diminish from — 0*00125 to — 0*00005 with increasing purity. Hence
it was inferred that the departure of the specific heat of ice from a linear function
of the temperature is less the purer the ice, consequently, the specific heat of pure
ice is assumed to be C=0-5057+0-001863^. F. G. Jackson obtained between 0°
and -78-4°, 0*424 ±0-002; and between 0° and -188°, 0*337 ± 0*001 ; P. Nord-
meyer and A. Bernoulli obtained analogous results ; and A. Bogojawlensky
worked with 5° intervals between —15° (0*500) and —50° (0*395). According to
W. Nernstand F. Koref, the relation between the molecular heat of ice and tempera-
ture is represented by the relations 8*47+0*0276^—14*0^-1.
The sudden rise in the specific heat of a substance just below its fusion tempera-
ture has given rise to some discussion. In some cases this is due to the presence of
an impurity. For example, if a salt be present as impurity in water then, above the
eutectic point for that salt, the apparent specific heat is increased by the fact that
some of the ice melts to produce a larger specific heat than that of solid ice. The
rise in the specific heat cannot be explained in this way ; silver iodide shows a
similar phenomenon, although it may be that difierent allotropic forms are present,
and that with crystals of a one-phase substance, the phenomenon would not
occur. G. N. Lewis and G. E. Gibson i^ found that with the exception of a small
variation of ice below the melting point, the specific heat of ice can be represented
by log Ct,=0*43(log T— 2*51) ; accordingly the entropy of ice at 0°, calculated
from absolute zero, <ji=zjVj^d log T, is 9*96 per gram-molecule; the entropy for
the liquefaction of ice is 1*58, at 0° ; and the increase of entropy of liquid
water from 0° to 25° is 1*58. Hence, the entropy of water at 25° is 16*8 per
gram-molecule.
The mean specific heat of water vapour between 100° and 800° and at a constant
pressure was found by L. Holborn and F. Henning n to be 0^=0*4460(1+0*000096^)
cals., and by A. Langen, for temperatures above 1100°, Cj,=0*44(l+0*00027^).
The specific heat of water vapour at constant volume Cv, and at the absolute tempera-
ture T, according to W. Nernst and H. von Wartenberg, is (7^=5*61 +0000717T
+0-06312T2 cals., or according to M. Pier, between 0° and ^°, 6-065+0*0005^
+0*0382^2^ from 1300° to 2500°.
The minimum in the specific heat curve of water is near 30°. W. Jager and
H. von Steinwehr give 33*5°. This, and other abnormal phenomena,, are in agreement
with the assumption that the observed specific heat of water is a complicated pheno-
menon involving both a true specific heat and an endothermal change of ice into water
molecules on a rising temperature. If the presence of a substance in solution
reduces the proportion of ice molecules in the liquid, it follows that the specific
heat of an aqueous solution will be smaller than that of the pure solvent under the
same conditions. This agrees with the observed facts, even when due allowance is
made for the specific heat of the solute — that is, the dissolved substance ; thus,
A. Jaquerod (1901) found for solutions of potassium chloride, KCl, at about 16°,
Percent. KCl ... 0 2-4 48 9-6 19-2 28-8
Specific heat . . . 1-000 0-968 0938 0882 0-790 0-720
WATER 471
K. Puschl (1901) has also shown that the specific heats of many aqueous solutions
are less than that which would be the case if solvent and solute were in the free
state.
According to E. Mallard and H. le Chatelier (1881),i2 the molecular specific heat
of water vapour at a constant volume, 0^, at a temperature 6, is Cy=5*91-|-O'OO3760
—O'OqISS^^ j and according to J. M. Gray, the specific heat at a constant pressure
is 03787. L. Holborn and L. W. Austin found the specific heat at constant
pressure, between 110° and 270°, to be 0-4639; between 110° and 440°, 0-4713;
and between 110° and 820°, 0-4881. The ratio of the two specific heats at 103°
or 104° lies between 1-25 and 1-35— say 1-3. W. Freitz found 13301 at 110°,
1-3129 at 120°, and 13119 at 130°.
The specific heat of water is higher than that of any other liquid, excepting
that of liquid ammonia. The general effect of the high specific heat of water is to
make the ocean, lakes, and streams absorb on heating or give up on cooling compara-
tively large amounts of heat which help to maintain the temperature more nearly
constant ; and to moderate the heat of summer, and the cold in winter. This is
shown by W. Zenker's comparison of the normal temperatures of continental and
marine climates at different latitudes.
Latitude
0"
10°
30°
50°
70°
90°
Continental
. 34-6°
33-5°
24-1°
5-0°
-19-0°
-26-1
Marine
. 26-1°
22-7°
18-8°
7-1°
-5-2°
-8-7'
The large capacity of water for heat also helps in the regulation of the temperature
by the transport of heat, so to speak, as ocean currents. Water being the chief con-
stituent of the living organism, also favours the regulation of the body temperature.
Thus, L. J. Henderson in his The Fitness of the Environment (New York, 1913) says :
Man is an excellent case in point. An adult weighing 75 kgrms. when at rest produces
daily about 2400 great calories, which is an amount of heat actually sufficient to raise the
temperature of his body more than 32° ; but if the heat capacity of his body corresponded
to that of most substances, the same quantity of heat would be sufficient to raise his tem-
perature between 100° and 150°.
The heat conductivity of liquid water, like that of other non-metallic liquids, is
low. Water is a bad conductor of heat.i^ The conductivity may be represented by the
number of calories transmitted per second per square centimetre through a centi-
metre layer with a difference of temperature of 1° between the two faces. The
conductivities of a few liquids contrasted with silver is as follows :
Silver. Mercury. Water. Alcohol. Ether. Benzene. Carbon disulphide.
1-530 0-0163 0-00152 0-00055 0-00038 0-00033 000027
Consequently, water is one of the best of liquids for conducting heat, but even
then, the thermal conductivity is small.
The thermal conductivity of water was determined by C. G. Lundquist in 1869, and
he obtained 0-00156 in C.G.S. units at 40-8° ; A. Winkelmann (1874), 0-00154 at 40° ;
H. F. Weber (1880), 0-00124 at 4°, and 0-00143 at 23*6° ; K. Weber (1903), 0-00131
(23°) ; S. R. Milner and A. P. Chattock, 0-001433 (20°) ; R. Wachsmuth, 0*00129
(4-1°); C. Chree, 000124 (18°) ; L. Graetz, 0-00158 (30°) ; R. Goldschmidt, 0-00150
(0°); and C. H. Lees, 000147 (11°), 0-00136 (25°) with an increase of —0-0055
per cent, per degree up to 45°. The heat conductivity of ice and snow is relatively
small, so that they protect the ground against the severe temperatures of a northern
winter. The ordinary nocturnal cooling of the soil by radiation under normal con-
ditions of soil exposure is of no significance when the ground is covered with snow.
The loss of heat from a river is retarded by the rigidity of the surface sheet of ice
which prevents direct contact of air and water. Although the conductivity of
ice is rather higher than that of water, the loss of heat by conduction is relatively
small in comparison with the heat losses by convection and wind currents which are
immediately stopped when a surface sheet of ice is formed. The protective action
473 INORGANIC AND THEORETICAL CHEMISTRY
of snow on the ground is of the greatest importance, for, when dry, as is the case in
the severest weather, a covering of snow is one of the best non-conductors of heat.
J. D. Forbes first determined the thermal conductivity of ice in 1874, and he found
that the conductivity is greater in the direction of the principal axis than it is when
perpendicular thereto. In the former case it is 00022 and in the latter 00021.
These numbers represent the number of calories of heat which flow per sq. cm. per
second in the direction of the fall of temperature when the temperature gradient is
1° per cm. A. C. Mitchell (1885) found the conductivity of ice to be 0"005 ;
F. Neumann, 000573 ; M. Straneo (1897), 030 to 00052 m the direction of the
principal axis, and 0005 when perpendicular to that direction. C. H. Lees found
the conductivity of ice to be 00052 at 3° ; 00058 at —57 ° ; and 0*0052 at —117°.
The heat conductivity of snow is much less than that of pure solid ice, and in
1885, T. Andrews found ice to conduct heat 122 per cent, better than snow. S. A.
Hjelstrom (1889) found the conductivity of snow to be 0*00051 ; and H. Abels
(1891) found the conductivity of snow to be proportional to the square of its density,
or K=0'006SD", and P. Jansson (1901) represented his results by the formula
^=0*00005-fO-0019D+0*006Z)2. T. Okada found that the density of snow varied
with its depth, for a depth of 10-20 cm., ^=000028 ; and for a depth 20-30
cm., ^=0*00045 ; his results agreed better with H. Abels' than with P. Jansson's
formula.
Optical properties.— Ice is optically positive. Its mean refractive index is
high ; and some have tried to show that Isaac Newton (1749) ^^ anticipated the
presence of combustible hydrogen in water, or at any rate the relation of water to
combustible substances, because he worked with thehypothesis that substances with
a high refractive index contained fatty, unctuous inflammable parts. According to
C. Pulfrich 15 the refractive index of ice is ai=l*30645, and €=1*30775 for the
5-line; w=l '30911, and €=1*31041 for the D-line ; and aj=l*31140, and
€=1*31276 for the E-line. According to A. Bertin, ice which is formed from water
at rest has its optical axes vertical to the cold surface, and F. T. Trouton explains
the greater heat conductivity of ice in the direction of the chief axis as the cause of
the orientation of ice crystals vertical to the cold surface. The index of refraction
of liquid water i6 at 16° is 1*3349 for the 5-ray ; 1*3322 for the D-ray ; 1 3358 for the
i^-ray ; and 1*3449 for the H-Ta,j.
The refractive index of most transparent substances for light waves of wave-
length within the limits of the visible spectrum, increases as the wave-length de-
creases— e.g. with water, alcohol, or carbon disulphide. The wave-length of violet-
light is shorter than red-light, and the index of refraction accordingly is greater for
violet than for red-light. With an alcoholic solution of fuchsine the reverse obtains,
for the violet rays are less refracted than the red rays. According to A. Kundt,
this anomalous phenomenon always accompanies great local absorption in the spec-
trum ; and wherever there is a strong absorption band in passing up the spectrum
from red to violet, the refractive index is abnormally increased below the band,
and abnormally diminished above the band. The refractive index of water de-
creases from about 1*4 to 1*3 in passing from the violet to the red end of the visible
spectrum. If the wave-length of the incident rays be increased upwards of 5 mm.,
the index of refraction increases to nearly 8*9 ; and generally with wave-lengths
between 6 metres and 6 millimetres, the refractive index is nearly 8*9. There is
therefore a big drop in the value of the index of refraction in passing from waves
5 mm. in length — and frequency 6x10"= — to the waves of red-light about j^th
mm. in length — and frequency 400x10^2. This anomalous behaviour is supposed
to be connected with the strong absorption band in the ultra-red spectrum of water.
Similar results are obtained with alcohol, the index of refraction of which drops from
about five to about half this value in passing from a wave-length 9 metres to about
8 mm.
According to J. H. Gladstone and T. P. Dale's data (1858) for the index
of refraction /x, the dispersion equivalent fiH—i^A and the dispersive power
I
WATER
473
(fjL^ — fi^)l{fjijy—l) of water are indicated in Table XVI. J. Jamin represents the
variation of the index of refraction of water with temperature by the formula
juo— 0-0412573— 0-05l929^2_there is no maximum at 4°. The index of refraction
for the extreme ultra-violet (214/x/x) is r40387 ; and for the ultra-red (1256/x/lc),
1-3210. E. van Aubel gives 1102 for the index of refraction of water at the
critical temperature. According to J. W. Briihl, the molecular refraction for
water by L. Lorenz and H. A. Lorentz's formula is 3282 for sodium light ; and
the value calculated from the atomic refractions of hvdrogen and oxygen
is 414.
Table XVI. — Optical Constants of
Water.
Index of refraction.
Temperature.
Dispersion
equivalent.
Dispersive
power.
fA
Mi,
f^M
0°
1-3291
1-3330
1-3438
0-0147
0-0429
10°
1-3288
1-3327
1-3434
0-0146
00439
20°
1-3279
1-3320
1-3427
0-0148
0-0445
30=
1-3270
1-3309
1-3415
0-0145
0-0438
40°
1-3257
1-3297
1-3405
00148
0-0449
50°
1-3241
1-3280
1-3388
0-0147
0-0448
60°
1-3223
1-3259
1-3367
0-0144
0-0441
70°
1-3203
1-3237
1-3344
0-0141
0 0435
80°
1-3178
■ — •
1-3321
0-0143
—
Assuming that water is a mixture of two substances, a comparison of the index
of refraction for water and ice shows that the specific refraction drops from
0-209680 for ice at 0° to 0*206342 for water at 0°, and subsequently, at 20°,
0-206208 ; at 60°, 0-206051 ; and at 100°, 0-206015. The increase in the value
of this constant for 100 parts of the following liquids heated from 10° to
20°, is
Ether.
0-08
Chloroform.
0-04
Ethyl iodide.
0-05
Ethyl acetate.
0-11
Carbon disulphide.
0-12
C. Cheneveau found the refractive index of ice and liquid water at 0° to be 1 '3095
and 1-3341 respectively, and the specific gravities 0-9176 and 0*99987. The
specific refractions by J. H. Gladstone and T. P. Dale's formula are respectively
0-3373 and 03341 ; and by L. Lorenz and H. A. Lorentz's formula 02097 and
02063 respectively. Similarly, for water at 100° the specific gravity is 0-95838,
the refractive index 1-3182, and the specific refractions 0-3320 and 0*2019 re-
spectively. The decrease with water is supposed to show that the normal increase
with temperature has superposed upon it a decrease due to a change in the mole-
cular constitution of the molecules of the water so that water is a mixture of two
substances in proportions varying with the temperature. The results calculated
by C. Cheneveau (1913) on this hypothesis are in agreement with observation.
J. Jamin found that difference between the refractive index of dry air and air
saturated with aqueous vapour to be 0-06726, an extremely small quantity.
The colour of water. — In 1828, H. Davy i^ described the water from snow and
glaciers in different parts of the Alps, as " pure water," and added that " its colour,
when it has any depth, or when a mass of it is seen through, is bright blue ; and,
according to its greater or less depth of substance, it has more or less of this colour."
In 1851 R. Bunsen's attention was also directed to the greenish-blue tint of the
Icelandic geysers, and he found that purified water in a glass tube blackened on
the inside, and two metres long, appeared distinctly blue, and he accordingly denied
that water is colourless, but is actually blue. Hence, it is generally considered
474 INORGANIC AND THEORETICAL CHEMISTRY
that the purest water is colourless in moderately thin layers but that it appears
faintly blue when viewed in thick layers- — say in a tube 2 metres long. According
to W. Spring, the blue tint can be closely imitated by a solution of cupric chloride
of the proper concentration. Lord Rayleigh believes that the blueness of water
at a depth of 4 metres is largely exaggerated by W. Spring, although possibly a
fully developed blue may be obtained at much greater thicknesses. Lord Rayleigh
says the colour of the transmitted light is a greenish-blue ; and he believes that the
pronounced blue colour reported by many observers is due to insufficient care being
taken to start with white light.
According to W. Spring, the faint blue tint of purified water seems to be dependent
upon the presence of oxygen. Liquid oxygen, O2, is distinctly blue ; liquid ozone,
O3, is intensely blue ; and hydrogen peroxide, H2O2, has rather a deeper blue
colour than water, H2O, so that in hydrogen peroxide the oxygen loses less of its
characteristic tint than it does in water. Many organic compounds containing
the hydroxyl OH-group are also blue — e.g. methyl and ethyl alcohols (CH3OH
and C2H5OH) are bluish-green when viewed in a long tube. As the number of carbon
atoms increase, making a longer chain, the colour changes into the golden-yellow
which is found in liquid hydrocarbon compounds free from hydroxyl. As the
carbon chain of the hydroxyl compound increases in length — C3H7OH, C4H9OH,
C5H11OH, etc. — the yellow colour becomes more and more pronounced — with
amyl alcohol, CsH^OH, the colour is yellowish-green — until finally the yellow
overpowers the blue altogether.
According to J. Aitken (1880), the blue colour of large bodies of water — e.g. in
china-clay settUng pits ; in tanks in which water is being softened by the addition
of milk of lime ; etc. — is an optical efiect due to the action of the fine particles
suspended in the liquid on the light. J. L. Soret (1869), E. Hagenbach (1870), and
J. Tyndall (1871) stated that the water from Lake Geneva is not optically empty,
but that the blue colour is possibly due to the scattering of light from numberless
colourless particles. H. St. C. Deville (1848) and G. C. Wittstein (1861) analyzed
a great number of natural waters, and concluded that the brown or yellow waters
contain more organic matter and less calcium salt than green waters, this organic
matter is brown, and the blue colour of natural waters changes to green, yellow,
brown, or black as the proportion of organic matter increases. W. Spring, however,
has pointed out that G. C. Wittstein' s data really show that the colour of natural
water stands in no direct relation with the organic matter or alkali concentration.
He showed that with water containing ferric oxide in solution or suspension, the
colour is dark mahogany-brown with a concentration of 1 : 10000 ; golden-yellow
with 1 : 1000000 ; grass-green with 1 : 8000000 ; and blue like pure water with
1 : 24000000. Similarly with humic matter, the colour is yellowish- brown with a
concentration 1 : 500000 ; green with 1 : 20000000 ; and blue with 1 : 50000000.
Consequently, W. Spring (1905) argues that the green colour of certain natural
waters is not due to dissolved calcium salts, but rather to an invisible suspension —
probably organic matter and silica. The brown or yellow colour of certain natural
waters is due to humus or salts of iron. According to^Lord Rayleigh, the apparent
colour of the sea is largely determined by the colour of the sky seen by
reflection.
A. Secchi found that the red and yellow rays are lacking in the absorption
spectrum of sea- water. 0. von Aufsess measured the transmission of light in various
parts of the spectrum, and found the principal absorption is in the red and yellow ;
and with the purest water he found practically no absorption above the i^-line,
and a high transparency in this region was attained by many natural waters. Hence,
if in sufficiently thick layers such waters must appear blue. The absorption
spectrum of water is indicated in the diagram. Fig. 23. The infra-red heat rays
are strongly absorbed. The maximum absorption is towards the red and orange,
the maximum transmitted is towards the blue and green. There are large absorp-
tion bands in the spectrum of water is at the approximate wave-lengths 15, 2"3,
>
WATER
475
4-75, and 6/x. Water is very opaque to the deep infra-red radiation and the
spectrum of numerous narrow absorption bands can be resolved only when the
substance is reduced to a highly attenuated vapour ; in the liquid state, these
groups of small bands coalesce into larger bands. In the visible spectrum, and as
far as 0-933/x in the infra-red, a thick layer of water is needed to produce absorption
bands. A layer 1 cm. thick absorbs all frequencies beyond 1-4/x ; a layer 05 mm.
thick is quite opaque beyond 2/x ; and a layer 01 mm. thick is quite opaque
beyond bfju. Beyond 8/i, water is transparent. According to W. W. Coblentz, if
infra-red absorption spectrum bands of a hydrated compound be present at wave-
lengths l-5/x,2'0)it,30/x, 4-75/x, and 6-0jLt,and the absorption bands at l-5/x,2-0/>t,
and 4:'75/x, are weak while the others are strong, the product is a hydrate because
these absorption bands are characteristic of water itself. The presence of a hydroxyl
group may cause an absorption band near 3/x, but if water molecules be absent,
the other characteristic bands are absent. G. Bode studied the infra-red spectrum
of ice.
Water vapour is more transparent than the liquid. About the middle of last
century there was an interesting controversy between J. Tyndall 19 and G. Magnus
on the absorption spectrum of water vapour. The former obtained a strong
absorption, the latter a negligibly small absorption. Other physicists investigated
the subject, and it has now been established not only that water vapour absorbs
heat rays but also what particular rays are absorbed, and how much of each.
Atniospheric water vapour transmits the sun's radiation as far as 11/x, while a layer
of liquid 2 cm. thick, and
equivalent to the water in
the earth's atmosphere, ab-
sorbs everything beyond
1-2 fJL.
According to M. Fara-
day, ice is positively electri- | °' '
fied by friction with water,
and on this fact, L.
Sohncke 20 has founded a
theory of atmospheric elec-
tricity. The electrical con-
ductivities of ice and water are very low, and this the more, the greater the
degree of purity. With an alternating current of 1000 cycles per sec, the
resistance of ice per centimetre cube is given as 7*22x10^ ohms; the corre-
sponding conductivity, as 1400x10" 11, where with a direct current the conduc-
tivity is 1-63x10-9. With an alternating current of 1000 cycles, the breakdown
voltage per cm. is 0*011 X 10^ ; 21 the specific inductive capacity is 86*4, and with an
alternating current of 15 cycles, 429*0. The electrical conductivity of ice, said G.
Foussereau (1884),22 is 15'000 times smaller than that of water, which is represented
to have an absolute resistance of 9400 ohms. Ice, at 0°, has an electrical resistance
of 4865 megohms, and at —17°, a resistance of 33,540 megohms. Dry ice indeed
is considered to be one of the most perfect of insulators. Telephone cables which
are defective in insulating properties through moisture may become all right when
the temperature falls below the freezing point.
The specific resistance of water at 15° was found to vary between 118,900 and
712,500 ohms, according to the degree of purity. The higher number was obtained
with water distilled three times in a platinum vessel. The resistance of water kept
for 24 hours in glass vessels at 15° was found to diminish about one-thirtieth owing
to the solvent action of water on the glass ; if kept in platinum vessels, the dimi-
nution is slower, and it is due to the slow absorption of salts and acid vapours from
the atmosphere. F. Kohlrausch and A. Heydweiller 23 purified some water by
distillation in vacuo which had a conductivity of 0*043x10"^ reciprocal ohm per
cm. cube at 18°, or 0*015 Xl0~6 reciprocal ohm at 0°. Twice-distilled water has a
0-3
0-2
0-1
n
Red.
Ora
nge.
Yellow
Green
"^
'
N
\
s
^^
,
^ 660 640 620
520 500 480/xfi
600 580 560 54-0
y^ai/'e Length. ^^
Fig. 23.— The Absorption of Light by Purified Water.
476 INORGANIC AND THEORETICAL CHEMISTRY
conductivity of from 1 to 2x10"^ reciprocal ohms. This, saidF. Kohlrausch and
A. Heydweiller, means that :
One millimetre of this water has at 0° a resistance equal to that of a copper wire of the
same cross-section 40,000,000 kilometres long, a wire that could therefore be wound a
thousand times round the earth. This water is probably the purest that has ever existed,
whether artificially prepared, or occurring ready formed in nature, not even excepting the
water precipitated in the form of clouds in the highest strata of the atmosphere. Simple
contact with the air for a short time raised its conductivity tenfold. The impurities still
present in the water might be estimated at a few thousandths of a milligram per litre.
F. Kohlrausch further estimates that the specific conductivity of absolutely pure
water at 18° is 0038x10"^ reciprocal ohm. J. Negreanu found the conductivity
of ordinary tap-water to be between 200 and 760 reciprocal ohms at 18°.
J. J. Thomson found the conductivity of electrolytes under very rapidly alter-
nating frequencies to be fairly constant up to 10^ cycles. J. A. Fleming and
G. B. Dyke found the conductivities of many solids increased rapidly with the
frequency — e.g. ebonite at 4600 cycles had a conductivity 6 4 times greater than
with 800 cycles. B. van der Pol found that the normal conductivity of sea- water
with steady currents is 5 XlO~ii, and for currents of frequency 275, 1070, and 3400,
the conductivities are respectively 1*005, 1'002, and 1*001 times the normal value,
thus showing that the conductivity of sea-water for all frequencies used in wireless
telegraphy is nearly equal to the value for steady currents to within less than
a half per cent. K. T. Compton calculated the ionizing potential for water vapour
from the formula 7=0-194(jfiL— 1)"* volts to be 864 volts, where Z is the dielectric
constant, and V the ionizing potential.
The velocity of migration of the H*-ion is i;*=318 at 18°, and for the OH'-ion,
«;'=174. The concentration C of the ions in gram-ions per litre Ch=Coh is
A/A^=A/(y+?;')=0-0384xlO-6/(318-fl74)=0-78xlO-7. The conductivity A in-
creases with temperature :
0° 2° 10° 18^ 26° 34° 42° 50°
AxlO'. . 0-0115 00133 00233 00361 00567 0*0833 0-1210 01690
The change in the concentration of the H*-ions is about 004 per degree.
If water be ionized H20^H'+0H', the equilibrium condition is [H'][OH']
=K'[H.20], where K is the ionization constant, and since the un-ionized water is in
very great excess, the term ^'[H20] is also constant, and hence [H*][OH']=-K',
where K is the so-called water constant. The ionization constant for water has been
worked out by several different methods. S. P. L. Sorenson reduced the more
important of these to their values at 18°. S. Arrhenius and J. Shields obtained
K=0'13XlO~^^ from measurements on the hydrolysis of sodium acetate;
J. J. A. van Wijs, 0*83 XlO~i* (hydrolysis of methyl acetate); H. Lunden
0*61 XlO~i* (hydrolysis of trimethyl pyridine 7?-nitrophenol) ; F. Kohlrausch and
A. Heydweiller, 0'63xlO~i* (conductivity of water) ; W. Ostwald and W. Nernst,
0'64xlO~i* (e.m.f. of hydrogen electrode in acid alkali cell) ; R. Lowenheiz used
an analogous process and obtained 0*74 Xl0~4, and g_ p^ jj Sorenson obtained
0*72 X 10-1* from the e.m.f. of the hydrogen electrode against O'OliV-KCl and the
calomel electrode. The best representative value is [H][OH']=0'73xlO-i* ;
hence [H']=[OH']=0*85xlO-'^. This is taken to mean that in a litre of water,
at 18°, 0*000085 milligram of hydrogen is present as free hydrogen ions. The
ionization constant for water changes rapidly with temperature. R. Lorenz and
A. Bohi computed values from their measurements of electrode potentials, and
F. Kohlrausch and A. Heydweiller and A. A. Noyes and co-workers from measure-
ments of the electrical conductivity. The results are by no means concordant.
R. Lorenz and A. Bohi give :
0°
18°
2.'",'
30°
.50°
70°
90°
99°
Kxm*
. 014
0-72
1-22
1-74
8-8
21-5
53-5
720
[H-]xl0' .
. 0-37
0-85
1-10
1-32
2-96
4-61
7-3
8-49
0°
18°
25°
100°
150°
218°
306°
0-30
0-58
0-91
6-9
14-9
21-5
130
8-089
0-46
0-82
48
223
461
168
WATER 477
According to C. W. Kanolt (1907), the ionization constant K for water at 0° is
0-089x10-14; at 18°, 0-46x10-1*, and at 25°, 0 82x10-1*, while A. A. Noyes.
Y. Kato, and R. B. Sosman give the concentration of the hydrogen ion, [H'J,
and the ionization constant K at different temperatures d, as
d
[H-]xlO' .
iiCxlOi*
The heat of ionization Q of one gram-molecule of water calculated from the usual
formula log {K,^IKt,)=Q(T2-T^)IRT^T2, is nearly 14 Cals. when 22=1-986
cals.
F. Kohlrausch assumes that each ion is surrounded by a shell or atmosphere of
the solvent which differs in some respects from the rest of the solvent ; the dimensions
of the atmosphere is determined by the character of the ions. In other words,
the ions are hydrated. The electrolytic resistance of an ion is a frictional re-
sistance which increases with the dimensions of the atmosphere. F. Kohlrausch
continues :
The relationship between the mobilities of the ions and their temperature coefficients
first led me to seek a general explanation for the electrolytic resistance in the idea of a water
atmosphere, in order to escape being compelled to explain this otherwise irreconcilable
fundamental characteristic of the ions as a deus ex machind. Assuming as the single
fundamental characteristic of each univalent monatomic ion the formation of a water
atmosphere, which varies according to the nature of the ion, the mobility of the complex
on the one side, and its temperature coefficient on the other, will be fiuictions of the atmo-
spheric formations, and therefore both quantities must hold functional relations to each
other. We know at present too little of the molecular forces to attempt to describe this
connection more exactly ; but for the case in which the water shell is so thick that the ion
exerts no force beyond it, the resistance to motion becomes simply a matter of water friction,
which explains the fact that the most sluggish ions have nearly the same temperature
coefficients as the viscosity. In the case of smaller aggregations, we must remain content
with the fact that we have at least the possibility of a fundamental explanation.
W. Nernst assumes that the solvent water is strongly contracted by the presence
of free ions, and the observed contraction which occurs during the dissolution of
ionized substances is smaller than the molecular volume of the solid — e.g. sodium
carbonate, magnesium or zinc sulphate, etc. The electrostriction is caused by
electrostatic fields of the ions which make the solvent contract in their immediate
vicinity.
In addition to the ionization H20^H*+0H' in which water acts as if it were a
monobasic acid, a second stage in the ionization is conceivable, H20^H0'+H'
^2H*4-0", where the water acts as if it were a dibasic acid. Nothing definite is known
about this second stage of ionization ; if it does occur at all, it must be in exceedingly
small proportions because the second stage in the ionization of a dibasic acid is
always more difficult than the first stage, and the first stage with water is very
small.
Many organic substances — fatty acids, oximes, alcohols, etc. — form complex
or polymerized molecules when dissolved in hydrocarbons, chloroform, carbon
disulphide, or carbon tetrachloride ; the complexes are usually broken down into
simpler molecules when these substances are dissolved in water, and to a less
extent when dissolved in alcohols, ethers, or phenols. The latter class of solvents
is said to be ionizing because when saturated with hydrochloric acid the liquids act
as conductors of electricity, whereas the former class of solvents is non-ionizing
because the liquids are virtually non-conductors under similar conditions. Water is
far excellence the ionizing solvent. J. W. Briihl 24 explains this by assuming that
water is an unsaturated compound containing quadrivalent oxygen H2==0=, or
H— 0— H, and that the latent valency of water is the cause of the formation of
molecular aggregates which in turn makes water an ionizing solvent. The organic
478 INORGANIC AND THEORETICAL CHEMISTRY
solvents, too, which act in a similar way, usually contain hydroxy lie oxygen, while
those solvents free from oxygen — hydrocarbons, chloroform, etc. — have usually
little or no ionizing power. J. W. Briihl cites in favour of the view that water has
an unsaturated molecule : (1) Nearly all substances capable of uniting with water
are hygroscopic ; (2) Numerous hydrates and compounds of water of crystallization
exist ; and (3) Water is an unusual solvent.
According to R. Abegg, the dielectric capacity 25 of ice at —18° is 316 for waves
approximately A=5xl03 cm. ; while U. Behn and K. Kiebitz give 1-76 and 1'88
for waves A=75 cm. at —190°. According to K. Badeker, the dielectric constant
of water vapour under 3 atm. pressure, at 145° is 1 '00705 (vacuum unity) or
1-00646 (air unity) ; and at 6° the dielectric constant is l-00705{ 1— 0-00014(145— ^) } .
M. Jona also measured the dielectric constant of water vapour at tempera-
tures ranging from 178° to 178-1°. According to Maxwell's rule, the square
of the index of refraction fi^ is equal to the dielectric capacity K for electric fields
alternating with a low frequency. C. B. Thwing (1894) found that the dielectric
capacity of liquid water with rapidly alternating oscillations (with a wave-length
over 10 m.) rises from 79*46 at 0° to a maximum 85-2 at 4°, and falls to 79-4 at 7°.
The refractive index fi for long electrical waves is given by the quotient yi=XjX\
where A and A' represents respectively the wave-lengths in air and in the compound
under investigation. For sufficiently long waves — about 70 cm. — the square of
the refractive index /x is equal to the dielectric constants, or ix^=K. At 17°-18°,
for water, H. Merczyng, A. CoUey, H. Rukop, J. F. Smale, B. B. Turner, and
P. Drude found :
A , . . .3-5 4-5 55-5-68-5 long c. 10*
[X . . . . 6-54 6-88 8-955 O'O —
/x2 . . . . 42-7 47-3 80-26 81-0 —
K .... — — 80-9-81-1 — 80-0
C. Niven found the dielectric constant decreased with increasing temperatures
at 0°, iC=90-36 ; at 7°, 80-06 ; at 33°, 69-31 ; at 58-32°, 59-5 ; and at 83°, 37-97.
B. Hopkinson and E. Wilson found that the dielectric constant of ice is nearly 80
with low frequency electrical oscillations between 10 and 100 per second. From
Maxwell's rule, the refractive index of ice is 1*41 corresponding with a dielectric
constant of about 2 ; and the dielectric constant of ice with oscillations of a
frequency of a million is a number less than 3. Thus, the refractive index of ice
for electro-magnetic waves falls to 1-4 for waves of even moderate frequency under
conditions where the refractive index of water still remains at 8-9. This illustrates
the general observation that when liquids with a high dielectric constant pass into
the sohd state, the abnormal refractive index is more easily reduced to approximate
with the value of fx for the visible spectrum by increasing the frequency. J. A.
Fleming and J. Dewar also found that the dielectric constant of ice falls from 80
at about 0°, to nearly 3 at very low temperatures, say — 190°, and the refractive
index is then 7-6 with waves of low frequency and wave-length 14 mm. The re-
fractive index of ice for light vibrations of wave-length from 14 mm. to 2088 cm.
progressively decreases from 1-76 to 1-50. The latter corresponds with a dielectric
constant 2-25. R. Blondlot obtained a value 2-6 for oscillations of still greater wave-
lengths. R. Abegg at -18°, found for A=:5xl03, i^=3-l ; for A==75 at -90°,
U. Behn and F. Kiebitz found ^=1*76 to 1-88 ; and E. Beaulard found at 0° Z=/x2
=1*71. There is a steady, almost linear, change of dielectric capacity with
temperature, such that the dielectric constant at 6° between 0° and 76° is
80-6{ 1-0 OO4583(17-0)-f 0-0000117(17-^)2;. With short waves (under 1 m.)
the dielectric constant is rather greater than with longer waves — at 17°, 80'6 with
a wave-length 200 cm., 81*7 for 74 cm., and 83*6 for 38 cm. E. A. Harrington
found the dielectric constant of aqueous solutions of sugar, and methyl alcohol
to be less the greater the concentration; with aqueous solutions of urea, the
reverse obtains.
W^ater.
Methyl alcohol.
Ethyl alcohol.
81-7
34
26
88
62
25
86
60
24
83
36
16
WATER 479
The high dielectric constant of water is supposed to give a hint as to the
cause of the great ionizing power, when contrasted with other solvents, e.g.
Dielectric constant ....
Per cent, ionization i^jjN) Potassium iodide
„ „ „ Sodium bromide
,, ,, ,, Potassium acetate
J. J. Thomson's explanation 20 how a high dielectric constant favours ionization is
as follows :
If we take the view that the forces which hold the atoms in the molecules together are
electrical in their origin, it is evident that these forces will be very much diminished when the
molecule is close to the surface of, or surrounded by, a conductor, or a substance like water,
possessing a very large specific inductive capacity (dielectric constant). Thus, let A, B,
Fig. 24, represent two atoms in a molecule placed near a conducting sphere, then the effect
of the electricity induced on the sphere by A will be represented by an opposite charge
placed at A', the image of A in the sphere. If A is very near the surface of the sphere,
then the negative charge at A ' will be very nearly equal to that at A .
Thus, the eSect of the sphere will be practically to neutralize the
electric effects of .4 ; as one of these effects is to hold the atom B in
combination, the affinity between the atoms A and B will be almost
annulled by the presence of the sphere. Molecules condensed on the
surface of the sphere will thus be practically dissociated. The same
effect would be produced, if the molecules were surrounded by a
substance possessing a very large specific inductive capacity. Since Fig. 24.
water is such a substance, it follows, if we accept the view that the forces
between the atoms are electrical in their origin, that when the molecules of a substance are
in aqueous solution, the forces between them are very much less than they are when the
molecule is free, and in a gaseous state.
P. Dutoit and E. A. Aston ^7 suggest that the ionizing power of a solvent is
dependent on its degree of polymerization ; they show that only polymerized solvents
conduct electricity.
Ice is diamagnetic.28 The coefficient of magnetization when referred to unit
mass is 07193 X 10-6 at 20° with a temperature coefficient of O'OOOIS at 20° ; P. Seve
found 0725x10-6 at 22°. A. Piccard found that the diamagnetic coefficient of
water at 0° is 07174x10-6, and it changes with temperature, until, at 100°, it
attains the constant value 07228 X 10-6. The curves showing the relation between
the variation of the coefficients of magnetization of water with respect to tempera-
ture, led A. Piccard (1912) 2^ to infer that " in water at ordinary temperatures there
are two substances in equilibrium." Determinations of the magnetic susceptibility
of water 30 by different investigators give numbers ranging from 6'4xlO-7 to
8*4:XlO-7 for temperatures approximating 20°. A. Piccard's value at 20° is
7-193x10-7, and C. H. Hayes' value at 24° is 7*26x10-7. The magneto-optic
rotation of liquid water for a wave-length A=0'2496jLt is 0*1042' per cm. per
unit magnetic field; for A=0-4046/x, the rotation is 0*0293'; for A=l*000/x, the
rotation is 0*00410' ; and for A=l-300, the rotation is 0*00264'. L. H. Siertsema's
value 31 for E. Verdet's constant for water at 13*4° and the D-line is 0*01302 ;
F. Agerer's value at 18°, 0*01309 ; G. Quincke's value at 21*81° is 0*01414 ; and
L. Arons' value at 23°, 0*01293. J. W. Rodger and W. Watson give for 0° between
3° and 98°, 0*01311 (1—0*0000305^—0*0000030502). l. H. Siertsema calculated
the ratio of E. Verdet's constant for light of wave-length A to the value for the
D-line at 20° ; for A=0*405jLt, the ratio is 2*218 ; for A=:0*589, unity ; and for
A=0*701/x, the ratio is 0*700. J. W. Rodger and W. Watson also calculated
the molecular rotation for the 2)-line between 4° and 90°, and they also found
that the magnetic rotary power of unit depth of water in a magnetic field of
unit strength at a temperature Q between 4° and 98° is 0*01311—0*0640—0*07402,
J. Kerr found that when the wire terminals of an induction coil were embedded in
a block of glass placed between crossed nicols, there was a restoration of the light.
The restored light could not be extinguished by rotation of the analyzer, and the
phenomenon was not therefore a simple rotation of the plane of polarization, but
480 INORGANIC AND THEORETICAL CHEMISTRY
an elliptical polarization resulting from the fact that the medium had become
doubly refracting. The phenomenon also occurs with isotropic liquids. It is
called Kerr's electro-optic effect. If D denotes the path-difference measured in
wave-lengths of the two components of the vibration, K the strength of the electric
field, then for unit length of fluid, D=jK, where _; is the so-called Kerr's constant. 32
For light of wave-length 680/x/Lt, at 20°, G. Lemoine^ found Kerr's constant to be
3*70 X 10"-^, which is nearly the value found by R. Leiser for water. W. Obolensky 33
found that water gives a maximum photo-electric effect with the extreme ultra-
violet rays in the neighbourhood of A=130/x/x ; and, becoming less as the wave-
length increases, vanishes completely, when A=202'5/x/x. Ice is more sensitive
than water ; with rays in the neighbourhood of A=130/x/x, the activity of water is
4J^jth of that of cupric oxide, and for rays approximating A=190/Lt/i, ice surpasses
cupric oxide.
References.
1 A. Schrotter, Sitzher. Akad. Wien, 10. 257, 1853 ; H. C. Sorby, Phil. Mag., (4), 18. ]05,
1859 ; J. Thomsen, Trans. Roy. Soc. Edin., 16. 575, 1849 ; R. Pictet, Archiv. Sciences Oentve,
59. 154, 1897.
2 G. liellmai.nn, Schneekrystalle, Berlin, 1893 ; Glaus Magnus, Historia d^e gentihus septentrio-
nalibus, Romae, 1555 ; J. Kepler, Strena seu de nive sexangida, Francofurt, 1611 ; W. A. Bentley,
Nature, 65. 234, 1901 ; J. E. Wolff, Proc. Amer. Acad., 33. 431, 1898; J. Schukewitsch, Bull,
Acad. St. Petersburg, 291, 1910 ; L. Milch, Neues Jahrb. Mm., 532, 1909 ; P. Groth, Tahellarische
Uebersicht der Mineralien, Braunschweig, 41, 1898 ; Physikalische Krystallographie, Braunschweig,
478, 1905; M. Trautz, Zeit. phys. Chem., 53. 9, 1905; R. Prendel, Zeit. KrysL, 22. 76, 1904;
K. Futterer, ib., 38. 510, 1904 ; A. Bertin, lyistit., 208, 1864 ; Ann. Chim. Phys., (5), 13. 283,
1878 ; F. Klocke, Neues Jahrb. Mm., i, 83, 1881 ; 272, 1879 ; i, 159, 1880 ; A. B. Dobrovolsky,
Arch. Kemi Min. Geol., 6. 1, 1916; A. E. von Nordenskiold, Oefv. Akad. Forh. Stockholm, 17.
439, 1860 ; Pogg. Ann., 114. 613, 1861 ; G. von Nordenskiold, Geol. Foren. Forh. Stockholm, 15. 146,
1893 ; 20. 163, 1898 ; Compt. Rend., 116. 770, 1893 ; Bull. Soc. Chim., (3), 16. 59, 1893 ; Albertus
Magnus, Opera omnia {Meteorol., 1. 10), Lugduni, 1651 ; J. Glaisher, Rep. Brit. Meteorol. Soc, 17,
1855; Jaurn. Micro. Soc, 3. 179, 1855; 4. 203, 1855; F. Rinne, Ber. Sachs. Ges. Wiss., 69. 57,
1917; A. St. John, Proc Nat. Acad., 4. 193, 1918. I am indebted to the pubhsher of
Hollman's Schneekrystalle for permission to use Fig. 21.
« W. C. Rontgen, Wied. Ann., 22. 510, 1885 ; R. Cohen, ib., 45. 666, 1891 ; E. Warburg and
J. Sachs, ib., 22. 518, 1885; T. E. Thorpe and J. W. Rodger, Phil. Trans., 185. A, 397, 1894;
L. Hauser. Drude's Ann., 5. 597, 1901 ; M. *de Haas, Versl. Akad. Amsterdam, 1. 23, 1894 ;
K. F. Slotte, Oefvens. Finsk. Vet. For hand., 32. 116, 1890; T. W. Richards and S. Palitzsch,
Journ. Amer. Chem. Soc, 41. 59, 1919 ; E. C. Bingham and G. F. White, Zeit. phys. Chem., 80. 670,
1912; R. Hosking, Phil. Mag., (6), 17. 502,'l909; (6), 18. 260, 1909; G. F. White and
R. H. Twing, Amer. Chem. Journ., 50. 380, 1913; G. Y. Williams and E. Washburn, ib.,
35. 737, 1906; E. C. Bingham, ib., 40. 277, 908; 43. 287, 1910; G. Zemplen and B. Pogany,
Ann. Physik, (4), 49. 39, 1916.
* K.Kundt and E. Warburg, Pogg. Ann., 155. 337, 1875; F. Houdaille, F&rtschr. Physik, 1.
442, 1896 ; Mesure du coefficient de la vapour d^eau dans Vatmosphere et du coefficient de frottement
de la vapour d*eau, Paris, 1896.
« R. M. T)eQ\ey, Geol. Mag., (4), 2. 408, 1895; Proc Roy. Soc, 81. A, 250, 1908;
J C. McConnel and D. A. Kidd, ib., 44. 331, 1888 ; J. C. McConnel, ib., 49. 323, 1891 ; J. F. Main,
ib., 42. 329, 1887 ; R. M. Deeley and P. H. Parr, Phil. Mag., (6), 26. 85. 1913 ; H. Morphy,
ib.. (6), 25. 133, 1913 ; J. Dewar, Chem. News, 91. 216, 1905 ; B. Weinberg, Journ. Russian Phys,
Chem. Soc, 38. 186, 250, 289, 329, 1906 ; Ann. Physik, (4), 18. 81, 1905 ; (4), 22. 331, 1907 ;
Zeit. Gletscherkunde, 1. 5, 1907; 4. 380, 1910; N. Slatowratsky and G. Tammann, Zeit. phys.
Chem., 53. 341, 1905; H. Hess, Ann. Physik, (4), 8. 405, 1902; G. Tammann, ib., (4), 7. 198,
1902 ; Flussige Kristalle, I^ipzig, 1904 ; 6. Miigge, Neues Jahrb. Min., ii, 2J 1, 1895 ; H. Reusch,
Pogg. Ann., 121. 573, 1864; R. Koch, Wied. Ann., 25. 438, 1885 ; R. Trowbridge, Amer. Jonrn.
Science, (3), 29. 349, 1885; H. Moseley, Phil. Mag., (3), 39. 1, 1870; T. Andrews, Proc. Roy. Soc.
40. 544, 1886; G. S. Turpin and A. W. Warrington, Phil. Mag., (5), 18. 120, 1884.
* L. J. Henderson, The Fitness of the Environment, New York, 1913.
' R. Eotv6=i, Wied. Ann., 27. 448, 1886 ; W. Ramsay and J. Shields, Phil. Tram., 184. A,
647, 1893; P. Volkmann, Wied. Ann., 11. 177, 1880; 17. 353, 1882; 56. 457, 1895; 53. 653,
1894 ; 66. 220, 1898 ; F. Poekels, ib., 67. 668, 1899 ; G. Timberg, ib., 30. 558, 1877 ; M. Cantor,
ib., 47. 408, 1892; A. Heydweiller, ib., 65. 311, 1898; T. Lohnstein, ib., 53. 1073, 1894;
P. I^nard, ib., 30. 232, 1887 ; Heidelberger Ber., 18. 1910 ; C. Brunner, Pogg. Ann., 70. 481, 1847 ;
L. Wilhelmy, ib., 119. 186, 1863 ; G. Quincke, ib., 139. 1, 89, 1870 ; 160. 337, 1877 ; Ann. Physik,
(4), 9. 1, 1902 ; F. von I^rch, ib., (4), 9. 434, 1902 ; W. Gallenkamp, ib., (4), 9. 475, 1902 ; R. H.
Webber, ib., (4), 4. 718, 1901 ; C. Forch, ib.. (4), 17. 750, 1905 ; L. Gnmmach, Abti. Eich Komm.,
3. 103, 1902 ; B. Weinberg, Zeit. phys. Chem., 10. 34, 1892 ; C. Watson, Phys. Rev., 12. 257,
WATER 481
1901 ; P. Hall, Phil. Mag., (5), 36. 385, 1893 ; Lord Rayleigh, ih., (5), 30. 38G, 1890 ; H. E.
Dorsey. ih., (5), 44. 349, 1897 ; W. F. Magie, ih., (5), 26. 162, 1888 ; Wied. Ann., 25. 421, 1885 ;
H. Sentis, Ann. Univ. Grenobh, 9. 1, 1897 ; L. Weinstein, Metron. Beitrdge, 6, 1889 ; C. Forch,
Wied. Ann., 68. 801, 1899 ; Ann. Physik, (4), 17. 744, 1905 ; A. Kalahne, ib., (4), 7. 440, 1902 ;
N. Bohr, Proc. noy.Soc.,S2. A, 19G, 1909 ; R. Magini, Ecvd. Accad. Lincei, (5), 19. ii, 184, 1911 ;
T. W. Richards and L. B. Combs, Journ. Amer. Chem. Soc, 37. 1656, 1915; A. Mayer, Amer.
Journ. Science, (4), 3. 253, 1897 ; Nature, 56. 21, 1895 ; L. Gav, Compt. Rend., 156. 1070, 1913 ;
F. M. Jfiger, Proc. Acad. AmHerdam, 17. 329, 1917; W. Sutherland, Phil. Mag., (5), 38. 188,
1894 ; (o), 40. 477, 1895 ; (5), 50. 460, 1900 ; A. M. Worthington, ib., (5), 20. 66, 1885 ; T. W.
Richards and J. H. Mathews, Journ. Amer. Chem. Soc, 30. 11, 1908; T. W. Richards and S.
Palitzsch, ih., 41. 59, 1919 ; A. Ritzel, Zeit. phys. Chem., 60. 319, 1907 ; K. Dnicker, ib., 64. 1,
1908 ; W. C. McC Lewis, ib., 74. 640, 1910 ; G. Metz, Wied. Ann., 41. 663, 1890 ; W. C. Rontgen
and J. Schneider, ib , 29. 165, 1886 ; 31. 1000, 1887 ; M. Schumann, ih., 31. 14, 1887 ; L. Kolow-
rat, Journ. Russ. Phys. Che.m. Soc, 36. 265, 1904; L Traube, Arch. ges. Physiol, 132. 511,
1910; 140. 109, 1911; G. Tammann, Ueber die Beziehungen zwischen den inneren Krdften
und Eigenachaften der Losungen, Leipzig, 178, 1907 ; F. Nansen, North Polar Expedition, Scieni.
Results, 10. 5l', 1900; J. M. Jager, Zeit. anorg. Chem., 101. 1, 1917.
8 A. Winkelmann, Wied. Ann., 22. 1, 152, 1884 ; 23. 203, 1884 ; 26. 105, 1885 ; 33. 445,
1888 ; 36. 93, 1889 ; A. Masson, Compt. Rend.., 44. 464, 1857 ; Phil. Mag., (4), 13. 533, 1857 ;
W. Jager, Wied. Ann., 36. 165, 1889 ; W. Treitz, Ueher die Fortpflanzungsgeschwindigkeit des
Schalles in einigen Ddmpfen, Bonn, 1903.
9 H. V. Regnault, Mim. Acad., 21. 730, 1847 ; H. T. Barnes, Phil. Trans., 199. A, 140, 1902 ;
W. R. and W. E. Bousfield, ib., 211. A, 199, 1911 ; H. L. Callendar, ib., 199. A, 142, 1902 ; Proc
Roy. Soc, 86. A, 254, 1912 ; H. T. Barnes and H. L. Cooke, Phys. Rev., 15. 65, 1902 ; C. Dieterici,
Ann. Phys., (4), 16. 593, 1905 ; Ber. deut. phys. Ges., 2. 228, 1904 ; C. E. Guillaume, Compt. Rend.,
159. 1483, 1914 ; W. Jager and H. von Steinwehr, Die Wdrmekapazitdt des Wasser zwischen
5° und 50° in irdernationalen Wattsekunden, Berlin, 1915 ; H. A. Rowland, Proc. Amer. Acad.,
6. 75, 1879 ; G. A. Liebig, Amer. Journ. Science, (3), 26. 57, 1883 ; A. Bartoli and E. Stracciati,
Nuovo Cimento, (3), 32. 19, 215, 1892 ; A. Cotty, Ann. Chim. Phys., (8), 24. 282, 1911 ; M. Marti-
netti, Atti Accad. Torino, 25. 565, 1890 ; J. Dewar, Proc Roy. Soc, 76. A, 325, 1905 ; J. Narbutt,
Phys. Zeit., 19. 513, 1918 ; F. G. Jackson, Journ. Amer. Chem. Soc, 34. 1470, 1912 ; P. Nord-
mever and A. L. Bernoulli, Ber. deut. phys. Ges., 5. 175, 1903 ; A. Bogojawlenskv, Schr. Dorpat
Nat. Ges., 13. 1, 1905; W. Nernst and F. Koref, Silzber. Akad. Berlin, 247, 262, 1910; ?L C.
Dickinson and N. S. Osborne, Journ. Franklin Inst., 179. 489, 1915 ; A. W. Smith, Phys. Rev,,
17. 193, 1903.
1" W. Nernst, Ann. Physik, (4), 36. 428, 1911 ; G. N. Lewis and G. E. Gibson, Jmbrn. Amer,
Chem. Soc, 39. 2554, 1917.
1^ W. Nernst and H. von Wartenberg, Zeit. fthys. Chem., 56. 543, 1906 ; A. Langen, Mitt.
Forsch. Ing. Wes., 8. 1, 1903 ; L. Holborn and F. Henning, Ann. Physik, (4), 18. 739, 1905 ; (4),
23. 809, 1907 ; W. Jager and H. von Steinwehr, Die Wdrmekapazitdt des Wasser zwischen 6° und
50° in internationalen Wattsekunden, Berlin, 1915 ; M. Pier, Zeit.Eiectrochem., 15. 536, 1909 ; A.
Jaquerod, Recherches sur les conductibilitks electriques, les densit4s,et les chaleurs specifiques des solutions
de chlorure depotassium et de potasse caustique, Geneve, 1901 ; K. Pnschl, Monatsh., 22. 77, 1901.
• 12 E Mallard and H. le Chatelier, Compt. Rend., 93. 1014, 1076, 1881 ; M. Berthelot and
P. Vieille, ib., 98. 545, 601, 770, 852, 1884 ; Ann. Chim. Phys., (6), 4. 13, 1885 ; J. M. Gray, Phil.
Mag., (5), 13. 337, 1882 ; G. de Lucchl, Atti 1st. Veneto, (5), 7. 1305, 1881 ; R. Cohen, Wied. Ann.,
37. 628, 1889 ; L. Holborn and L. W. Austin, Phys. Rev., 21. 209, 1905 ; Wiss. Abhand. phys.
chem. tech.- Reichsanst., 4. 133, 1905; W. Treitz, Ueher die Fortpflanzungsgeschwindigkeit des
Schalles in einigen Ddmpfen, Bonn, 1903 ; W. Zenker, Die Vertielung der Wdrme auf der Erd-
oberfldche, Berlin, 1888,
13 M. Straneo, Atti Accad. Lincei, (5), 6. 262, 299, 1897 ; C. H. Lees, Proc Roy. Soc, 74. 337,
1905 ; Phil. Tran.9., 191. A, 399, 1898 ; R. Goldschmidt, Phys. Zeit. 12. 417, 1911 ; A. Winkel-
mann, Pogg. Ann., 153. 481, 1874 ; J. T. Bottomlev, Proc Roy. Soc, 28. 462, 1879 ; C. Christian-
s?n, Overs, dansk. Vid. Selsk. Forh., 183, 1889 ; J. D. Forbes, Trans. Roy. Soc Edin.,23. 133, 1862 ;
24. 73, 1865; A. C. Mitchell, ib., 34. 535, 1887; 35. 947, 1890; Proc Roy. Soc Edin., 13. 592, 1886; 17.
300, 1890 ; F. Neumann, Wied. Ann., 66. 286, 1898 ; T. Andrews, Proc Roy. Soc, 40. 544, 1881 ;
H. Abels, Rep. Meteorol. Acad. St. Petersburg, 16. 53, 1892 ; S. A. Hjelstrom, Oefvers. Akad. Stock-
holm, 46. 669, 1889 ; P. Jansson, ib., 58. 207, 1901 ; C. G. Lundquist, Arskrift Univ. Upsala,
1, 1889 ; H. F. Weber, Wied. Ann., 10. 103, 1880 ; 11. 345, 1880 ; S. R. Milner and A. P. Chat-
tock, Phil. Mag., (5), 48. 46, 1899 ; R. Weber, Ann. Physik, (4), 11. 1047, 1903 ; L. Graetz, Wied.
Ann , 18. 79, 1883 ; R. Wachsmuth, ib., 48. 158, 1893 ; C. Chree, Proc Roy. Soc, 43. 30, 1888 ;
T. Okada, Journ. Meteor. Soc. Japan, 24. 1, 1905.
1" Isaac Newton, Opticks, London, 249, 1750,
1^ A. des Cloizeaux, Manuel de miniralogie. Paris, 7, 1862 ; E. Reusch, Pogg. Ann., 121. 575,
1864 ; C. Meyer, Wied. Ann., Zi.S2\. 1887 ; E. Ketteler, ib., 33. 608, 1888 ; C. Pulfrich, ib.. 34
326, 1888 ; L. Lorenz, ib., 11. 82, 1880 ; A. Bertin, Ann. Chim. Phys., (5), 13. 283, 1878 ; F. T.
Trouton, Proc Roy. Soc Dublin, 8. 691, 1898.
i« M. Croullebois, Ann. Chim. Phys., (4), 22. 139, 1871 ; Compt. Rend., 70. 847, 1870 ;
C. Cheneveau, ib., 156. 1972, 1913 ; A. Cornu, ib., 70. 989, 1870 ; J. Jamin, ih., 70. 966, 1870 ;
43. 1191, 1856; J. H. Gladstone and T. P. Dale, Phil. Trans., 153. 323, 1863; 148. 887, 1858;
VOL. I. 2 I
482 INORGANIC AND THEORETICAL CHEMISTRY
176. 887, 1886; W. J. Pope, Zeit KrysL, 28. 116, 1897 ; J. W. Briihl, Ber., 24. 1, 648, J891;
Zeit. phys. Chem., 7. 1, 1891 ; P. Schult, ib., 5. 358, 1890 ; V. S. van der Willigen, Arch.
Musie Teyler, 1. 115, 1868; J. Kannonikofi, Joiirn. prakt. Chem., (2), 31. 352, 1885; E. van
Aubel, Phys. Zeit., 14. 302, 1913; W. Ruhlmann, Pogg. Ann., 132. 1177, 1867; H. A. Lorenz,
Wied. Ann., 11. 70, 1880; E. Kettler, ib., 33. 353, 506, 1888; B. Walter, ib., 38. 107, 1889;
46. 422, 1892; C. Bender, ib., 39. 89, 1890; 68. 343, 1899; 69. 676, 1899; H. Dufet, Journ.
Phys., (2), 4. 401, 1885; J. Conroy, Proc. Boy. 8oc., 58. 228, 1895; J. W. Gifford, ib., 70. 329,
1902; 78. 406, 1907; G. P. Baxter, L. L. Burgess, and H. W. Daudt, Journ. Amer. Chem, Soc.,
33. 893, 1911; F. A. Osborn, Phjs. Rev., (2), 1. 198, 1913.
" W. Spring, Arch. Sci. phys. nat, (4), 20. 101, 1905 ; Bvll. Acad. Belgique, (3), 5. 555, 1883 ,
(3), 31. 94, 256, 1896 ; (3), 34. 578, 1897 ; (3), 36. 266, 1898 ; Bee. Trav. Chim. Pays-Bas, 17.
202, 359, 1898 ; 18. 1, 1899 ; V. Meyer, Ber., 15. 297, 1882 ; 0. von Aufsess, Die Farbe der Seen,
Munehen, 1903 ; J. Aitken, Proc. Boy. Soc. Edinburgh, 11. 472, 637, 1882 ; R. Abegg, Naturwiss.
Bund., 13. 14, 1898 ; H. Davy, Collected Works, 9. 199, 1840 ; R. Bunsen, Pogg. Ann., 83. 197,
1851; LieMg's Ann., 72. 44, 1847; J. L. Soret, Ann. Chim.. Phys., (4), 17. 517, 1869;
E. Hagenbach, ib., (4), 20. 225, 1870; H. St. C. Deville, ib., (3), 23. 32, 1848; J. Tyndall,
Naturjorscher, 4. 1, 1871; A. Secchi, ib., 1. 149, 1868; G. C. Wittstein, Vierteljahr. prakt.
Pharm., 10. 342, 1861 ; Lord Rayleigh, Nature, 83. 48, 1910; W. D. Bancroft, Journ. Franklin
Inst., 187. 268, 459, 1919; E. Bourcart, Arch. Sciences Geneve, (4), 17. 169, 1904; A. A. Hayes,
Amer. Journ. Science, (2j. 49. 180, 1870; Compt. Bend., 68. 911, 1869 ; F. Boas, Beitrdge zur
Erkenntnis der Farbe des Wassers, Kiel, 1881 ; J. Duclanx and E. Wollman, Journ. Phys., (5),
2. 263, 1912.
^* F. Boas, Beitrdge zur Erkenntnis der Farbe des Wassers, Kiel, 1881 ; J. L. Soret, Arch.
Sciences Phys. Nat., (3), 11. 276, 1884 ; P. Desains, Compt. Bend., 94. 1144, 1882 ; W. C. Rontgen,
Wied. Ann., 23. 1, 259, 1884; A. Secchi, Naturjorscher, 1. 149, 1868; 0. von Aufsess, Die
Farbe der Seen, Miinchen, 1903 ; W. H. Julius, Verh. Akad. Amsterdam, 1. 1, 1892 ;
F. Paschen, Wied. Ann., 53. 334, 1894 ; E. Aschkinass, ib., 55. 406, 1905 ; W. de W. Abney and
E. R. Festing, Phil. Trans., 172. 887, 1882 ; W. W. Coblentz, Journ. Franklin Inst., 172. 309,
1911 ; G. Bode, Ann. Physik, (4), 30. 326, 1909.
" G. Magnus, Sitzber. Akad. Berlin, 246, 1861 ; 1128, 1861 ; 572, 1862 ; 149, 1863 ; 73,
1866 ; Pogg. Ann., 112. 497, 1861 ; 114. 635, 1861 ; 118. 575, 1863 ; 121. 186, 1864 ; 127. 613",
1866 ; 130. 207, 1867 ; J. Tyndall, Proc. Boy. Soc, 10. 37, 1860 ; 11. 558, 1861 ; 30. 10, 1879 ;
31. 307, 478, 1881 ; 35. 21, 1883 ; Phil Mag., (4), 22. 377, 1861 ; (4), 23. 252, 1862 ; (4), 26. 44,
1863 ; (4), 32. 118, 1866 ; (4), 33. 425, 1867 ; Phil. Trans., 151. 1, 1861 ; 152. 69, 1862 ; 153.
1, 186.3 ; 154. 201, 1864 ; 154. 327, 1864 ; 173. 291, 1882 ; H. Wild, Pogg. Ann., 129. 57, 1866 ;
J. L. Hoorweg, ib., 155. 385, 1875 ; H. Buff, ib., 158. 177, 1876 ; P. M. Garibaldi, Nuovo Cimento,
(2), 3. 231, 1870 ; S. A. HiU, Proc. Boy. Soc, 33. 216, 435, 1881 ; J. G. MacGregor, Proc Boy.
Soc Edin., 12. 24, 1882 ; G. Neumayer, Phil. Mag., (4), 31. 510, 1886 ; E. Lecher and F. Pemter,
Sitzber. Akad. Wien, 82. 265, 1880 ; E. Lecher, ib., 82. 851, 1880 ; AV. C. Rontgen, Wied. Ann.,
12. 155, 1881 ; 23. 1, 259, 1884 ; H. Heine, ib., 16. 441, 1882 ; H. Haga, Ueber die Absorption
der strahlenden Wdrme durch WasserdampJ, Leiden, 1876 ; K. Angstrom, Wied. Ann., 39. 267,
1890 ; Ann. Physik, (4), 6. 163, 1901 ; S. P. Langley, Proc Nat. Acad. Sciences, 4. 197, 1889.
2» L. Sohncke, Wied. Ann., 28. 550, 1886 ; J. Elster and H. Geitel, ib., 32. 74, 1887,
21 P. Thomas, Journ. Franklin Inst., 176. 283, 1913.
22 G. Foussereau, Compt. Bend., 99. 00, 1884 ; F. Beijerinck, Neues Jahrb. Min. B. B., 11. 403,
1897.
2» P. Kohlrausch and A. Heydweiller, Zeit. phys. Chem., 14. 317, 1894 ; Wied. Ann., 53. 209,
1894 ; F. Kohlrausch, Proc Boy. Soc, 71. 338, 1903 ; W. Nernst, Theoretical Chemisiiy, London,
420, 1910 ; Zeit. phys. Chem., 14. 155, 1894 ; W. Ostwald, ib., 11. 52, 1893 ; R. Lowenherz, ib.,
25. 283, 1896 ; S. Arrhenius, ib., 11. 805, 1893 ; J. J. A. van Wijs, ib., 11. 492, 1893 ; 12. 514,
1893 ; 14. 789, 1894 ; A. A. Noyes, Y. Kato, and R. B. Sosman, Journ. Amer. Chem. Soc, 32.
154, 1900 ; C. W. Kanolt, ib., 29. 1402, 1907 ; H. C. Jones and E. C. Bingham, Amer. Chem.
Jo^trn., 34. 481, 1905; J. Negreanu, Bull. Soc Bucuresei, 15. 271, 1908; J. J. Thomson, Proc
Boy. Soc, 45. 269, 1889 ; J. A. Fleming and G. B. Dyke, Journ. Inst. Elect. Eng., 49. 323, 1912 ;
B. van der Pol, Phil. Mag., [6), 36. 88, 1918; J. Shields, ib., (6), 35. 365, 1893; H. Lunden,
Affinitdtsmessungen an schwachen Sduren und Basen, Stuttgart, 1908; S. P. L. Sorensen,
Biochem. Zeit., 21. 191, 1909; S. P. L. Sorensen and S. PaUtzsch, ib., 24. 387, 1910; R. Lorenz
and A Bohi, Zeit. phys. Chem., 66. 733, 1909 ; L. Michaelis, Die Wasserstoffionenkonzentration,
Berlin, 1914; K. T. Compton, Phys. Bev., (2), 8. 412, 1916.
" J. W. Bruhl, Ber., 28. 2866, 1895.
" K. Badeker, Zeit. phys. Chem., 36. 305, 1901; M. Jona, Phys. Zeit., 20. 14, 1919;
R. Abegg, Zeit. phys. Chem., 29. 491, 1899; J. A. Fleming and J. Dewar, Proc. Boy. Soc, 61. 2,
1897; G- Gutton, Compt. Bend., 130. 1119, 1890; C. B. Thwing, Zeit. phys. Chem., 14. 280,
1894 ; F. Heerwagen, Wied. Ann., 49. 272, 1893 ; P. Drude, ib., 59. 17, 1896 ; Zeit. phys.
Chem., 23. 267, 1897 ; E. Bouty, Journ. Phys., 1, 1892 ; B. Hopkinson and E. Wilson,
Phil. Trans., 189. A, 109, 1897; R. Blondlot, Compt. Bend., 119. 95, 1894; A. Perrott, 16.
119, 601, 1894; H. Merczvng, Ann. Physik, (4), 34. 1015, 1911 ; H. Rukop, ib., (4), 42. 489,
1913 ; A. Colley, Phjs. Zeit., 10. 471, 1909 ; Journ. Bussian Phys. Chem. Soc, 38. 431, 1906 ;
J. F. Smale, Wied. Ann., 57. 215, 1897; R. Abegg, ib., 65. 229, 1898; P. Drude, ib., 59. 17,
1897; Zeit. phys. Chem., 23. 267, 1897 ; B. B. Turner, ib., 35. 185, 1900; E. Beaulard, Compt.
WATER 483
Rend.y 144. 904, 1907 ; C. Niven, Proc. Roy. Soc, 85. A, 139, 1911 ; U. Behn and K. Kiebitz,
Boltzmann's Festschrift, Leipzig, 610, 1904; W. D. Coolidge, Wied. Ann., 69. 125, 1899;
K. Cohn and L. Arons, ib., 33. 13, 31, 1888; E. A. Harrington, Phys. Rev., (2), 8. 681, 1916.
2« J. J. Thomson, Phil. Mag., (5), 36. 320, 1893 ; W. Nemst, Zeit. phys. Chem., 13. 531, 1894.
27 P. Dutoit and E. A. Aston, Compt. Rend., 125. 240, 1897 ; P. Dutoit and L. Friderich,
Bull. Soc. Chim., (3), 19. 321, 1898.
28 C. Brunner, Pogg. Ann., 79. 173, 1850.
29 A. Piccard, Compt. Rend., 155. 1497, 1912 ; R. Weiss and A. Piccard, ib., 155. 1234, 1912 ;
P. Seve, Ann. Chim. Phys., (8), 27. 189, 1913.
3» H. C. Hayes, Phys. Rev., (2), 3. 295, 1914 ; H. D. Stearns, ib., 16. 1, 1903 ; A. P. Wills,
ib., 20. 188, 1905; Phil. Mag., (5), 45. 432, 1898; H. du Bois, Wied. Ann., 35. 137, 1888;
J. Koenigsberger, ib., 66. 098, 1898; Ann. Physik, (4), 6. 500, 1901 ; G. Quincke, Wied. Ann., 24.
347, 1885 ; 34. 401, 1888; G. Jaeger and St. Meyer, ib., 67. 427, 1899; Sitzber. Akad. Wien, 106.
594, 623, 1897 ; 107. 5, 1898 ; J. A. Fleming and J. Dewar, Proc. Roy. Soc, 60. 283, 1896 ; 63.
311, 1898 ; G. Piaggesi, Phys. Zeit., 4. 347, 1904 ; A. Scarpa, Nuovo Cimento, (5), 10. 155, 1905;
U. Meyer, Ann. Physik, (4), 30. 630, 1909.
31 L. H. Siei^tsema, Proc. Acad. Amsterdam, 5. 131, 1887; Arch. Nierl., (2), 6. 830, 1901;
F. Agerer, Sitzber. Akad. Wien,lU. 830, 1905 ; G. Quincke, Wied. Ann., 24. 609, 1885 ; L. Arons,
ib., 24. 609, 1885 ; J. W. Rodger and W. Watson, Zeit. phy.i. Chem., 19. 357, 1896 ; U. Meyer,
Ann. Physik, (4), 30. 639, 1909; J E. H. Gordon, Phil. Mag., (5), 1. 73, 1876.
32 J. Kerr, Phil. Mag., (4), 50. 337, 1875; G. Lemoine, Compt. Rend., 122. 835, 1896;
R. Leiser, Elektrische Doppelbrechung der Kohtenstojfverbindun^entliallea,. S., 1910 ; J. W. Rodger
and W. Watson, Zeit. vhys. Chem., 19. 357, 1896.
33 A. L. Hughes, Phil. Mag., (6), 24. 380, 1912; W. Oholensky, Ann. Physik, {i), 39. 961.
1912.
§ 9. The Chemical Properties of Water
Water is formed by the direct union of the two elements, such as when one
of the elements is burnt in the presence of the other; by the combustion of
compounds containing hydrogen in air or oxygen ; by the action of hydrogen
on certain oxides or compounds containing oxygen ; and by the decomposition
of many organic and inorganic compounds containing hydrogen and oxygen.
Water formed by the combustion of hydrogen in oxygen is reported to contain
sometimes hydrogen peroxide, and ozone ; and if the oxygen contains nitrogen,
both nitric acid and ammonium nitrite may be formed. M. Berthelot i has
reported that when a gram of hydrogen was burnt in air, 0"000075 grm. of nitric
acid, HNO3, was simultaneously formed ; and in oxygen containing 8 per cent, of
nitrogen, 0'017 grm. of nitric acid was obtained with a small flame, and 0"071 grm.
with a large flame.
A mixture of two volumes of hydrogen and one volume of oxygen is variously
styled detonating gas, electrolytic gas, gaz tonnant, and Knallgas. The constituents
of detonating gas unite to form water (i) when heated ; (ii) when exposed to an
electric spark, or to the silent electrical discharge ; 2 (iii) when merely in contact with
certain metals — particularly the platinum family — charcoal, etc. ; (iv) when
directly exposed to radium rays ; ^ but, according to W. P. Jorissen and W. E.
Ringer, there is no perceptible action if the detonating gas in a glass vessel be
exposed to radium rays. S. C. Lind found that the velocity of combination of
hydrogen and oxygen under the influence of radium emanation is proportional to
the amount of emanation present at any time, and to the gas pressure such that
log {plpQ)=kE{e~^*—l), where Eq denotes the initial amount of emanation expressed
in curies which is decaying proportionally to the factor e~^^ ; A; is a constant ; and
Pq and p mm. are respectively the initial pressure and the pressure at the time t.
Increasing the volume of the gas decreases the velocity constant such that in a
spherical bulb of diameter d, k=84:'ld"^. An excess of oxygen gives a velocity
constant higher than normal ; an excess of hydrogen acts in the opposite direction.
For each pair of ions produced by the emanation, about 3*9 molecules of water
are formed. From A. T. Cameron and W. Ramsay's measurements, S. C. Lind
showed a parallelism between chemical action and ionization, but 0. Scheuer found
about 5"5 molecules of gas combined for each pair of ions produced, and E. Wourtzel
found that in general, the amount of reaction is in excess of ionization. This led
484 INORGANIC AND THEORETICAL CHEMISTRY
A. Debierne to reject the assumption that ionization is the primary cause of the
reaction, and he suggested that the passage of an a-particle through a gas may
thermally decompose molecules lying outside the ionization zone. S. C. Lind
explains the apparent discrepancy on the assumption that the recoil atoTns contribute
to the ionization. When an atom emits an a-particle at a high speed, the residual
atom recoils with a velocity about ~th of that of the emitted particle, and the
velocity and corresponding kinetic energy suffice to give the recoil atoms ionizing
properties. The resulting calculation shows that there is a statistical agreement
between the actual number of ions produced and the number of reacting molecules.
According to T. de Saussure,* (v) various organic substances (peas, corn, and
humus) in the act of decomposition may stimulate the union of hydrogen and
oxygen gases, (vi) J. B. Biot (1805) found that detonating gas combined under
pressure in an iron tube, but heat developed by the compression may have
raised the mixture to the temperature of ignition, for F. de la Roche (1811)
observed no combination under a pressure of 50 atm. applied gradually ; nor
did A. F. E. Degen (1836) observe any signs of the recombination of the gases
from the electrolysis of sulphuric acid, at a pressure of 150 atm. H. N. Warren
(1893) states that an explosion with the production of flame occurs under these
conditions at 180 atm. pressure, but it is not clear from his account whether or not
combination occurred — his tubes always burst — he may have simply compressed
the gas to the bursting pressure of the tube, (vii) Electrolytic gas may be confined
an indefinite time over water or mercury, at ordinary temperatures, in darkness,
with no tangible sign of chemical action, but according to H. B. Baker, if detonating
gas be exposed to sunlight ^ there is a slow combination. This no doubt explains
the contradictory observations of B. Hooke (1803) and T. de Saussure (1815), for
the former claimed that the gases slowly combine on standing some months, while
the latter contradicted this statement.
V. Meyer and W. Raum ^ could detect no sign of combination when a mixture
of hydrogen with half its volume of oxygen was heated in glass bulbs for 218 days
at 100° ; at 300° the formation of water could just be detected after 65 days ; at
350°, in four bulbs, between 0*5 and 1*9 per cent, had combined in 5 days, and in
one bulb 16'4 per cent, in the same time. W. A. Bone attributes this to a slight
devitrification of the glass because he found similar accelerations in the speed of
the reaction with bulbs partially devitrified. At 400°, says W. A. Bone, we are on
the border line where the formation of water may be recognized within a week, but
hardly within three days. The reaction then progresses the faster the higher the
temperature. H. Helier passed the mixed gases through a glazed porcelain tube
packed with pieces of porcelain so as to present a large surface to the gas ; he could
detect but a slight combination at 180°, and found that the reaction progresses
faster and faster as the temperature is raised. At any assigned temperature, there
is a limit beyond, which no further combination ensues. For example, the percentage
amount of water formed at difierent temperatures was found to be as follows :
200=^
260°
300°
311°
376°
498°
637°
825°
012
1-6
3-8
9-8
251
56-4
85-6
96-1
At 840° the mixture exploded. Water vapour does not dissociate at these tempera-
tures, and accordingly there appear to be two equilibrium states for the reaction :
2H20^2H24-02, according as equilibrium is approached by decomposing water
vapour, or by uniting the elementary gases. Does not this conflict with the
general theory of mass action ? M. Bodenstein believes that a true state of
equilibrium was not obtained by H. Helier, and that if he had heated his mixtures
longer, no discrepancy would have been detected. On the other hand, P. Duhem
believes that the system in H. Helier's experiment adquired a passive condition
called a state oi false equilibrium. Great differences in speed, of the reaction have
been obtained by different experimenters, and M. Berthelot has shown that this
must be largel}i due to the nature of the surface exposed to the gas. Indeed, there
WATER 485
are many reasons for supposing that the reaction takes place only on the surfaces
of the solid ; but with the facts available it is at present impossible to infer, with
any degree of certainty, what would be the course of the reaction in a vessel with
walls absolutely inert.
The union of hydrogen and oxygen, and the decomposition of water vapour
under the influence of ultra-violet light in quartz vessels at 150° and at 240° shows
that a definite equilibrium — the same at both temperatures— is attained, 2H2+O2
^2H20. The work done by the light rays in the reaction 2H20^2H2+02 is
about 44 "5 Cals. per gram-molecule of water. Small quantities of hydrogen dis-
appear owing to the reduction of silica by moist hydrogen when stimulated by active
light rays. No signs of hydrogen peroxide have been detected in the equilibrium
mixture, but if a fast stream of the two gases passes through the apparatus, hydrogen
peroxide can be detected in the water which is formed ; this is thought to indicate
that the union of hydrogen and oxygen takes place in two stages : 2H2+02=H202
H-H2=2H20.
Ignition temperatures. — H. Davy,^ and T. von Grotthus found that a mixture
of hydrogen gas and air heated to a temperature below visible redness, rapidly
unites to form water without the evolution of light or heat. The temperature at
which detonating gas inflames has been measured by E. Mallard and H. le Chatelier
(1880) and many others. The numbers are widely divergent. By plunging a
bulb containing mixed gases in a bath at a constant temperature, numbers ranging
from 518° to 650° have been obtained, and higher results are obtained if an excess
of either gas be present ; by passing a stream of the mixed gases through a tube
in a bath heated on a gradually rising temperature, numbers ranging from 550° to
845° have been reported ; and by measuring the adiabatic compression required
to just ignite the gas, and calculating the corresponding temperature, K. G. Falk
obtained 540° for the mixture 2H2+O2 ; 514°, for H2+O2 ; and 530°, for H2+2O2.
According to H. B. Dixon (1910), the ignition temperatures of mixtures of 100 vols,
of hydrogen with n vols, of oxygen, by adiabatic compression, are :
Vols, of oxygen
. 33-33
40
50
100
200
300
400
Ignition point
. 557°
542°
536°
530°
520°
512°
507°
Hence, the most easily ignited mixture is not one in which the proportion of hydrogen
to oxygen is as 2 : 1, but when the ratio is 1 : 4. H. B. Dixon and H. F. Coward
showed that the ignition temperature falls with increasing pressures ; and that
a stream of hydrogen at a normal pressure ignites between 580° and 590° in an
atmosphere of oxygen, and virtually the same in air. The hydrogen gas was
led through a narrow tube in the axis of a larger porcelain tube through which a
slow current of oxygen or air was passed. The outer tube was heated by an electric
current traversing a coil of wire. A constant ignition point was obtained when the
diameter of the outer tube and the speed of the current of gas surpassed a certain
minimum value. With an orifice 1 mm. diameter, and an outer tube 4*5 cm.
diameter, the gas had to be passed at a rate exceeding 9 c.c. per minute to give
constant results. The ignition temperatures obtained with the gases in sealed
bulbs are rather lower than the ignition points of flowing gases. According to
H. F. Coward, C. Cooper, and C. H. Warburton, a flame which filled a 570 c.c. globe
has been obtained with the electrolytic gas at a pressure of 5 mm., and a flame
travelled through a cylinder 2 metres long with a gas at 8 mm. pressure. These
pressures are lower than have been obtained by others, for previous records gave
minimum pressures between 34 and 146 mm. The ignition temperature is largely
determined by the nature of the spark discharge. The small amount of gas— 0*3
to 5'5 per cent. — which remained uncombined, presumably owing to the cooUng
of the walls, varied inversely as the original pressure of the gas, and was greater
the larger the globes.
Lord Rayleigh (1875) has shown that energy is dissipated when a mixture of hydrogen
and oxygen at atmospheric pressure is exploded by an electric spark. The spark itself
486 INORGANIC AND THEORETICAL CHEMISTRY
can be neglected because any given spark can explode an indefinitely large quantity of
the mixture. If the gas is expanded at constant temperature before the explosion, and,
after the explosion, brought back to its former volume, the pressure required to compress
the steam will be less than that exercised by the same volume of uncombined gas, and
accordingly work is gained in the operation, for less energy is dissipated during the explosion
of the expanded gas than is dissipated by the explosion of the condensed gases. If the
expansion be increased without limit, the amount of energy dissipated during the explosion
becomes indefinitely small ; otherwise expressed, the tendency to combine diminishes the
rarer the gas, and there must be a point where the gas becomes so rare that the explosion
will not take place.
According to H. B. Dixon (1884:),7the speed of the explosion wave in long tubes
containing mixtures of hydrogen and oxygen, WH2+O2, is
w . . 8 6 4 2 1 i i
Speed . 3532 3527 3268 2821 2328 1927 1707 metres per sec.
The speed is thus greater in an excess of the specifically lighter gas, hydrogen, than
in an excess of the specifically heavier gas, oxygen.
The catalytic action of metals, etc., by contact. — The combination of hydrogen
and oxygen can be inaugurated at ordinary temperatures where the most refined
observations show no signs of chemical combination, and, where the temperature
is such that the gases are actually combining, the speed is accelerated by the mere
presence of a number of different substances — by finely divided platinum in
particular. The phenomenon was discovered by H. Davy in 1817,^ for he remarked
that platinum foil or wire heated to a temperature " short of redness " will induce
the combination of oxygen with other inflammable gases or vapours ; and
P. Erman (1819) showed that a temperature of 50°-51° suffices. After
E. Davy (1820) had shown that when finely divided platinum damped with spirits
of wine became incandescent owing to the heat generated by the oxidation of the
alcohol, J. W. Dobereiner described in a brochure, Ueher neu entdeckte hochst
merkwiirdige Eigenschaften des Platins (Jena, 1823), how finely divided platinum
will spontaneously induce the rapid combination of hydrogen and oxygen gas.
A. Pleischl noticed that when platinum wire has been in the hydrogen flame for
some time it becomes corroded, dull, and dark grey ; and that the corroded part
subsequently gets hot quickest in the jet of hydrogen gas. W. A. Bone also found
that the appearance of a piece of silver gauze before and after it has acted as a
catalytic agent for hydrogen and oxygen at 400° changed so that the wiies originally
smooth became quite rough. These facts are taken to show that the reacting gases
during the catalytic action are in very intimate contact with the surface film of
metal. P. L. Dulong and L. J. Thenard found that while a coil of new platinum
wire — 0*05 mm. diameter — must be heated to 300° to make it effective, after
several ignitions, it acts as low as 50° or 60°. If the wire be immersed in hot or
cold nitric acid, and dried at 200°, it acts at the ordinary temperature of the air,
and becomes red hot in a jet of a mixture of air and hydrogen ; sulphuric acid acts
similarly but is less effective ; and hydrochloric acid is still less effective. Potassium
or sodium hydroxide destroy the activity of the metal.
P. L. Dulong and L. J. Thenard reported that the property acquired by platinum,
by the acid treatment, persists but a few hours when the metal is exposed to the
air, and 24 hours if confined in a vessel. The treated metal loses its property if
exposed for five minutes to a current of dry air, oxygen, hydrogen, or carbon
dioxide ; ammonia or the alkali hydroxides deprive the metal of its peculiar
power. Platinum foil and platinum filings when fresh and clean exploded a
mixture of hydrogen and oxygen. They lose this quality if exposed to air for an
hour or two, but the property is recovered if the metal be ignited in a covered
crucible. Platinum filings prepared under water have no action. The less active
forms of the metal may induce the combination of the mixture of hydrogen
and oxygen without causing explosion. M. Faraday showed that a platinum
plate must have its surface rigorously clean if it is to effect the combination of
WATER 487
detonating gas, and lie described several methods of cleaning platinum plates in
order to make them active stimulants of the union of hydrogen and oxygen. He
also found that the surface of the plates loses its activity on exposure to air for
24 hours, but regains it when gently ignited. A platinum plate is active when it
has been used as the anode in sulphuric, nitric, oxalic, tartaric, citric, or acetic
acid, or in a solution of potassium phosphate, chlorate, or nitrate, or of sodium or
copper sulphate ; it acquires less power in hydrochloric acid, still less in potassium
or sodium carbonate, and none at all in potassium hydroxide.
According to J. W. Dobereiner, spongy platinum is effective at ordinary tempera-
tures in inducing the combination of hydrogen and oxygen ; the action is at first
slow, but as the temperature rises, the action is very fast ; while A. Pleischl, and A. de
la Rive and F. Marcet found the ash of filter paper saturated with ammonium
chloroplatinate is even more efiective than spongy platinum, for the action was
then apparent at —20°. E. Davy and J. W. Dobereiner found that very finely
divided platinum — ^platinum black — is still more active. Hence, the more finely
divided the platinum, the greater its efficiency in stimulating the union of a mixture
of hydrogen and oxygen. Platinum black which has not been freed from oxygen
can induce the union of the two elements at the temperature of liquid air, —190°.
J. W. Dobereiner, P. L. Dulong and L. J. Thenard, M. Faraday, A. Pleischl,
W. C. Henry, and A. de la Rive and F. Marcet obtained similar results with iridium
and palladium ; osmium acts at 40° or 50° ; spongy rhodium at 240° ; gold leaf at
260° ; gold paper ash at 50° ; and silver leaf below 357° ; silver paper ash, 120°
to 150°. W. C. Henry found finely divided copper at 264° does not set fire to a
stream of hydrogen but the copper oxidizes ; at higher temperatures the powder
becomes red hot. E. D. Campbell found an alloy of copper with one per cent, of
palladium — palladized copper — stimidates the union of hydrogen and oxygen ;
cohalt and nickel behave similarly; while P. L. Dulong and L. J. Thenard say copper
and nickel act at 300°. W. C. Henry found that iron reduced from the oxide induces
rapid combination of hydrogen and oxygen at the temperature of reduction. Copper
or iron turnings, zinc foil, and charcoal were reported by W. C. Henry to have no
action, but P. L. Dulong andL. J. Thenard found charcoal, pumice stone, porcelain, rock
crystal, and glass to act below 350° ; the action of fluorspar is very feeble. P. L.
Dulong and L. J. Thenard say that angular pieces of glass are twice as active as
rounded pieces of equal surface in stimulating the union of hydrogen and oxygen.
Mercury at its boiling point has no appreciable action. Devitrified glass bulbs
were found by W. A. Bone to accelerate the reaction at about 300° ; and M. Berthelot
found barium hydroxide, alkali salts, and manganese salts raised the speed of the
reaction between 250° and 300°.
R. Bottger found that ammoniag&s destroys the activit3^of platinum — even a drop
of a solution of ammonia evaporating in a room suffices. The vapour of nitric acid
or chlorine restores the activity removed by ammonia. J. S. C. Schweigger adds
that hydrogen sulphide, ammonium sulphide, and particularly carbon disulphide
render the metal inactive, and the activity is not restored by nitric acid or chlorine.
W. Art lis showed that the activity may be also destroyed by the traces of hydrogen
sulphide present as impurities in hydrogen gas, and M. Faraday showed that
hydrogen prepared by the decomposition of water by red hot-iron is not affected
by either spongy platinum or platinum foil — presumably because the activity of
the catalyst is quelled by the impurities in the gas. According to E. Turner, finely
divided platinum does not become inactive when confined 24 hours over mercury
in an atmosphere of oxygen, hydrogen, carbon dioxide, or air ; it loses part of its
power after 5 min. exposure to hydrogen chloride, still more in ethylene or coal
gas ; and it becomes inactive in sulphur dioxide, hydrogen sulphide, or ammonia.
A little moisture favours the action, for W. French (1900) showed that if the
mixture be thoroughly dried, finely divided platinum does not start the reaction.
If the metal be wet with water, the action is feeble at first but gradually increases
as the water evaporates ; similar remarks apply if the metal be wetted with alcohol
488 INORGANIC AND THEORETICAL CHEMISTRY
or ether ; and if wetted with sulphuric, nitric, or hydrochloric acid, the metal is
inert. W. Henry also found that spongy platinum produces its effect if one volume
of detonating gas be mixed with ten volumes of oxygen, hydrogen, nitrogen, or
methane ; or with six volumes of hydrogen chloride ; but it is either prevented
or very much retarded by 11 vols, of nitrous oxide ; 3 vols, of carbon dioxide ;
1 J vols, of ethylene ; 1 vol. of cyanogen ; or 0*5 vol. of carbon monoxide. Spongy
platinum is inactive in a mixture of detonating gas with an equal volume of carbon
monoxide, hydrogen sulphide, or ethylene, but platinum black becomes red hot
and combination rapidly occurs. E. Turner, M. Faraday, and W. C. Henry also
made observations on the retarding effects of carbon monoxide, sulphur dioxide,
hydrogen sulphide, hydrogen chloride, ammonia, nitrous oxide, carbon disulphide,
and ethylene on the union of detonating gas by finely divided platinum.
According to T. Graham, only impure ethylene retards the activity of spongy
platinum ; the purified gas does not affect the metal, and he states that in a mixture
of ethylene, hydrogen, and oxygen, spongy platinum acts only on the hydrogen, not
on the ethylene, so that the two gases may be separated by this agent. W. Henry
also noticed that if oxygen be mixed with hydrogen, carbon monoxide, methane,
and nitrogen gases, and passed over spongy platinum at 177°, the hydrogen and
carbon monoxide are alone oxidized. W. Hempel based a method for analyzing
certain mixtures of gases on the power possessed by finely divided palladium or
platinum of inducing -pTeieTential fractional combustion in this manner. H. Landolt
also found that hydrogen burns more readily than methane or ethylene in a flame
when in the presence of platinum, but W. A. Bone has shown the need for a repetition
of these experiments. W. A. Bone and his co-workers found that in explosive
reactions the affinity of methane is at least 20 to 30 times greater than that
of hydrogen for oxygen. At high temperatures the specific nature of the solid
catalytic agent becomes negligible because the reaction takes place so rapidly in
the gas phase. W. A. Bone found that hydrogen burns more readily than methane
at 500° in the presence of broken firebricks (grog) ; in borosilicate glass tubes,
at 300°- 400°, methane, ethylene, and acetylene are burned more quickly than
hydrogen or carbon monoxide ; E. Jager says that in the presence of copper oxide
at 250°, all the hydrogen in a mixture of hydrogen and methane can be burned
without decomposing the methane. W. D. Bancroft adds that at low temperatures
the nature of the catalytic agent may determine which of two combustible gases
will bum the more readily. Since charcoal causes the oxidation of ethylene to
carbon dioxide and water, and since charcoal has very little effect on a mixture of
hydrogen and oxygen, it is probable that charcoal will cause the preferential
combustion of ethylene in a mixture of ethylene and hydrogen.
There have been many attempts to explain the mechanism of the catalytic
activity of platinum. J. W. Dobereiner and J. S. C. Schweigger vaguely attributed
the phenomenon to the electrical reactions between the different substances con-
cerned in the action. One of the oldest hypotheses is due to A. Fusinieri (1825)
and M. Faraday (1834). It refers the action to the absorption or the condensation
of the reacting gases on the surface of the metal, and the enhanced faculty of combi-
nation possessed by the gases in this condensed state. J. Babinet, L. Meyer,
G. H. Quincke, and J. J. Thomson also attribute the action in part to the change
in the physical condition of the molecules of the reacting gases in contact with the
catalytic agent. M. Bodenstein's measurements of the rate of the reaction at
the higher temperatures (short of explosion) correspond with a termolecular reaction,
2H2+02=2H20, whereas at lower temperatures the reaction is monomolecular,
for the rate of the reaction is then proportional to the pressure of the gas — the rate
of absorption of each gas is also proportional to its pressure. Hence it is inferred
that the main reaction at low temperatures takes place on the surface of the metal,
and is dependent on the rates of absorption of the two gases by the metal, and also
on the rate of diffusion of the products of the reaction away from the seat of the
reaction on the surface of the metal. C. Ernst also found that the rate of combination
WATER 489
of hydrogen and oxygen dissolved in water in contact with electrolytic gas is
proportional to the rate of solution of the mixed gases, which in turn is proportional
to the pressure of the gases lying above the surface of the water. M. Bodenstein
found similar results for the action of Bredig's colloidal platinum in water.
This agrees with the observations of W. A. Bone and R. V. Wheeler, W. A.
Rowe, etc.
Another hypothesis, suggested by A. de la Rive and F. Marcet (1839) and by
C. Engler and L. Wohler (1901), assumes that there is a rapidly alternating series of
oxidations and reductions of the surface of the metal. A platinum oxide is first
supposed to be formed : Pt+%02=Pt02n ; the oxide is immediately reduced :
Vt02n \-2nH.2=2nI{20-\-Tt ; the reduced metal is re-oxidized, to be reduced again,
and so on indefinitely. In support of the view that the formation of an oxide of
platinum is an intermediate stage of the catalyzed reaction, oxidized platinum is
stated to be a more active catalytic agent than platinum alone. M. Berthelot and
P. Sabatier suggested that a hydride, not oxide of platinum, is formed as the inter-
mediate stage in the reaction.
How the metal carries oxygen to the hydrogen or hydrogen to the oxygen has
not been definitely established. W. D. Bancroft sums up the evidence by stating
that while the effect of platinum may be due to an oxidation, it is doubtful if this is
the case with charcoal and oxygen. In general, with contact catalysis : (i) onlv
those substances which are absorbed by a solid are catalyzed by it ; (ii) while the
catalytic action of solids may be solely the result of the increased surface concentra-
tion in some cases, this is not always the only factor ; (iii) a solid catalytic agent
may be considered as equivalent to a solvent and may therefore displace the
equilibrium ; (iv) as a result of selective adsorption we may get different reaction
products with different catalytic agents ; (v) a catalytic agent tends to produce
the system which it absorbs the most strongly.
I. Langmuir^ mounted a short filament of platinum in a 4-litre bulb ; this was
electrically heated in a mixture of hydrogen and oxygen in the absence of water
vapour, at low temperatures and low pressures, and he found that the rate of the
reaction V is directly proportional to the pressure of the oxygen, ^2j ^^^ inversely
proportional to the pressure of the hydrogen pi, so that V=kif2lPi, where ki is
a constant ; while at higher temperatures the rate varies with the partial pressures
of the two gases, so that the speed of the reaction ^=^-2^1^25 where Jc2 is a constant.
I. Langmuir here assumed that a certain fraction of the hydrogen molecules which
strike the surface of the platinum are condensed, and the layer of hydrogen thus
formed distils off at a certain rate ; he further assumes that the reaction occurs
when the hydrogen molecules strike oxygen on the surface, but not when oxygen
molecules strike hydrogen molecules.
The energy of the reaction between hydrogen and oxygen.— According to
W . G. Mixter ( 1 903) , the heat o! combustion of a gram of hydrogen at constant pressure
to form liquid water at 0° is 33993 cals. ; and the mean value of the observations
of other recent observers is 34022 cals. with a possible error of about y'^jth per cent.
The heat of formation Q, of a gram-molecule of water, H2H-i02=H20, at atmo-
spheric pressure, and 0°, is, according to M. Berthelot and C. Matignon 10 (]893),
70*4 Cals. for ice ; 69*0 Cals. for the liquid ; and 58-] Cals. for the vapour all at 0°.
The change in the heat of combination of hydrogen and oxygen to form a gram-
molecule of liquid water per degree change of temperature, can be calculated
from G. Eirchoff's equation dQldT=E{C—Ci), where EC denotes the sum of the
specific heats of the reacting products, and EC-^ that of the end-products. Taking the
molecular heat of hydrogen as 6*8 cals. : of oxygen, 6-96 cals. ; of water, 18 cals.,
it follows that (^Q/(^r=6-8-f 3-48— 18, and the heat of formation of a gram-molecule
of water decreases 7*72 cals. per degree rise of temperature. The heat of formation
Q of water vapour decreases perceptiblv with a rise of temperature; it is 50*5
Cals. at 2000°, and 371 Cals. at 40(X)°. " W. Nernst and H. von Wartenberg found
that at constant pressure, Qp=57200-fl •37^-0-000365 r2_o-Oe312r3 cals. per
490 . INORGANIC AND THEORETICAL CHEMISTRY
gram-molecule of the gas, and F. Haber and L. Bnmer, at a constant volume,
gp=57084-2976T-000125T2 cals.
A platinum plate charged with hydrogen, and a platinum plate charged with
oxygen, behave in conducting liquids as if they were electrodes made of the
respective gases which conduct electrically. If these electrodes be immersed in a
solution of an electrolyte, there is formed a voltaic combination, H2 | Aqueous
solution I O2 known as Grove's gas cell, which gives an electromotive force of ri5
volts, but the calculated value from II. von Helmholtz's equation (1847) furnishes
1'23 volts at 7° on the assumption that the hydrogen and oxygen are under a pressure
of one atmosphere. The free energy of the reaction H2+J02=H20 is therefore
equivalent to 475,000 joules. The discrepancy has been attributed to the formation
of an oxide of platinum at the oxygen electrode. E. Bose obtained 1-1392+0'015
volts from the gas battery, and this makes the heat of formation of water vapour
52*654 Cals. This agrees with G. Preuner's value. M. de K. Thompson calculated
from potential measurements 57'5 Cals. for water vapour, and 67*6 Cals. for liquid
water at 0° ; J. Thomsen gives 68-3 Cals., C. von Than, 68-43 Cals. ; A. Schuller
and V. Wartha, 68-25 Cals. for liquid water at 0°. The reported values for the heat
of formation of water are therefore very concordant, and G. N. Lewis considers
the best representative value to be 68'47 at 0°. J. E. Mills calculates for ice at
-273°, 71-4 Cals.
According to W. Nernst and H. von Wartenberg (1906), the free energy of the
reaction H2+J02=H20 when the two gases are at atmospheric pressure, and at
1000° K. is 90-6 Cals., showing that there is a large positive chemical affinity between
those gases at this temperature. G. N. Lewis and M. Randall calculate the free
energy of the reaction H2+402=H20 to be #=— 57410+0-94T log T+0-00165T2
— 0-00000037T3-f3-66T. This makes the free energy of formation of a gram-
molecule of gaseous water at 25°, i.e. 298° K., —54590 cals. If the vapour
pressure of water be 23*8 mm. at 25°, barometer 760 mm., the free energy =
—RT log (760/23*8) = — 2053 ; the free energy of formation of liquid water at
25° becomes— 56640 cals. The value calculated from the heat of dissociation of silver
oxide is —56530 cals. ; from the dissociation of mercuric oxide, —56650 cals. ;
and from the e.m.f . of the hydrogen : oxygen cell, —54567. The free energy of the
formation of a gram-molecule of ice is —56478 cals. ; and the free energy of fusion
H20soiid=H20iiquid is i^=— 1022-9-OT log T+54-230T, or -141*6 cals. at 25°
when the heat of fusion is 79-7 cals. per gram, or 1436 cals. per gram-molecule.
The decomposition of water. — Water is fairly stable, and even at the high
temperature of the oxyhydrogen flame — estimated to be over 3000° — the amount
decomposed is small although quite appreciable. Water is decomposed (i) by exposure
to very high temperatures ; (ii) by the electrolysis of the liquid ; (iii) by passing
a series of electric sparks or the silent discharge n through the vapour ; (iv) by the
formation of an electric arc under the liquid — as shown by W. Loeb ; (v) by the action
of the alkali metals at ordinary temperatures or of other metals at higher tempera-
tures whereby the oxygen of the water is fixed by the metal, and hydrogen gas escapes ;
and (vi) by exposing water to the action of fluorine at ordinary temperatures, or
of chlorine or bromine light or heat whereby the hydrogen of the water is fixed by the
halogen, and oxygen escapes. Again, (vii) W. Ramsay found that water is decom-
posed by the dissolution of radium salts in the water, whereby a stream of hydrogen
is continuously evolved ; and (viii) M. Kernbaum, that water is decomposed by
exposure to ultra-violet light : 2H20=H202+H2.
W. Ramsay and F. Soddy decomposed water acting as solvent with a radium
salt as solute, and W. H. Bragg noted what he called " a curious parallelism in
numbers " in that the number of molecules of water decomposed was almost exactly
equal to the number of ions which would have been produced in air by the emanation
employed. K. Bergwitz noted the decomposition of water by the a-rays of polonium
deposited on copper, and M. le Blanc showed that the results closely approximated
to the requirements of Faraday's law — the ionization and chemical action are of
WATER 491
the same statistical order. W. Duane and 0. Scheuer also found a close equivalence
between ionization and the amount of water decomposed. S. C. Lind showed that
the recoil atoms also play a part in producing ionization, and that if ionization by
a-particles be alone considered, the chemical effects produced appear greater than
corresponds with ionization.
In 1847, W. E. Grove described in a paper, On the decomposition of water into
its constituent gases, hy heat,^^ an experiment in which hydrogen and carbon dioxide
were heated in a tube by means of a wire heated white hot by means of an electric
current. Carbon monoxide and water were formed, C02+H2=C0+H20. If
carbon monoxide be similarly heated under precisely similar conditions, carl3fen
dioxide and hydrogen are formed C0+H20=C02+H2. Now, added W. R. Grove,
It appeared to me ultimately that the ignited platinum had no specific effect in producing
either composition or decomposition of the water, but it simply rendered the chemical
equilibrium unstable, and that the gases then restored themselves to a stable equilibrium
according to the circumstances in which they were placed, with regard to surrotmding
affinities ; that if the state of mixed hydrogen and oxygen were, at a certain temperature,
more stable than that of water, ignited platinum would decompose water as it does ammonia.
... It now appeared to me that it was possible to effect the decomposition of water by
ignited platinum : that supposing the atmosphere of steam in the immediate vicinity of
platinum were decomposed, or the affinities of its constituents loosened, if there were any
means of suddenly removing this atmosphere, I might get the mixed gases ; or secondly,
if quantity had any influence, that it might be possible to so
divide the mixed gases by a quantity of neutral ingredient as
to obtain them by subsequent separation (or as it were filtra-
tion) from the neutral substance.
W. R. Grove then related that when the incandescent
wire converted water into steam, some steam was always
decomposed, and a small bubble of mixed hydrogen and
oxygen gases was formed.
A glass tube with a piece of platinum wire sealed at the upper
end and filled with water, is placed in an inclined position, and
heated in its upper part by means of a spirit-lamp as shown in FiG. 25. — W. R. Grove's
Fig. 25. The platinum wire is heated to incandescence by a Experiment,
suitable batterj^. In a few moments the lamp is removed, and
the water is again allowed to fill the tube. A bubble of mixed gas remains in the tube,
and it can be examined. However long the heating be continued no further decomposition
occurs, for the equilibrium conditions between the water vapour and the products of decom-
position are established immediately the wire becomes incandescent. The experiment was
repeated with a similar result when electric sparks were passed from one wire to another
inside the tube ; and also when a platinum tube was heated with a blowpipe flame, and
no electrical heating employed.
W. R. Grove's work was undoubtedly the starting point of the modern theory
of dissociation, even though his explanations are couched in different terms. Ten
years after W. R. Grove, the subject was resumed by H. St. C. Deville,i3 who, in 1857,
repeated Grove's experiments on a larger scale, and introduced many other ingenious
modifications. He was able to collect a relatively large amount of the mixture of
oxygen and hydrogen by pouring one to three kilograms of molten platinum into
a vessel of water.
H. St. C. Deville also passed the vapour of water through a porous tube of earthenware,
placed in the axis of a wider glazed porcelain tube, and passed a current of carbon dioxide
in the annular space between the two tubes. The whole was heated in a furnace to about
1300°. The gases from the water dissociated in the inner tube, passed through the porous
walls at different rates, and were carried away before they had the chance of recombining
in the cooler parts of the tube. The carbon dioxide was removed by absorption in potash.
There are two important objections to this experiment- — first, it is doubtful if the temperature
was high enough to give an appreciable amount of decomposition, and second, there is a
possible reaction between the hydrogen and carbon dioxide. In A. W. Hofmann's experi-
ment (1890), a stream of water vapour is quickly passed through a tube containing a strip
of platinum foil heated to incandescence by an electric current, and the resulting hydrogen
and oxygen are prevented from recombining on cooling. The mixed gases are collected in
a cylinder over water.
492
INORGANIC AND THEORETICAL CHEMISTRY
The word dissociation was coined by H. St. C. Deville (1857) and was used
synonymously for decomposition, but three years later, he used the term to
characterize the temporary disjunction of the molecules of certain bodies into
their elements at elevated temperatures. He assumed that bodies possess, at
a temperature below their decomposing point, a certain tendency to decompose
which he called their dissociation tension.
E. Mallard and H. le Chatelier (1881) showed that the pressure developed during
the explosion of a mixture of hydrogen and oxygen is rather less than that calculated
o^ the assumption that combustion is complete. It was therefore inferred that a
small proportion remains uncombined. It will also be observed that the back-
reaction, 2H20=2H2+02, the un-burning, so to speak, sets a limit to the tempera-
ture attainable in a combustion dependent on the reaction symbolized : 2H2-fC)2
=2H20. W. Nemst and H. von Wartenberg were able to show that steam dis-
sociates by heat, and that the reaction is balanced by the recombination of the
products of decomposition to re-form water vapour : 2H20=H2+02. Only about
0-00003 per cent, is dissociated at 1000° and 3-98 per cent, at 2500°. The equilibrium
constant, K, for the reaction 2H20^2H2+02 is [H2]2[02]=K[H20]2, where the
bracketed terms denote concentrations — say partial pressures. If a gram- molecule
of water be decomposed into two gram-molecules of hydrogen, and one of oxygen,
and if x denotes the degree of dissociation, there will be present 1—x gram-molecules
of water, x of hydrogen, and ^x of oxygen. The total volume will be l—x-\-x-{-^x
=l-f Ja:, and if p denotes the total pressure, the partial pressure of water will be
[R20]=p{l—x)l{l-\-ix); of hydrogen, [}l2]=pxl(lflx); and of oxygen, [Og]
=ipxl{l-\-^x). By substitution, therefore, the equilibrium equation reduces to
K-
px^
(2-{-x)(l-x)^
If x is small, K=^px^, or log K=3 log a;— log 2, at atmospheric pressures when
y=l. In that case W. Nernst gives 3 logio a;=— 25050T-i+r75 logio T-f 0-00028T
+0-1. - - - .
W. Nernst and H. von Wartenberg ^^ further obtain
logZ=ir46-
25030
-2-381og
1000
■1-38 X 10-4(r-1000) -6-85 X 10-8(T2_ioo02)
The calculated degrees of dissociation at different pressures and temperatures are
indicated in Table XVII.
Table XVII. — Degree of Dissociation of Water at Different Temperatures.
rc
Degree of dissociation, x per cent.
3)»»10 atm.
p — 1 atm.
2>=»01 atm.
p= 0-01 atm.
727°
1227°
1727°
2227°
0-0000139
00103
0-273
1-98
0-00003
00221
0-588
3-98
0-0000646
0-0476
1-26
8-16
0-000139
0-103
2-80
16-6
The values agree satisfactorily with W. Nernst and H. von Wartenberg's observa-
tions below 2000° and with N. Bjerrum's above that temperature :
1124°
1288°
1984°
2369°
2488°
2656°
0-0073
0-034
0-77
4-3
8-6
• 11-1 per cent.
1. 1. Andreeif found that a mixture of hydrogen and oxygen, when exposed to the
ultra-violet light from a mercury lamp, unite to form water at a constant speed,
which is independent of the concentration of the reacting gases, but is almost
WATER 493
proportional to the intensity of the light. Water is also decomposed under the
influence of ultra-violet light. The same condition of equilibrium is attained whether
a mixture of hydrogen with half its volume of oxygen, or water- vapour be exposed
to the light. With an increase in the intensity of light, the equilibrium is displaced
in favour of dissociation.
The electrolysis of steam. — In 1858, A. Perrot is noticed that when an electrical
discharge is passed through steam, there is a kind of electrolysis, and J. J. Thomson
found that with short sparks, the gases which collect at the negative electrode contain
an excess of oxygen, and those at the positive, an excess of hydrogen — the total
gas collected corresponds with that obtained in the voltameter exactly as would be
obtained with a true electrolysis. It was therefore inferred that the current is
conveyed through water vapour in the same manner as through water. With long
sparks this state of things is reversed and the excess of hydrogen changes over to
the negative electrode, and the excess of oxygen to the positive electrode — the hydro-
gen collected is less than would correspond with a true electrolysis. Hence, adds
J. J. Thomson,
There is this remarkable difference between the electrolysis of steam and water, that
whereas in the case of water, the hydrogen always comes off at the negative, and the oxygen
at the positive electrode ; in the case of steam, the hydrogen and oxygen come off sometimes
at one terminal, sometimes at the other, according to the nature of the spark.
According to D. L. Chapman and F. A. Lidbury, the electrolysis of water vapour
does not take place in such a manner that the hydrogen appears at one pole and
oxygen at the other, but when water vapour is decomposed bj^ electric sparks, the
hydrogen separates at both electrodes, and it has a tendency to accumulate in the
neighbourhood of the two electrodes, while the oxygen is driven towards the middle
of the spark gap. The nature of the gases collected at the respective electrodes is
determined by the velocity of the current of steam, as well as by the nature of the
sparks. Hydrogen can diffuse against the stream of water vapour much faster than
oxygen. When electric sparks are passed through steam (or other compound gases),
the distribution of the products of decomposition is not always the same as when
the corresponding liquid is electrolyzed. The decomposition may occur in the entire
path of the spark as well as in the neighbourhood of the electrodes. The distribution
of the decomposition products about the two electrodes is determined by the position
of the tube bringing the current of steam to the electrodes and the relative rates of
diffusion of the constituent gases.
Water is jtar excellence a solvent. Aqueous solutions of certain salts can dissolve
some substances not soluble in pure water — e.g. salts of benzoic, salicylic, and
benzenesulphonic acids on certain alkaloids, fats, alcohols, and carbohydrates. The
phenomenon is termed hydrotropism by C. Neuberg (191 6). ^^ The metals act as
reducing agents towards water forming an oxide — usually basic — and liberating
hydrogen ; or reciprocally, water acts as an oxidizing agent on the metals — e.g.
it has been shown that with sodium, 2Na-f 2H20=2NaOH4-H2 ; and with
iron, 3Fe-}-4H20=Fe304-|-4H2. The behaviour of the metals towards water
was used as an important criterion in L. J. Thenard's and H. V. Kegnault's systems
of classifying the metals. i^ They first attempted a dichotomous separation into
metals whose oxides can be decomposed by heat alone, and those whose oxides
cannot be so decomposed. The metals were then arranged in five sub-groups :
(1) metals which decompose cold water — e.g. the metals of the alkalies and of the alka-
line earths ; (2) metals which decompose hot water between 50° and 100° — e.g.
beryllium, magnesium, etc. and a number of the rare earth metals, etc. ; (3) metals
which decompose water at a red heat — e.g. zinc, cadmium, tin, iron, nickel, cobalt,
chromium, osmium, etc. ; (4) metals which decompose water only at a white heat —
e.g. copper, lead, etc. ; (5) metals which do not decompose water at any temperature —
e.g. mercury, silver, gold, and members of the platinum family excepting osmium.
The halogens act as oxidizing agents on water liberating oxygen ; or reciprocally,
494 INORGANIC AND THEORETICAL CHEMISTRY
water acts as a reducing agent on the halogens — e.g. with chlorine, 2CI2+2H2O
=4HCl+02. With sulphur and phosphorus the hydrogen of the water forms a
hydride and the oxygen forms an oxide which may unite with water furnishing an
acid : e.g. with sulphur over 100°, 2H204-3S=2H2S+S02 ; and with phosphorus
at 250°, 3H20+2P=PH3+H3P03 — if air be present the action occurs at a lower
temperature. Some of the non-metals act like the metals and form an oxide — usually
acidic — and liberate hydrogen — e.g. carbon and boron at a red heat. The reaction
with carbon is symbolized: C+H20=CO+H2 and C+2H20=C02+2H2 ;
there is also a reversible side reaction, C02+H2^H20+CO. Some oxides
react with water liberating hydrogen — for instance, in the reaction last symbolized
carbon monoxide is decomposed by water vapour forming carbon dioxide and hydro-
gen ; similarly, manganous oxide forms the brown oxide, Mn304, and hydrogen,
3MnO+H20=Mn304-f H2 ; uranium oxide, UO2, gives the green oxide,
U3O8, and hydrogen ; chromous oxide in acid solution is oxidized by water and
hydrogen is slowly evolved ; potassium cobaltocyanide, K4CoCy6, on oxidization by
water to the cobalticyanide, K3CoCy6, gives ofE hydrogen ^^ ; when an alkaline
solution of molybdenous chloride, M03CI6, is warmed, a black precipitate of
Mo(0H)3 is produced, and hydrogen is evolved: 2Mo3Cl6+18H20=6Mo(OH)3
4-12HC1+3H2. Neither iodine nor aluminium alone acts chemically on water,
but remarkably enough, conjointly these elements attack water with the evolution
of hydrogen. J. H. Gladstone and A. Tribe i^ consider that there is first formed a
little aluminium iodide, AII3, which is immediately decomposed by water whereby
aluminium hydroxide, A1(0H)3, and hydriodic acid, HI, are produced : AII3
-}-3H20=Al(OH)3+3HI ; the hydriodic acid immediately attacks the metal
re-forming aluminium iodide with the evolution of hydrogen : 2Al+GHI=2All3
+3H2. Water also reacts with many metal dioxides forming hydroxides. For
example, with the oxides of the alkalies and alkaline earths a base is formed — thus,
with calcium oxide the reaction is symbolized, CaO+H20=Oa(OH)2. With
the non-metaUic oxides an acid is formed — thus, the products with sulphur
trioxide, SO3, and water, are sulphuric acid, H2SO4, in symbols : S03+H20=H2S04.
Water reacts with many metal phosphides, siUcides, or carbides, respectively
forming hydrogen phosphide, silicide, or carbide (hydrocarbon), and the metal
oxides or hydroxides : e.g. calcium carbide, CaC2, gives acetylene, C2H2, and calcium
hydroxide : CaC2+2H20=C2H2+Ca(0H)2 ; similar remarks apply to the
carbides of the alkali and alkaline earth metals ; aluminium or beryllium carbides
give methane, CH4 ; and other carbides give various mixtures of hydrogen and
hydrocarbons. The sulphides and selenides of boron, silicon, aluminium, and
magnesium are decomposed by water furnishing hydrogen sulphide or selenide,
and the metal hydroxide, or, in the case of the non-metals, an acid and the metal
hydroxide. The nitrides of the metals usually give ammonia with cold or hot water ;
the metal hydrides give hydrogen under similar conditions— the metal hydroxide
is formed simultaneously. The organo-metaUic compounds of the more oxidizable
metals give the corresponding hydrocarbon and the metal hydroxide — e.g. zinc
methide, Zn(CH3)2, gives methane, CH4, and zinc hydroxide, Zn(0H)2.
Many halogen compounds give the haloid acid and either an oxyhaloid or
hydroxide of the other element, e.g. with phosphorus pentachloride, PCls, hydro-
chloric acid, HCl, and phosphoric acid, H3PO4, are formed : PCl5-l-4H20=5HCl
-{-H3P04 ; with antimony chloride, antimony oxychloride, SbOCl3, and hydrogen
chloride, HCl, are formed by a reversible reaction : SbCl5+H20=SbOCl3
+2HC1. The esters— studied by M. Berthelot and L. P. de St. Giles 20 in their
Recherches sur les affinites (1861-3) — are broken down by water — preferably in dilute
acid solution ; e.g. ethyl acetate, CH3COOC2H5 forms acetic acid, CH3CO.OH, and
alcohol, C2H5OH. The reaction gradually slows down, and finally comes to a stand-
still, when a certain proportion of the four components of the reaction are present :
CHgCOIOCaHs+HjOH = CH3CO.OH + CgHgOH
Ethyl acetate. Water. Acetic acid. Ethyl alcohol.
WATER 495
The distribution of these four components when the system is in equilibrium depends
on their concentration, and on the temperature. The work on this reaction is
classical, and it played an important role in the evolution of the law of mass action.
Reactions like the so-called bydrolysis of potassium cyanide, KCy, symbolized,
KCy-f HgO^KOH+HCy ; and of ammonium chloride, NH4Cl-fH20
=NH40H4-HC1, are similar in character, and the ionic hypothesis attempts to
describe the mechanism in still more detail. Water also forms a series of hydrated
compounds with the elements— e.^. Br2-flOH20=Br2(H20)io, or Br2.10H20 —
or with compounds— e.^r. FeS04.6H20+H20=FeS04.7H20 ; and also Na2S04
-|-10H20=Na2S04.10H20. The relations of the water in these compounds to the
rest of the molecule has been much discussed.
The different behaviour of one of the two hydrogen atoms in water towards
sodium, ethyl iodide, C2H5I, and phosphorus trichloride, PCI3, led W. Ganswindt
(1891)21 to assume that water is hydrogen hydroxide, H— (OH). The argument,
however, is rather weak, because it is possible that there is an equal proba-
bility of, say, sodium displacing either of the hydrogen atoms in water, H.O.H,
but that as soon as one hydrogen atom has been replaced by sodium, the molecule
with its one hydrogen atom becomes more resistant. J. W. Briihl (1896) assumed
that water contains quadrivalent unsaturated oxygen because water exhibits, beyond
most other substances, an unsaturated character — as exemplified by its faculty of
forming hydrated and crystalline compounds, and its great solvent and ionizing
powers. Hence, J. W. Briihl represents the molecule of water by the formula
HOH.
The adsorption of water by solids. — Probably many substances adsorb water
vapoux, so that after exposure to air, they are covered with a film of moisture which
they retain so very tenaciously that it can be removed only by heating to a tempera-
ture short of dull redness in vacuo. R. Bunsen estimated that 2 '11 sq. metres
of glass surface, dried at 20°, lost 22*3 mgrm. of water when heated to 500°. The
adsorption of water vapour by glass surfaces has been studied by R. Bunsen,22
E. Warburg and T. Ihmori, C. J. Parks, etc. ; quartz by A. von Dobeneck,
L. J. Briggs, etc. ; and charcoal by H. W. Foote. The penetration of adsorbed
water into insulating materials has been studied by C. J. Rottmann. T. Ihmori also
investigated the adsorption of water vapour by platinum, shellac varnish, brass, etc.
It has long been known that old soda-glass tubing which has stood in the laboratory
some time, becomes very rough when suddenly heated owing to the development
of innumerable spits. This is probably due to adsorbed water.
E. Warburg and T. Ihmori divide the water film condensed on the surface of
glass into a temporary portion which disappears when the vapour pressure is reduced
to zero, and a permanent portion which remains. According to L. J. Briggs, the
adsorption of water vapour by quartz is less than with amorphous silica. At 30°,
the permanent film condensed on quartz from an atmosphere within one per cent,
of saturation, corresponds with a film 2"66xlO~6 cm. thick on the assumption that
the film is uniform and has unit specific gravity. C. J. Parks found for glass wool
in a saturated atmosphere at 15° a thickness 13*3x10-6 cm. H. ^Y. Foote and
B. Saxton found that the water adsorbed by lampblack is essentially the same as
other water, only it does not freeze in the capillary pores until a low temperature,
about —35°, is attained.
Pouillet effect.— In 1822, G. S. N. Pouillet 23 showed that porous substances-
paper, wool, etc. — and fine powders — glass, charcoal, alumina, etc. — become heated
when they are wetted with water — a liquid which exerts no solvent or chemical
action on the solid. The phenomenon is now known as the Pouillet effect. The
rise of temperature amounts to about 1°, and by suitably varying the conditions a
rise of nearly 30° has been recorded. One gram of powdered charcoal; according
to P. Chappius, evolves 7' 425 cals. when wetted with water, and 24*36 cals. when
wetted with carbon disulphide, while a gram of powdered alumina evolves 2' 747
cals. when wetted with water. For the same liquid and the same powder, the
496 INORGANIC AND THEORETICAL CHEMISTRY
quantity of heat evolved is proportional to the mass of the powder. According to
G. Schwalbe, 10 grms. of washed sand gives 0*3 cal., when wetted with 0"5 grm. of
water, andO'8 cal. when wetted with 2 grms. of water, but no further rise of tempera-
ture occurred when more water was used ; with silicic acid, 189 cals. were developed
with 5 grms. of water, and 6" 16 cals. with 20 grms. of water. T. Tate referred the cause
of the Pouillet effect t6 chemical action, but this hypothesis was considered to be out
of the question ; G. S. N. Pouillet himself referred the effect to capillary action ;
C. G. Jungk showed that the effect is possibly due to the exercise of a pressure at the
surface of the powder and liquid. C. Cantoni (1866) and L. Meslens (1874) assumed
that the water passes into a different state of aggregation when it wets a powder —
either into the solid state or some intermediate state between solid and liquid.
T. Martini holds that just as a gas becomes a liquid when dissolved by a liquid, so
does a liquid become solid when dissolved by a solid ; and accordingly, the Pouillet
effect is due to the latent heat of solidification. This hypothesis was discountenanced
whenM. Bellati and L. Finazzi (1902) showed that the specific heat of the water was
not diminished by adsorption, as would have been the case if the water was solidified.
They also stated that the grain-size of the powder had no influence on the result, but
this statement proved to be erroneous ; for C. J. Parks (1902) showed that
when powdered and dry silica, sand, or glass are wetted with water at the
same temperature the heat evolved is proportional to the exposed area of the
solid, and is nearly equal to 0*00105 cal. per sq. cm. when the temperature is 7°.
G. Schwalbe (1905) then demonstrated that if the temperature is below 4°, there is a
negative Pouillet effect, for the water is cooled and not heated ; and at 4°, there is
neither heating nor cooling. This is in agreement with Lord Kelvin's proof that the
temperature change dT due to the change of pressure djp on a liquid whose coefficient
of thermal expansion is a, specific gravity D, absolute temperature T, and specific
heat at constant pressure Cp, is
Since the coefficient of expansion of water changes sign at 4°, there should be a
change of sign in the Pouillet effect at the same temperature. Lord Kelvin computes
there is a cooling of 0*00026° when a cubic centimetre of water at 0° is subjected to
a pressure of 10 atm., and at 10° a rise of temperature of 0*0040°. It is also possible
to calculate the compression of the w^ater adsorbed by the powder from the Pouillet
effect. The formation of ice within fissures and caves whose mean temperature
is not below the freezing point of water, has also been explained by assuming that
water below 4° is cooled instead of heated in percolating through the sandy walls of
the cave.
Refebences.
1 R Bunsen, Onsometrische Methoden, Braunschweig, 1877; H. Kolbe, Liehig\s Ann., 119.
176 18(51 ; A. W. Hoffman, Ber., 3. 658, 1870; R. Bottger, Jouryi. prakt. Chem'., (1), 85. 390,
1862 ; A. R. Leeds, Chem. News, 49. 237, 1884 ; M. Berthelot, Compt. Reyid., 130. 1002, 1900.
* M Bertbelot, Ann. Chim. Phys., (5), 17. 142, 1879 ; P. P. Deherain and L. Maquenne,
Compt. Rend., 93. 895, 963, 1021, 1881 ; P. J. Kirby, Phil. Mag., (6), 7. 223, 1904 ; W. G. Mixtcr,
Amer. Journ. Science, (4), 4. 51, 1898.
» B Davis and C. W. Edwards, Journ. Soc. Chem. Ind., 24. 206, 1905; C. Pickel, Zeit.
anorg. Chem., 38. 307, 1904 ; W. P. Jorissen and W. E. Ringer, Ber., 39. 2093, 1900 ; S. C.
Lind Jmirn. Amer. Chem. Soc, 41. 631, 651, 1919; Jmirn. Phys. Chem., 16. 664, 1912; Trans.
Amer Electrochem. -Soc, 24. 339, 1913; Radium, 11. 108, 1914; Zeit. phjs. Chem., 84. 759, 1913;
A. T. Cameron and W. Ramsay, Jmirn. Chem. Soc, 93. 966, 1909; 91. 1260, 1907; W. Ramsay,
ib 91. 931, 1907; W. Duane and O. Scheuer, Radium, 10. 33, 1913; O. Seheuer, Compt. Rend.,
159. 423, 1914; W. Duane and A. Laborde, ih., 150. 1421, 1910; E. Wurtzel, ib., 157. 929,
1913; Jaarn. Russian Phys. Chem. Soc, 47. 210, 493, 494, 1915.
4 T. de Saussure, Journ. prakt. Chem., (1), 14, 152, 1838; C. F. Schonbein, ib., (1), 89, 344,
1863; B. Hooke, Nicholson's Journ., 8. 228, 1803 ; T. de Saussure, Gilbert's Ann., 47. 103, 1815 ;
H. b! Baker, Proc Chem. Soc, 18. 40, 1902 ; J. B. Blot, Gehlen's Journ., 5. 95, 1805 ; F. de la
Roche, Schweigger's Journ., 1. 172, 1811 ; A. F. E. Degen, Pogg. Ann., 38. 454, 1830; H. N.
Warren, Chtm. News, 67. 195, 1893.
WATER 497
6 V. Meyer and W. Raum, Ber., 28. 2904, 1896 ; H. Helier, Ann. Chim. Phys., (7), 10. 621,
1897 ; W. A. Bone, Joitrn. Chew. Soc, 81. 535, 1902 ; 86. 694, 190^1 ; A. Gautier and H. HeUer,
Compt. Rend.., 122. 566, 1896 ; M. Bodenstein, Zeit. phys. Chem., 29. 147, 1899 ; P. Duhem, ib.,
29. 711, 1899 ; M. Berthelot, Compt. Rend., 124. 1275, 1897.
8 H. Davy, Phil. Trans., 96. 7, 1816 ; 97. 45, 1817 ; E. Mallard and H. le Chatelier, Compt.
Rend., 91. 825, 1880 ; Bull. Soc. Chim., (2), 39. 2, 1883 ; Lord Rayleigh, Proc. Roy. Inst., 7.
386, 1875 ; Nature, 11. 454, 1875 ; F. Freyer and V. Meyer, Ber., 25. 622, 1892; A. MitscherUch,
ib., 26. 160, 399, 1893 ; V. Meyer and A. Munch, ib., 26. 2421, 1893 ; V. Meyer, G. Krause,
and P. Askensay, ib., 25. 622, 1892 ; 26. 429, 1893 ; Liebig's Ann., 264. 85, 1891 ; 269. 49, 1892 ;
Zeit. phys. Chem., 11. 28, 1893 ; M. Bodenstein, ib., 29. 665, 1899 ; K. G. Falk, Journ. Amer.
Chem. Soc, 28. 1527, 1906 ; 29. 1856, 1907 ; P. Emich, Monatsh., 18. 6, 1897 ; 19. 299, 1898 ;
21. 1061, 1900; R. W. Bunsen, Qasometrische Methoden, Braunschweig, 1857 ; A. Gautier and
H. Helier, Compt. Rend., 122. 566, 1896 ; H. Helier, Ann. Chim. Phys., (7), 10. 521, 1897 ;
H. B. Dixon and H. F. Coward, Journ. Chem. Soc, 95. 514, 1909 ; H. B. Dixon, ib., 97. 661.
1910 ; H. F. Coward, C. Cooper, and C. H. Warburton, ib., 101. 2278, 1912 ; H. F. Coward,
C. Cooper, and J. Jacobs, ib., 105. 1069, 1914 ; H. F. Coward and F. Brinsley, ib., 105. 1859,
1914 ; L. Meyer and K. Seubert, ib., 45. 581, 1884 ; H. B. Dixon, Phil. Trans., 175. 634,
1884 ; A. de Hemptinne, Bull. Acad. Belg., 761, 1902 ; F. Fischer and M. Wolf, Ber., 44. 2956,
1911 ; T. von Grotthus, Gilbert's Ann., 33. 212, 1809; 58. 345, 1818; 69. 241, 1821.
7 H. B. Dixon, Phil. Trans., 184. 97, 1893 ; Ber., 38. 2419, 1905 ; M. Berthelot, Compt. Rend.,
93. 18, 1881 ; E. Mallard and H. le Chatelier, Recherches experimentales et theoriques sur la
combustion des melanges gazeux explosives, Paris, 1883.
8 H. Davy, Phil. Trans., 97. 45, 77, 1817 ; E. Davy, ib., 100. 108, 1820 ; M. Faraday, ib.,
114. 55, 1834 ; W. A. Bone and R. V. Wheeler, ib., 206. 1, 1906 ; P. Erman, Abh. Akad. Berlin,
368, 1919 ; T. von Grotthus, Gilbert's Ann., 33. 212, 1809 ; 58. 345, 1818 ; 69. 241, 1821 ; Liebig's
Ann., 1. 29, 1832 ; 14. 10, 1835 ; K. Karmarsch, ib., 75. 80, 1850 ; G. F. F. Chladni, ib., 61. 346,
1847 ; 75. 98, 1850 ; J. W. Dobereiner, ih., 74. 269, 1850 ; Ueber neu entdeckte hochst merkumrdige
Eigenschaften des Platins, Jena, 1823 ; Kastner's Arch., 2. 225, 1824 ; Journ. prakt. Chem., (1),
17. 158, 1839 ; (1) 1. 114, 1834 ; Schweigger's Journ., 34. 91, 1822 ; 38. 321, 1823 ; 39. 4, 142,
1823; 42. 60, 1824; 47. 133, 1826; 63. 465, 1833; A. Pleischl, Journ. prakt. Chem., (1), 39.
142, 201, 351, 1846 ; C. G. Gmelin, ib., 38. 515,1846 ; F. Pfaff, i6., 40. 1, 1847 ; J. S. C. Schweigger,
ib., 39. 223, 1846 ; 40. 10,237, 1847 ; R. Bottger, ib., 63. 370, 1854 ; 68. 390, 1856 ; P. L. Dulong
and L. J. Thenard, Ann. Chim. Phys., (2), 23. 440, 1823 ; (2), 24. 380, 1823 ; A. de la Rive and
F. Marcet, ib., (2), 39. 328, 1828 ; J. Babinet, ib., (2), 37. 183, 1828 ; A. de la Rive, Pogg. Ann.,
46. 489, 492, 1839 ; 54. 386, 397, 1841 ; L. Meyer, ib., 104. 189, 1858 ; G. H. Quincke, ib., 150.
118, 1877 ; M. Berthelot, Compt. Rend., 94. 1377, 1882 ; 125. 271, 1897 ; Ann. Chim. Phys., (5),
30. 519, 1883 ; K. Stratmgh, Repert. Pharm., 21. 410, 1865 ; P. van Dijk, ib., 21. 235, 1825 ;
E. Turner, Edin. Phil. Journ., 11. 90, 311, 1824; 12. 311, 1825 ; W. Henry, Phil. Trans., 114.
266, 1824 ; Ann. Phil., 21. 364, 1824 ; 25. 416, 1825 ; Phil. Mag., 6. 364, 1836 ; W. C. Henry, ib.,
6. 354, 1835 ; 9. 324, 1836 ; T. Graham, Quart, Journ. Science, 6. 354, 1829 ; W. Artus, Journ.
prakt. Chem., (1), 6. 176, 1835 ; F. Mohr, Liebig's Ann., 23. 228, 1856 ; A. Fusinieri, Giorn. di
Fisica, 8. 259, 1825 ; J. J. Thomson, Application of Dynamics to Physics and Chemistry, London,
206, 236, 1888 ; E. D. CampbeU, Amer. Chem. Journ., 17. 496, 1895 ; W. Hempel, Ber., 12. 1006,
1879 ; E. Harbeck and G. Lunge, Zeit. anorg. Chem., 41. 764, 1904 ; E. Richardt, ib., 38. 65, 1904 ;
E. Jager, Journ. Gasbeleuchtung, 41. 764, 1898 ; W. French, Chem. News, 81. 292, 1900 ;
M. Bodenstein, Zeit. phys. Chem., 29. 665, 1899 ; 46. 725, 1903 ; A. W. Rowe, ib., 59. 41, 1907
C. Ernst, ib., 37. 448, 1901 ; C. Engler and L. Wohler, Zeit. anorg. Chem., 29. 1, 1901 ; L. Wohler,
Ber., 36. 3475, 1903 ; V. Meyer and W. Raum, ib., 28. 2804, 1895 ; P. Sabatier, La catalyse, Paris,
1913 ; H. Landolt, Pogg. Ann., 99. 411, 1856 ; W. A. Bone, Journ. Chem. Soc, 81. 535, 1902 ;
85. 694, 1904; Ber., 46. 14, 1913; Proc Roy. Inst., 19. 82, 1908; Phil. Trans., 215. A, 298,
1915 ; W. D. Bancroft, Journ. Phys. Chem., 21. 644, 573, 1917.
" I. Langmuir, Journ. Amer. Chem. Soc, 37. 1164, 1915; M. Bodenstein and F. Ohlmer,
ZeU. phys. Chem., 53. 166, 1905 ; M. Bodenstein and C. G. Fink, ib., 60. 46, 1907.
10 M. Berthelot and C. Matignon, Ann. Chim. Phys., (6), 30. 563, 1893 ; Compt. Rend., 115.
347, 1892; J. Thomsen, Ber., 5. 769, 1872; 6. 2553, 1873; 15. 2998, 1882; Thermochemische
Untersuchungen, Leipzig, 1. 52, 1882 ; A. Schiiller and V. Wartha, Wied. Ann., 2. 52, 1877 ;
C. von Than, i6., 14. 422, 1881 ; W. Nernst and H. von Wartenberg, Zeit. phys. Chem., 56. 534,
643, 1906 ; F. Haber and L. Bruner, Zeit. Elektrochem., 12. 78, 1906 ; W. G. Mixter, Amer.
Journ. Science, (4), 16. 214, 1903 ; M. de K. Thompson, Journ. Atner. Chem. Soc, 28. 731, 1906 ;
E. Bose, Zeit. phys. Chem., 38. 1, 1901 ; N. T, M. Wilsmore, ib., 35. 291, 1901 ; G. W. Preuner,
ib., 42. 50, 1903 ; G. N. Lewis, ib., 55. 463, 1906 ; Journ. Amer. Chem. Soc, 28. 1390, 1906 ;
C. N. Lewis and M. Randall, ib., 36. 1967, 1914 ; F. Haber, Thermodynamik technischer Gasre-
aktionen, Munchen, 1905 ; Zeit. anorg. Chem., 51. 250, 1906 ; J. E. Mills, Chem. News, 105. 18,
1912.
11 P. P. Deherain and L. Maquenne, Compt. Rend., 93. 895, 963, 1031, 1881 ; P. and A. Thenard,
ib., 76. 1508, 1873 ; A. Debieme, ib., 148. 703, 1909 ; Ann. Chim. Phys., (9), 2. 97, 1914 ;
M. Kembaum, Bull. Acad. Cracow, 683, 1911 ; W. Loeb, Ber., 34. 917, 1901 ; W. Ramsay and
F. Soddy, Proc. Roy. Soc, 72. 204, 1903 ; W. H. Bragg, Phil. Mag., (6), 13. 356, 1907 ; K. Bergwitz,
Phys. Zeit., 11. 273, 1910; M. le Blanc, Zeit. phys. Chem., 85. 611, 1913; W. Ramsay, Journ.
Chem. Soc, 91. 931, 1907 ; A. T. Cameron and W. Ramsay, ib., 91. 1266, 1907 ; 93. 966, 1908 ;
VOL. I. 2 K
498 INORGANIC AND THEORETICAL CHEMISTRY
W. I>uane and 0. Scheuer, Radium, 10. 33, 1913 ; O. Scheuer, Compt. Rend., 159. 423, 1914;
E. Wonrtzel, ih., 157. 929, 1913 ; Journ. Russian Phys. Chem. Soc, 47. 210, 493, 494, 1915 ;
S. C. Lind, Journ. Amer. Chem. Soc., 41. 631, 551, 1919.
" W. R. Grove, Phil. Trans., 137. 1, 1847 ; Phil. Mag., (3), 31. 20, 91, 96, 1847 ; G. Wilson
ib., (3), 31. 177, 1847.
" H. St. C. Deville, Co7npt. Rend., 45. 857, 1857 ; 56. 195, 729, 1864 ; 59. 873, 1865 ; 60.
317, 1865 ; Repert. Chim. Pure, 2. 37, 1860 ; Le<;ons sur la dissociation, Paris, 1866 ; H. Debray,
Compt. Rend., 64. 603, 1867 ; E. Mallard and H. le Chatelier, ib., 93. 1014, 1076, 1881 ;
R. Bunsen, Oasometrische Methoden, Braunschweig, 1877 ; A. W. Holmann, Ber., 23. 3310,
1890.
1* H. von Wartenberg, ih., 56. 513, 1906 ; Verh. deut. phys. Oes., 8. 97, 1906 ; W. Nernst and
H. von Wartenberg, Zeit. phys. Chem., 56. 534, 1906 ; N. Bjerrum, ib., 79. 513, 1912 ; E. Bose
ib., 34. 701, 1900; 38. 1, 1901 ; G. N. Lewis, ib., 55. 465, 1906; L. Lowenstein, ib., 54. 715
1906 ; Beitrdge zur Messung von Dissociationen bei hohen Temperaturen, Gottingen, 1905 ;
F. Haber and L. Bruner, Zeit. Elektrochem., 10. 697, 1904 ; 12. 79, 1906 ; F. Haber and F. Fleisch
mann, Zeit. anorg. Chem., 51. 245, 1906 ; F. Haber and G. W. A. Foster, ib., 51. 289, 1906 ,
F. Fleischmann, UrUersuchungen iiber die Knallgaskette bei hoherer Temper atur unter Benutzung
von Glas und Porzellan als Elektrolyt, Karlsruhe, 1907 ; F. Haber, Ann. Physik, (4), 26. 942, 1908
W. H. Patterson, Phil. Mag., (6), 13. 181, 1907 ; A Holt, ib., (6), 13. 630, 1907 ; I. I. Andreeff,
Journ. Russian Phys. Chem. Soc., 43. 1342, 1911 ; I. Langmuir, Journ. Amer. Chem. Soc, 28. 1357
1906 ; W. Nernst, Experimental and Theoretical Applications of Thermodynamics to Chemistry
London, 1907 ; F. PolHtzer, Die Berechnung chemischer Affinitdten nuch dem Nernstschen Wdrme
theorem, Stuttgart, 1912 ; I. W. Cederberg, Die thermodynamische Berechnung chemischer Affini-
tdten, Berlin, 1916 ; F. Haber, Thermodynamik technischer Gasreaktionen, Miinchen, 1905.
15 D. L. Chapman and F. A, Lidbury, Journ. Chem. Soc, 81. 130, 1902 ; A. Perrot, Compt.
Rend., 46. 180, 1858 ; 47. 35S, 1858 ; Ann. Chim. Phys., (3), 61. 161, 1861 ; J. J. Thomson,
Recent Researches in Electricity and Magnetism, Cambridge, 559, 1893,
" C. Neuberg, Biochem. Zeit., 76. 107, 1916.
1' L. J. Thenard, Traite de chimie dimerdaire, Paris, 1816 ; H. V. Regnault, Cours ilementaire
de chimin, Paris, 1840.
" M. Berthelot, Compt. Rend., 127. 24, 1898 ; A. Deschamps, ib., 67. 330, 1868 ; R. Peters,
Zeit. phys. Chem., 26. 195, 1898 ; Pharm. Centrh., 39. 69'o, 1898 ; H. V. Regnault, Ann. Chim.
Phys., (3), 62. 349, 1836 ; W. Muthmann and W. Nagel, Ber., 31. 2009, 1898.
" J. H. Gladstone and A. Tribe, Chem. Neus, 42. 2, 1880.
2» M. Berthelot and L. P. de St. Giles, Compt. Rend., 53. 474, 1861 ; Ann. Chim. Phys., (3),
65. 385, 1862 ; (3), 66. 5, 68, 1863 ; (3), 68. 225, 1863 ; W. Osticald's Klassiker, 173, 1910.
" A. Ganswindt, Pharm. Centrh., 32. 291, 1891 ; J. W. Bruhl, Ber., 28. 2866, 1895 ; A. Wurtz,
La theorie atomique, Paris, 243, 1893.
22 R. Bunsen, Wied. Ann., 20. 545, 1883 ; 24. 321, 1885 ; T. Ihmori, ib., 31. 1006, 1887 ;
E. Warburg and T. Ihmori, ib., 27. 481, 1886 ; C. J. Parks, Phil. Mag., (6), 5. 517, 1903 ; A. von
Dobeneck, Forsch. Qebiete Agrik. Physik, 15. 163, 1892 ; L. J. Briggs, Journ. Phys. Chem., 9.
617, 1905 ; H. W. Foote and B. Saxton, Journ. Amer. Chem. Soc, 38. 588, 1916 ; 39. 627, 1917 ;
C. Rottmann, Journ. Franklin Inst., 188. 409, 1919 ; E. Priwoznik, Zeit. anorg. Chem., 9. 289,
1895 ; A. von Schrotter, Sitzber. Akad. Wien, 63. 2, 1871 ; B. Moore and J. W. Mellor, Trans.
Cer. Soc, 7. 1, 1908.
23 G. S. N. Pouillet, Ann. Chim. Phys., (2), 20. 141, 1822 ; L. Meslens, ib., (5), 3. 522, 1874 ;
P. Chappius, Wied. Ann., 19. 21, 1883; F. Meissner, ib., 29. 114, 1886; E. Wiedemann and
C. LiJdeking, ib., 25. 145, 1885 ; G. Schwalbe, Ann. Physik, (4), 16. 42, 1905 ; C. G. Jungk,
Pogg. Ann., 125. 292, 1865; G. Rose, ib., 73. 1, 1848; A. Mousson, ib., 105. 161, 1858;
L. Dufour, ib., 114. 530, 1861 ; F. Meissner, Ueber die beim Benetzen pulverfor^niger Korper
auftretende Wdrmetonung, Strassburg, 1886 ; G. Gore, Phil. Mag., (5), 37. 306, 1894 ; (5), 44.
205, 1897 ; Lord Kelvin (W. Thomson), ib., (4), 15. 540, 1858 ; T. Martini, ib., (5), 48. 329, 1899 ;
(5), 50. 618, 1900 ; Atti 1st. Veneto, 59. ii, 615, 1900 ; Nuovo Cimenio, (4), 7. 396, 1898 ;
G. Ercolini, ib., (4), 9. 110, 1899 ; S. Lussana, ib., (4), 2. 233, 1895 ; M. Bellati and L. Fmazzi, Atti
1st. Veneto, 61. 503, 1902 ; Phil. Mag., (6), 4. 240, 1902 ; A. Kirschmann, Phys. Zeit., 4. 797,
1903 ; W. Spring, Bull. Soc Geol. Belgique, 17. 13, 1903 ; G. J. Parks, Proc Phys. Soc, 18. 253,
1902 ; Phil. Mag., (6), 4. 240, 1902 ; T. Tate, ib., (4), 20. 508, 1860 ; S. Lagergren, Beh. Vet.
Akad. Handl., (2), 24. 14, 1899 ; M. Ventzke, Dingier' s Journ., 129. 144, 1853 ; C. Cantoni, Rend.
Real. Inst. Lombardo, 8. 135, 1866 ; P. G. Tait, Proc. Roy. Soc Edin., 11. 51, 217, 1881 ; D. H.
Marshall, C. M. Smith, and R. T. Omond, ih., 11. 626, 809, 1881 ; G. Creelman and J. Crocket,
ib., 13. 311, 1885 ; E. H. Amagat, Compt. Rend., 116. 946, 1893 ; C. E. Linebarger, Phys. Rev.,
13. 48, 1901 ; G. F. Fitzgerald, Nature, 49. 293, 316, 1894 ; J. P. Joule, Phil. Trans., 149. 135,
1859.
§ 10. Hydrates and Hydrated Salts
The term hydrate is not used very definitely in chemistry. It is sometimes
loosely employed in contrast with anhydrous. In addition, there are at least three
WATER 499
difierent meanings to the word : (1) Hydrated colloids or colloidal water. — Silica,
and the hydroxides of many of the metals — iron, chromium, aluminium, etc. — pass
out of solution in a more or less gelatinous or colloidal condition associated with an
indefinite amount of water. When dried, the water usually passes ofE continuously
with rise of temperature without any signs of the existence of definite chemical
compounds. (2) Hydrated salts or water of crystallization. — The very definite amount
of water which is contained in many salts which crystallize from aqueous solutions —
e.g. sodium carbonate, Na2CO3,10H2O, etc. — is called water of crystallization, and
several other synonyms have been employed. It is often stated that the molecules
of water in compounds containing water of crystallization belong to the molecular
structure or else exist in them entirely among the other molecules, and belong only
to the crystalline structure ; on the other hand, the molecule of water of constitution
is not supposed to exist as such but to be formed when the mineral is decomposed
owing to the union of the contained hydrogen with oxygen or hydroxyl groups
contained in the molecule. (3) Hydroxides — acids, alkalies, alcohols, etc. — The water
of hydration appears to be an integral part of the molecule, and it cannot usually
be removed without changing the character of the substance. Examples are indicated
above, to these are sometimes added the alcohols which are related to water and the
ethers as previously indicated by graphic formulae.
Both hydrate and hydroxide contain the elements of water which can usually
be recovered as water when the substance is heated or subjected to the action of a
dehydrating agent. It is sometimes assumed that the molecules of waterin a hydrated
colloid or salt have entered into combination as a whole ; while in hydroxides, the
constituent atoms have been rearranged, to form quite a different compound. The
two terms are therefore used to distinguish two conceptions regarding the relation
between the properties and composition of compounds containing water. In the
case of hydroxides, the atoms of water and of the oxide are rearranged during the
union so as to form hydroxyl or HO -groups. In hydrates the atoms of water either
combine as a whole with the compound or enter into some new relation with the
rest of themolecule'which is different from that which obtains in the case of hydroxides.
It is difficult to apply the definitions, hydrate and hydroxide, in practice, for they
are largely theoretical ; a true hydroxide may be so unstable that it decomposes into
water, etc., far more readily than a true hydrate. For example :
Hydroxide .
. Au(OH)
Ag(OH)
T1(0H)3
Zn(0H)2
A1(0H)3
NaOH
Decomposes at
15°
150°
230°
585°
850°
very high
T. Camelley and J. Walker i tried to measure the affinity of the oxides for water
by measuring the temperature of dehydration of the hydroxides. They showed
that the members of the even series of Mendeleeff's table, exhibit an increase in the
temperature of dehydration with an increase in atomic weight ; and with the odd
series, there is a decrease in the temperature of dehydration with increasing atomic
weight. In the same series of elements, the temperature of dehydration of the
hydrated oxides diminishes to the middle member and then increases.
To distinguish hydroxides from hydrates. — The specific or molecular volume of
compounds formed by the interaction of water has been suggested as a means of
throwing light on the difference in the effect produced by water in forming hydrates
and hydroxides. The specific or molecular volume is the quotient of the molecular
weight by the specific gravity. F. W. Clarke2 found that in the hydrated chlorides,
MCI2WH2O, the difference in the molecular volumes of the hydrates and anhydrous
salts divided by the number of molecules of water in the hydrated salt, varies from
125 to 150 — average 1376. If oxides be substituted for chlorides, the corre-
sponding quotient varies from 7 '4 to 19 "4. Hence, it was inferred that while the
specific volume of each H2O molecule in hydrated chlorides has a mean value 13*74,
no simple relations can be traced with the hydrated oxides in which presumably
a rearrangement of the atoms of the water molecule has taken place.
W. J. Perkin 3 attempted to distinguish the two types of combined water by
500 INORGANIC AND THEORETICAL CHEMISTRY
measurements of the magnetic rotatory power. If the value of this constant for water
be unity, and for an anhydrous compound be ic, then if the molecular rotatory power
of the hydrated compound be increased by unity for each molecule of combined water,
it might be concluded that the compound is a hydrate. Thus, the product of the
union of formic acid, H.COOH — molecular rotatory power 1-67 — with water,
1*676 4-0*995, and hence the product is a hydrate. With sulphuric acid
H2SO4 H2SO4+H2O HaS04+2H20
Molecular rotatory power , . . 2*315 3*188 4-113
Difference per molecule of H2O , . 0*873 0*925
Hence, W. J. Perkin concluded that the first hydrate of HoSO, viz. H2SO4.H2O, is a chemical
compomid — a hydroxide — which he symbolized, SO(OH)4, while the second hydrated
compound, HaS04.2H20, is a hydrate : SO(OH)4.H20.
The period of free vibration of molecular aggregates can be studied by absorption
or emission spectra. The high temperature generally required to produce emission
spectra may complicate the phenomena owing to the dissociation of the radiating
body ; on the other hand, absorption spectra can be obtained at low temperatures,
and this gives more definite knowledge of the molecular complexity. The period
of free vibration of an atomic group or radicle is not always greatly modified as the
molecular complexity of a series of compounds containing that radicle increases.
In 1882, W. de W. Abney and E. R. Festing * found that certain radicles showed
distinctive absorption bands at particular parts of the spectrum ; and ten years
later, W. H. Julius demonstrated that a chemical atom may lose its identity in a
compound, for the effect is not additive but constitutive ; and the absorption spec-
trum of a compound cannot be predicted from the spectra of the constitutive elements.
The intra-molecular character of the certain absorption bands, however, is evidenced
by the fact that the important groups of atoms and radicles which occur in chemical
compounds have a definite or specific effect upon radient energy, particularly upon
the so-called low frequency or infra-red radiation.. This effect is manifest by intense
absorption bands which occur in definite positions characteristic of the radicle or
group of atoms causing these bands.
W. W. Coblentz further argues that the absorption spectrum of a compound
with water of constitution will not be the composite spectrum of the anhydrous
substance plus water, because the combination with water completely changes the
physical character of the molecule ; on the other hand, if the molecules or groups of
atoms which cause the absorption band undergo no physical change when they com-
bine to form a crystal, or when they enter into solution, the absorption spectrum
will be a composite of the absorption bands of the constituents. The heat of hydra-
tion, in this latter case, might lead to the inference that a new compound is formed
although the bands due to the anhydrous substance and to water are the same in
magnitude and position as that which the constituents possessed before they united.
The radiometric test makes no distinction between water of crystallization, dissolved
or absorbed water, and water of solid solution. The behaviour of water in crystals
and in solid solution is identical with that of water in its free liquid state. In
illustration, the absorption spectra of selenite, opal, and the zeolites are identical
with those characteristic of free liquid water, and therefore it is inferred that the
water in these minerals remains intact as a group of molecules ; on the other hand,
in brucite, Mg(0H)2 ; diaspore, AIO(OH) ; bauxite, Al20(OH)4 ; gothite, reO(OH) ;
muscovite mica, H2KAl3(Si04)3 ; tremoHte, CaMgsJSiOs).!, the water is not present
in solid solution or as absorbed water, but is rather constitutional, being an integral
part of the molecule. Many of these minerals show a clear absorption band charac-
teristic of the OH group in alcohol.
The infra-red absorption spectra show that the following compounds probably contain
water of crystallization : heulandite ; stilbite ; potassium alum ; natrolite ; scolecite ;
analcite ; colemanite ; hexahydrated calcium chloride — CaCla-CHgO ; trihydrated
potassivun ferrocyanide, K4FeCye.3H20 ; apophyllite; deweylite; thomsonite; gismondite^
WATER 501
blodite ; thaumasite ; hydrotalcite ; varescite ; wavellite ; vivianite ; mellite ; and
Rochelle salt ; while the following compounds contain water oj constitution : manganite ;
gOthite ; bauxite ; turquoise ; lazulite ; hydrargillite ; diaspore ; datolite ; azurite ;
brucite ; prehnite ; hydronephelite ; pectolite ; chloritoid ; clinochlore ; penninite ; tour-
maline ; the micas ; muscovite ; biotite ; serpentine ; talc ; epidote ; sodium metaphos-
phate ; meta- and ortho-phosphoric acids.
C. Schaefer and M. Schubert ^ have studied the natural infra-red vibrations of
solids by the reflection method in preference to the absorption method. They
obtained characteristic reflection maxima for the S04-group in 34 sulphates ; and
for the COs-group in 15 carbonates. In agreement with W. W. Coblentz, they also
found that substances containing water of hydration show a reflection maximum
between 30 and 3"5/>t, and a long-wave maximum between 15 and 20/x. The
former is displaced in salts like cupric sulphate and the alums containing water of
hydration. Water of hydration seems to adapt itself to the symmetry of the
crystal, and it is anisotropic in anisotropic crystals, uniaxial in uniaxial crystals,
and biaxial in biaxial crystals. L. Vegard and H. Schjelderup studied the
structure of crystals by means of the X-ray reflection method with a view of
elucidating the role of the water of crystallization.
References.
1 T. Camelley and J. Walker, Jmirn. Chem. Soc, 53. 59, 1888.
2 P. W. Clarke, Amer. Journ. Science^ (3), 8. 428, 1874.
3 W. J. Perkin, Journ. Chem. Soc, 49. 777, 1886.
* W. de W. Abney and E. R. Festing, Phil. Trans., 172. 887, 1882 ; W. H, JuUus, Verh.
Akad, Amsterdam, 1. 1, 1892; W. W. Coblentz, Journ. Franklin Inst., 172. 309, 1911;
Phys. Rev., 16. 35, 1903 ; 20. 252, 1905 ; 23. 125, 1906 ; 30. 322, 1910 ; BuU, Bur. Standards,
2. 457, 1907 ; 4. 392, 1907 ; Jahrh. Bad. Eleh., 3. 397, 1907 ; Investigations of Infra-red Spectra,
Washington, 1905-8.
5 C. Schaefer and M. Schubert, Ann. Physik, (4), 50. 283, 339, 1916; (4), 55. 397, 1918;
L. Vegard and H. Schjelderup, ih., (4), 54. 146, 1918.
§ 11. The Vapour Pressmre o! Hydrated Salts
If a substance can form a number of definite hydrates, at any given temperature,
each hydrate has its own characteristic vapour pressure. For instance, from the
work of W. Miiller-Erzbach, H. Lescoeur, and J. L. Andreae,i it appears that copper
sulphate forms three hydrates with vapour pressures, at 50° :
Water. CUSO4.5H2O CuSO^.SHaO CUSO4.H2O
Vapour pressure . . 92-0 47-0 30*0 4-5 mm.
Suppose water vapour be slowly admitted to a suitable vessel containing
anhydrous copper sulphate at 50°, The two-component system has two solid
phases and one vapour phase ; accordingly, the system will be univariant, and the
vapour pressure will remain constant arid fixed at 4'5 mm. until all the anhydrous
copper sulphate has been transformed into a monohydrate : CUSO4-I-H2O
^CuS04.H20 ; the system is now bivariant because but one solid phase — the mono-
hydrate — is present, and this can exist in contact with a vapour pressure which
varies within certain limits. Hence, further addition of water vapour will be followed
by a rise in the vapour pressure. At 30 mm. pressure the bivariant system becomes
univariant owing to the appearance of a second phase — the trihydrate. The vapour
pressure will then remain constant and fixed at 30 mm. until all the monohydrate
has been transformed into the trihydrate : CuS04.H20-|-2H20=CuS04.3H20.
Further additions of water vapour will be attended by a rise of the vapour pressure
to 47 mm., and the vapour pressure will then remain stationary until all the trihydrate
has been transformed into the pentahydrate : CuS04.3H20-f 2H20^CuS04.5H20.
Any further addition of water vapour will raise the vapour pressure until the water
condenses to a liquid and gives the vapour pressure of a saturated solution of copper
sulphate at 50°. Further additions of water vapoui will simply increase the amount
502
INORGANIC AND THEORETICAL CHEMISTRY
of condensed water. When all the copper sulphate has dissolved, the vapour pressure
will be determined by the concentration of the solution of copper sulphate.
Conversely, starting with the pentahydrate, the reverse change will occur if
water be continually withdrawn from the system. The vapour, pressure of the
pentahydrate, 47 mm. at 50°, will be maintained as long as the system contains any
of the pentahydrate. When all the pentahydrate has decomposed, the vapour
pressure will drop suddenly to 30 mm. and remain stationary until all the trihydrate
has decomposed; the vapour pressure will then fall abruptly to 45 mm. and remain
at that value until all the monohydrate has decomposed into anhydrous copper
sulphate. These changes are usually shown diagrammatically by curves resembling
Fig. 26. Amounts of water, expressed in gram-molecules, are represented on the
horizontal axes, and the vapour pressures along the vertical axes. The constancy
of the vapour pressure of each hydrate is emphasized by the horizontal terraces on
the vapour pressure curve.
This step by step dissociation of the hydrates furnishes a method which is some-
times available 5or deciding whether or not definite compounds exist at definite
temperatures. If definite compounds are produced, the gradual addition or
removal of water vapour will alter the vapour pressure curve until a pressure is
reached which remains constant for a certain period, and then suddenly assumes
a new constant value. It must be added that some
(e.g. P. Blackman, 1911) consider the dehydration of
copper sulphate pentahydrate is comparable with the
removal of water from an ordinary aqueous solution ;
that the alleged breaks in the curve are due to a mal-
interpretation of imperfect experiments ; and that the
observed vapour pressures are points on a continuous
curve. This, however, does not interfere with the
principle involved.
By measuring the rate of decomposition of hydrated
aluminium and ferric hydroxides, W. Ramsay inferred
the non-existence of definite hydrates because no
signs of any discontinuity was observed in the rate of evolution of water ; but he
concluded that two were formed with lead oxide— 2PbO.H20 and SPbO.HgO.
J. M. van Bemmelen, for similar reasons, believed in the non-existence of definite
hydrates of silicic acid. The underlying hypothesis is that differences in the tenacity
with which the constituents of water are retained by the molecule, as evidenced
by the temperatures at which the water is expelled, are the result of differences in
the structure ; and that molecules of water which are simultaneously expelled at
a definite temperature occupy a similar position in the molecule.
Each hydrate has its own specific vapour pressure at a particular temperature,
The average vapour pressure of the water vapour in atmospheric air is equivalent
to 8 or 9 mm. of mercury. If the vapour pressure of the hydrate be greater than
the vapour pressure of the atmospheric moisture, the hydrate will lose water on
exposure to the air — ^in other words, the salt will be efflorescent ; on the contrary,
if the vapour pressure of the hydrate be less than that of the atmospheric moisture,
the salt will absorb moisture from the atmosphere, and be deliquescent. A few
illustrations are indicated in Table XVIII.
"" i 1 1 1 1 1 ■■■ "
-. A7mm.. . . . . _ .
\cus04., 5H2O ::::_:___::_::_:
. . . . . _ _ 30 mm. . .
I": :: : : CuSO^ 3H2O '- :: :
CilSQ.H,0
5 4. 3 2 1 0
Fig. 26. — Vapour Pressure
Curvea of the Hydrates of
Copper Sulphate.
Table XVIII.— Vapour Pressures of Hydrates.
Salt.
Vapour pressure, mm.
Property.
CaCla.GHaO
FeCla-GHaO
Na^SO^.lOHjO
NajCO,.10H8O
3-2
6-0
27-8
24-2
Deliquescent
Deliquescent
Eflfiorescent
Efflorescent
WATER 603
Whether a salt effloresces or deliquesces depends on the humidity of the atmo-
sphere in which it is confined. Thus, A. Vogel found that while hydrated copper
sulphate may be preserved unaltered for years, it effloresces rapidly in air dried by
sulphuric acid or calcium chloride. H. Watson also showed that sodium carbonate
does not effloresce between 6° and 12° in air with the dewpoint at 3° or 4°, but it
does effloresce in air at 14*4:° when the dew point is at 8"9° ; Glauber's salt effloresces
in air at 144° when the dewpoint is at 9*4°, but not when the dewpoint is over 10°.
According to H. Lescoeur,^ the vapour pressure for the deliquescence of hexahydrated
strontium chloride is 11*5 mm., and for efflorescence 5"6 mm. at 20°. It is
therefore possible to predict whether or not a salt will have a tendency to
efflorescence or deliquescence in an atmosphere of known humidity when the
vapour pressure of the hydrate is known,
A. C. Gumming, J. R. Partington, W. N. Rae, and F. Ephraim have shown that
when certain crystalline salts are dehydrated, the vapour pressure may rise very
slowly or even remain constant for a time, and then rise rapidly to the normal value
required for the given temperature. The suspended transformation, lagging, or
period of induction is shown by copper sulphate, CUSO4.5H2O ; sodium carbonate,
NagCOs.lOHgO ; barium chloride, BaCl2.2H20 ; racemic acid, C4H6O6.H2O ;
the double sulphates, M'2S04.M"S04.6H20 ; some ammino-salts ; etc. M.
Faraday showed that a perfectly sound crystal of hydrated sodium carbonate,
phosphate, or sulphate does not effloresce readily on exposure to the atmo-
sphere ; indeed, M. Faraday found such crystals may be kept for years in an open
dish without efflorescing. If, however, the change has commenced at any point,
it will spread quickly throughout the whole mass ; and this is in accord with the
phase rule F=C—P-\^2. In the perfect crystal there are two phases P, namely,
sodium sulphate, Na2S04.10H20, and water vapour ; and two components C,
namely, Na2S04 and H2O. Hence the system is bivariant (^=2), so that the
pressure of the water vapour and the temperature can be arbitrarily within certain
limits without altering the state of the system. If, however, some efflorescent salt
be present, there will be three phases, and the system will be univariant (^^^=1),
so that for every temperature there is one and only one vapour pressure for equili-
brium. The phase rule is only concerned with the conditions of equilibriimi, and
has nothing to say about how that state will be obtained.
The study of copper sulphate, which is typical of numerous other hydrates,
shows that the molecules of the combined water may differ in the tenacity with which
they are retained by the molecule of copper sulphate. The older books on chemistry
drew fine distinctions between the different combinations of water with a salt ; it
is now believed that the water of one hydrate does not differ in kind from that of
the other hydrates. The water is sometimes conventionally styled " constitutional
water," " water of crystallization," " water of hydration," or " water of com-
bination." The actual term used does not matter very much provided it is not mis-
understood. The proportion of water in the different hydrates is in accord with
the law of multiple proportions— if not, the water is arbitrarily said not to be chemi-
cally combined. The mode of writing the formulae — CUSO4.5H2O ; Na2SO4.10H2O
— and the ease with which the hydrates dissociate into water, etc., might give rise
to the idea that the water molecule exists in the hydrate ready-made. There
is, however, no evidence how the elements of water are combined in the hydrate,
and consequently, following the old adage— truth is to be found within ourselves ; it
takes no rise from outward things — many suggestions have been made to represent
the constitution of the hydrates graphically. The application of the theory of
valency to explain the composition of salts does not usually include water of
crystallization ; and in order to extend that theory to salts with water of crystalli-
zation, it is usually assumed that all or some of the contained oxygen atoms are
quadrivalent. For instance, assuming that in copper sulphate (1) the oxygen is
quadrivalent ; (2) one water molecule is associated in the molecule differently from
the other four molecules because it is not expelled except at a much higher
504 INORGANIC AND THEORETICAL CHEMISTRY
temperature ; (3) two of the remaining four molecules are more strongly attached
than the other two as is evidenced by the step by step dehydration of the
pentahydrate->trihydrate->monohydrate->anhydrous salt, one of the many
formula which can be devised is :
H
O^ .. ^0-0=H2
2
Cu<^>s<^_5ZS2
A Hg
HOOK
If all four water molecules were united in the same way it is further assumed that
they would be liberated together, or set free individually one after the other forming
respectively penta-, tetra-, tri-, di-, and mono-hydrates. There is some evidence
to show that when a double salt is formed containing the solvent, the latter is an
essential constituent of the salt, for the substitution of one solvent for another
may prevent the formation of a particular double salt. Thus H. W. Foote (1910) ^
found that potassium and mercuric chlorides form the following double salts with
water, alcohol, C2H6O, and acetone, CsHeO :
Water. Alcohol. Acetone.
... ... KCl.SHgClg.CaHoO
KCl.2HgCla.2H2O ... ...
5KC1.6HgCl2.2C2HeO 5KC1.6HgCl2.2C,HeO
KCl.HgCl2.H2O ... ...
2KCl.HgCl2.H,0 ... ...
We know very little beyond the simple facts that (1) water is a product of the
dissociation of the hydrates ; (2) the water of the hydrate is given off at com-
paratively low temperatures ; (3) the water is not an essential part of the reacting
unit in its most characteristic transformations ; (4) the water is not generally
necessary for the formation of the salt itself since the water of crystallization can
generally be removed by suitable means leaving behind the anhydrous salt ; and
(5) the water can often be replaced by an organic solvent so that a salt crystallizing
with a definite nimiber of molecules of water at a given temperature will crystallize
from one organic solvent with the same or a smaller number of molecules of the
solvent.
There is much cogent evidence leading to the inference that water is a ternary
substance containing molecules H2O, (H20)2, and (H20)3, and A. Rosenstiehl (1911)
considers that in salts containing water of crystallization, the water may be present
as H2O, (H20)2, or (H20)3, or a mixture of these different groups. The effect of
temperature on the dehydration of the hydrated salts will show the state of the
polymerization of the bound water. Salts containing 1, 2, or 3 molecules of water
of crystallization usually lose water in one step. Salts with 3, 6, 9, 12 molecules
of water on hydration lose water as 3H2O or multiples of this ; and salts with 4, 5,
7, 8, or 10 molecules of water furnish evidence that two kinds of molecules of
water are involved. For instance, the dehydration of CUSO4.5H2O behaves as
if the molecule contained CuS04.H20-f2(H20)2 ; MgS04.7H20, as if it were
MgS04.H2O+3(H20)2 ; while Na2C03.10H20 behaves as if the molecules
contained Na2C03.H204-3(H20)3.
If the term " water of crystallization " be carelessly employed it may suggest that
crystallization is somehow dependent on the presence of water, and this the more, as
efflorescent salts appear to lose their crystalline character when water is lost. Crystals of
gypsum — CaS04.2H20' — 'form a white chalky powder when the water is driven off ;
crystalline sodium carbonate, and Glauber's salts, likewise produce white powders when
their combined water is expelled. The powdered dehydrated substances are all crystalline.
Several zeolites may lose their combined water without losing their crystalline form. In
fact, practically all chemical compounds can be crystallized. Crystallization is not dependent
upon the presence of water. Sulphur, common salt, iodine, potassium chlorate, potassium
sulphate, and numerous other crystalline substances do not contain the elements of water.
WATER 505
Again, crystalline calcspar does not contain the elements of water, and yet when calcined
it gives a white powder. The calcspar loses carbon dioxide, not water.
Alcohol, C2H5OH, has a constitution similar to water, but one of the hydrogen
atoms of water is replaced by the radicle C2H5. Alcohol, ammonia, and hydrogen
peroxide can combine with certain other molecules to form complexes, and thus we
speak of " alcohol of crystallization," " ammonia of crystallization," " hydrogen
peroxide of crystallization," etc.
References.
1 W. Miiller-Erzbach, Zeit. phys. Chem., 17. 446, 1895 ; 19. 146, 1896 ; Wied. Ann., 23. 607,
1884 ; 31. 75, 1887 ; 32. 313, 1887 ; 34. 1047, 1888 ; Ber., 20. 1152, 1887 ; H. Lescoeur, Bull.
Soc. Chim.y (2), 46. 285, 1886 ; (2), 47. 30, 377, 1887 ; P. Blackman, Journ. Phys. Chem.,
15. 871, 1911 ; J. L. Andreae, Zeit. phys. Chem., 7. 260, 1891 ; J. M. van Bemmelen, Zeit. anorg.
Ghent., 13. 234, 1896 ; W. Ramsay, Journ. Chem. Soc, 32. 395, 1877 ; J. B. Hannay, ib., 32.
381, 1877 ; Min. Mag., 1. 106, 1877 ; G. Wiedemann, Journ. praU. Chem., (2), 9. 338, 1874 ;
A. Newmann, Ber., 7. 1573, 1874 ; A. F. Weinhold, Pogg. Ann., 149. 227, 1873 ; G. Tammann,
Wied. Ann., 33, 322, 1888 ; L. Schneider, Monatsh., 11. 166, 1890.
2 H. Lescoeur, Compt. Rend., 103. 1260, 1886 ; N. A. Rae, Journ. Chem. Soc, 109. 1229,
1917 ; A. 0. Gumming, ib., 97. 593, 1910 ; J. R. Partington, ib., 99. 466, 1911 ; F. Ephraim and
S. Millmann, Ber., 50. 529, 1917 ; F. Ephraim and P. Wagner, ib., 50. 1088, 1917 : M. Faraday, Pogg.
Ann., 33. 186, 1834 ; H. Watson, Phil. Mag., (3), 12. 130, 183S ; A. Vogel, Schweigger's Journ.,
22. 160, 1818.
3 A. Rosenstiehl, Compf. Rend., 152. 598, 1911 ; Bull. Soc. Chim., (4), 9. 281, 284, 1911 ;
H. W. Foote, Journ. Amen Chem. Soc, 32. 618, 1910.
CHAPTER X
SOLUTIONS
§ 1. The SolubUity oJ Solids in Water
In a strictly scientific sense of the word, insolubility does not exist. Even those
substances which are characterized by the most obstinate resistance to the solvent action
of water should probably be regarded as extraordinarily difficult of solution, but not
insoluble.— O. N. Witt (1905).
Water is one of the most active of solvents, and, in consequence, it has been styled
the universal solvent, but not in the same sense as the visionaries' alcahest (universal
solvent) so often mentioned in the writings of mediaeval alchemy. It is remarkable
that a belief in this, Paracelsus' menstruum, was fairly prevalent towards the
beginning of the eighteenth century. R. Boyle i said : " He that hath seen it,
hath more reason to believe it, than he that hath not." J. Kunckel 2 said that
some derive the term alcahest from the Latin alkali est, others from the German
all geist — all spirit —and yet others from the German allesest — it is all. J. Kunckel
also expressed his surprise that it does not seem to have occurred to the old
alchemists that no vessel on earth could hold their universal solvent, because a
universal solvent would also dissolve its containing vessel.
For convenience, the dissolved substance is often called the solute, and the
liquid in which the solute is dissolved the solvent. When no solvent is mentioned,
water is usually understood ; the list of possible solvents is almost as extensive as
the list of chemical compounds. If potassium chloride be added to water kept
at a constant temperature the salt is gradually dissolved, and the process of solution
continues until a definite amount has dissolved. Any soUd in excess of this will
remain an indefinite time without further change, provided the temperature remains
constant, and no solvent is lost by evaporation, or gained by absorption. The
solid and solution are then in equilibrium. As in the analogous case of the pressure
of a liquid, the equiUbrium between a saturated solution and a solid is dynamic,
not static. Accordingly, with the preceding notation the equilibria with solid
and liquid solutes respectively are represented :
SolutesoUd^SolutCdissoived ; or Soluteiiquid^Solutedissoived
A solution in equilibrium with its solid is said to be saturated with the solid at the
temperature of experiment. The weight o£ salt dissolved by 100 c.c. of the solvent
so as to make a saturated solution at any assigned temperature is called the
solubihty oi the salt. Thus, 100 c.c. of water at 20° will dissolve 35 grams of potas-
sium chloride, and accordingly, 35 is the solubility of potassium chloride in water
at 20°.
J. L. Gay Lussac (1819) ^ expressed the solubility as parts of the substance in 100
parts of water. Other modes of representing solubility are more convenient in special
cases — e.g. the percentage amount of salt in a given weight of the solution may be employed,
and A. Etard (1884) represented the solubility in terms of the weight of the solvent in
100 parts of solution. With the latter mode of representation, the solubility curves are
usually straighter, and very great solubilities cannot exceed 100, whereas with J. L. Gay
Lussac's method they may become infinite. For example, at 310° a saturated solution of
sodium hydroxide has 22,222-2 parts of the solid in 100 parts of water, and 99-45 grms. of
the hydroxide in 100 parts of solution. E. Cohen and E. H. Biichner, however, have shown
506
SOLUTIONS 507
that A. Etard's rule (1898)— to the effect that if the solubility be defined as the weight of
salt in 100 grms. of saturated solution, the solubility- temperature curves are straight
lines- — is not in accord with facts.
It is common in studying the physical properties of solutions, to represent the ratio
in terms of the number of gram-molecules of the dissolved substance per 100 of the solvent,
or of the solution ; or as the molecular fraction as it is called, that is, the ratio of the number
of gram-molecules of the solute to the number of gram -molecules of the solvent and solute.
If S denotes the quantity of the substance by weight dissolved in 100 parts by weight of the
solvent, and W the quantity by weight in 100 parts by weight of the solution ; then, if
S parts of the substance are contained in 100 parts of the solvent, the S parts of the substance
are contained in 100 ^-'S' of the solution ; and consequently, 100 parts of the solution will
contain W = 100S/( 100 +/S) parts of the dissolved substance. Conversely S = 100 Wl{ 100 - W).
This gives a relation between S and If, provided no marked change in volinne occurs on
solution.
There are also many ways of respresenting the concentration of a solution, for example,
the concentration of ordinary Sulphuric acid can be represented: (1) By the specific
gravity (or density) 1'161. GHiis mode of representation m\ist be supplemented by tables
relating concentration and specific gravity. (2) By the stoichiometric proportion of
H2SO4 — (i) 22-27 per cent, by weight of H2SO4; (ii) 258 grms. of H2SO4 per litre;
(iii) 258-^98 = 2-63 gram-molecules per litre; or (iv) H2SO4 + I9H2O. (3) By the
normality of the solution 258-^49 = 5-26iV— that is, 5*26 equivalents of H2SO4 per litre.
(4) Molecular fraction 0*05, meaning that 0*05 gram-molecule of H2SO4 is mixed with
I— 0*05 =0-95 gram-molecule of water.
It is important in measuring the solubility of a salt to make sure that the solution
is really saturated because some salts dissolve very slowly. Many of the older
determinations are vitiated by failure to guard adequately against supersaturation,^
and by using inadequately purified salts.
The concentration of a solution is determined by the relative amount of solute
in solution — if but a small proportion is present, the solution is said to be weak
or dilute ; if a relatively large amount of solute is dissolved, the solution is said to
be strong or concentrated. We can thus see with C. L. Berthollet (1803) a close
analogy between the solution of a salt in water, and of water by air. In each, the
quantity dissolved at a given temperature is always the same — in the case of a
solution of salt in water, this constant is called the solubility of the salt, and in the
case of a liquid in air, the vapour pressure of the liquid.
J. H. van't HofE's definition (1890) 5 of a solution is one of the best yet suggested.
It runs somewhat as follows : A solution is a homogeneous mixture of two or more
substances ; the composition of the mixture can vary within certain Umits — the
Hmits of its existence ; or, as C. L. Berthollet expressed it in 1803 : Les sels
s^unissent d Veau en toute proportion, jusqu'au point de la saturation. It has been
conventionally agreed to call solutions mixtures because their composition can
vary in the way just described and not per saltum as is characteristic of that mode
of chemical combination defined by the laws of constant and multiple proportion.
F. Wald distinguishes a chemical individual as a substance which persists as a
phase of constant composition when the conditions of temperature, pressure, and
composition of the other phases present, undergo continuous alteration within
certain limits — the limits of existence of the substance.
Pliny commented on the limited solubility of salts in water. In his Historia
naturalis (31. 34) of the first century, he said :
It is a singular fact that if more than one sextarius of salt be put into four sextarii of
water, the solvent action of the water will be overpowered, and no more will dissolve.
N. le Febure ^ in the seventeenth century stated the law of saturation very clearly :
Digest four ounces of ordinary salt in eight ounces of water, and you will find that the
water will dissolve three ounces of the salt, and that it will take up the other fourth if the
water be boiled and the liquid agitated. . . . When a menstruum is fully saturated' — either
cold or hot — it is impossible by any art to go further, because it is charged conformably
with le poids de nature, which cannot be transgressed.
The numbers expressing the solubility of a salt were thus regarded as natural
constants — le poids de nature.
Speaking in terms of the phase rule, the solubility of a solid in a liquid is the
508 INORGANIC AND THEORETICAL CHEMISTRY
saturation concentration, and at an assigned constant temperature the system is
invariant ; and, with liquid-liquid systems, two liquid layers are necessary for
invariance ; and, in dealing with gas-liquid systems, the pressure of the gas must
be specified. It is also necessary to consider the formation of compounds of solute
and solvent, for, at a given temperature, some compounds with the solvent may be
stable, others unstable. When a compound is decomposed by water, its solubility
has no more meaning than to speak of the solubility of zinc in dilute sulphuric acid.
The term solubihty is loosely appHed to both phenomena because in each case the
soUd phase disappears, and the material passes into the liquid until the Hquid is
saturated. In a rough way the term dissolution is apphed when the substance
dissolved is decomposed by the solvent, and solution when it is not decomposed.
Are all substances soluble in water ? — Excluding chemical action, so-called,
there are three possible ways in which two substances can behave : (1) One
substance may be quite insoluble in the other — e.g. platinum in water ; (2) One
substance may be partially soluble at a given temperature — e.g. salt in water ;
(3) The two substances may be completely miscible in all proportions — e.g. alcohol
and water ; fused cobalt and nickel. It might be argued that a substance must
either be soluble or insoluble in a given menstruum — either it will be diminished
in mass by the solution of a portion in the menstruum, or it will suffer no change
after prolonged contact therewith. Very exact investigations have shown that
few substances considered by the chemist to be insoluble really are so. The so-
called insoluble substances obtained as precipitates in analysis are in reahty sub-
stances with a very low solubility. It is all a question of measurement. As the
methods of observation become more and more precise, so does the list of insoluble
substances grow less and less. The general use of the term insoluble must, in
consequence, give way to sparingly soluble. In illustration, the three precipitates
obtained in the first group of the regular scheme for qualitative analysis are usually
said to be insoluble, but they are not really insoluble in water because their solu-
bilities, per 100 c.c. of water, at 20°, are represented by the following numbers :
silver chloride, 0*00016 gram ; mercurous chloride, 0'00031 gram ; and lead chloride,
1*18 gram. In some cases the alleged solubility — e.g. platinum in water — cannot
be proved directly, but requires involved reasoning which appears to be subtle
sophistry of no substance or profit.
The influence of the grain-size of solids on the solubihty.— In 1813, W. H.
Wollaston noted that finely-divided substances suspended in a solvent not only
dissolve more rapidly but they may have an even greater solubility than coarse-
grained powders ; in 1870, E. Divers also made the same observation with respect
to calcium carbonate ; and G. A. Hulett found that a litre of water at 25° will dissolve
2 '085 grms. of gypsum when particles have an average diameter of about O'OOOl cm.
and 2-476 grms. when the average diameter is about 0'00006 cm. The theory was
worked out by J. W. Gibbs ^ in 1876, and by J. J. Thomson in 1888. J. W. Gibbs'
theor}^ is embodied in the expression
log ^2^ Wi_ 1 \
where R denotes the gas constant, viz. 8'315xl0'^ ergs per degree ; T, the absolute
temperature ; D, the specific gravity of the solid ; M, the molecular weight of the
solute ; o- the energy per unit area of the surface of separation between the solid
and solution ; and Si and S2 denote the concentrations of saturated solutions in
contact with spherical particles of the respective radii r^ and ^2- If ^2 ^^ indefinitely
large, the expression reduces to that employed by G. A. Hulett, for S2 then denotes
the ordinary or normal solubility of the substance. For calcium sulphate, where
D=2 33 ; M=136 ; (r=1050 ergs per sq. cm. at 25°,
, Si 2 Ma- , >Si 14-69x10-4
SOLUTIONS
509
if D and o- are independent of temperature. On calculating the solubilities of
gypsum for different values of / and T it is found that the solubility curve for r
=500/x is virtually the same as when r is infinite. The results shown in Fig. 1
represent the solubiHty curves (milligram-molecules per litre) for r=0*5)it ; r=l'0/Lt ;
/•=3'0/>t ; and r=50'0)Lt.
A solution in equilibrium with fine-grained particles, say 0*5/x, is super-saturated
with respect to coarser-grained particles, say 50/x. Consequently with a mixture
of coarse and fine grains, the coarse grains will grow at the expense of the fine
grains. In illustration, a fine-grained precipitate, after standing some time in contact
with its solution, becomes coarser-grained, so that the freshly-made precipitate
readily passes through the filter paper, while the older precipitate does not
pass.
E. Podszus found that certain oxides — alumina, thoria, and zirconia — usually not
acted upon by the hydrochloric acid are dissolved by this reagent when they have
been reduced to a fine state of subdivision so that the particles have a diameter of
the order l/x. The dissolution of the oxides in hydrochloric acid is a phenomenon
different in kind from the solution of, say, gypsum in water. W. Herz calculated
the molecular diameter, d, of liquids from the equation d=2yv/Lj where y denotes
the capillary constant, v the specific volume, and L the latent heat of evaporation
per gram. He then examined the relation between this magnitude and the solu-
bility of the liquids in water, and found that in general
the solubiHty is greater the smaller the diameter. The
rule can be entirely altered by specific chemical properties.
The effect of grain size on solubility recalls the fact
that when drops of liquid are suspended in air or other
gas, the smaller drops of liquid grow smaller and dis-
appear, so that the larger drops grow larger at the expense
of the smaller drops. The vapour pressure of a liquid
depends on the curvature of its surface; the greater the
curvature the greater the vapour pressure, and hence the
vapour from the smaller drops is distilled on to the larger
drops — 1. 9, 6. The two phenomena are not strictly Fig. 1. — Solubility
analogous except in this way. The boundary-surface ^^™, °^, t?-«'''''"J
V J. T • 1 J ^' ^ • j.-^ 1. £ _x • J. Sulphate of Different
between a liqmd and a soiid is the seat of a certam amount GrcSn-size.
of energy — the so-called free surface energy of the liquid.
The greater the curvature of a liquid, the greater the surface energy. The greater
the free surface energy of a substance, the greater the solubiHty — e.g. the allotropic
forms of a substance have different solubilities, the less stable is always the more
soluble. Hence, P. Curie inferred that the greater the free surface energy between
a solid and its solution, the greater the solubiHty.
Is water in aqueous solutions identical with water alone ? — When the absorption
of light by a given layer of an aqueous solution is compared with that of a layer
of water of the same depth, it cannot be assumed that the water in the aqueous
solution absorbs as much light as pure uncombined water ; and that the difference
between the light absorption of the aqueous solution and of pure water is due to
the dissolved substance. The different transparency of the water in a solution as
compared with water alone must be ascribed to a relation between the dissolved
substance and the solvent water ; part, at least, of the water must be different from
water alone, and the most probable hypothesis is that the water is partly de-
polymerized by the solute or that part of the water present in a solution is in com-
bination with the dissolved substance. Of the salts examined by H. C. Jones
(1913) and his co-workers, those which do not form hydrates absorb practically
the same amount of light as a corresponding layer of water. A difference in Hght
absorptive power is only exhibited by solutions of those substances which form
hydrates ; this is taken to mean that the difference between Hght absorbing power
of solutions of hydrated salts and the corresponding amount of the solvent is not
40* 60*
Temperatures
510
INORGANIC AND THEORETICAL CHEMISTRY
fully explained by the depolymerization or the breaking down of associated
molecules of water by the dissolved substance.
The influence of temperature on solubility.— The solubility of most substances
increases with the temperature ; the higher the temperature, the greater the
solubility. Graphs obtained by plotting the relation between the solubility of
sohds and temperature are called solubiUty curves. The solubility curve presents
a graphic picture which enables the relation between solubility and temperature
to be seen at a glance. In illustration, the upward left-to-right slope of the
solubility curve of calcium sulphate shows that the solubility of this salt
increases with a rise of temperature up to about 40°, and the downward left-to-right
slope over that temperature shows that the solubility then decreases as the tempera-
ture rises. Sodium chloride is but slightly more soluble in hot than in cold water.
The solubilities of a few typical salts at 0°, 50°, and 100° are as follows :
Solubility of
Potassium hydroxide, KOH
Sodium chloride, NaCl.
Calcium hydroxide, Ca(0H)2
Calcium chromate, CaCrOi .
Cerium sulphate, CegiSO^jg .
The solubility of a substance depends on so many complex factors that a
satisfactory quantitative theory has not yet been established. E. Clapeyron and
R. Clausius' equation can be written :
0°
50°
100"
. 9700
140-00
178-00
. 35-63
36-67
39-12
. 0-14
0-10
0-06
. 4-50
1-12
0-42
. 1909
4-78
0-78
d log S Q
dT
RT^
S dT RT^
1
or -^ ^7;;= —
(1)
so as to show the relation between the absolute solubility, S, and the temperature
coefficient, dSjdT, of the solubility — i.e. approximately the change in solubility
per degree — and the reversible heat of solution, Q. The gas constant R is nearly
2 calories. It is usual to represent the observed data between the concentration
S and the temperature 6° by an empirical formula of the type, S=ad-{-hd^
+c03_|_, , ,^ where a, b, c are constants to be evaluated from the measurements
of the solubilities S at temperatures di, d^, 6^, . . . R. T. Hardmann and J. R.
Partington ^ used the empirical expression log S=A—BT-^—C log T, which
contains three constants like the simpler relation, S=a-\-W -\-cd'^ .
Starting from F. M. Raoult's vapour pressure law, G. Bodlander calculated
the solubilities of some very sparingly soluble salts from the heats of formation Q
of an equivalent amount of the salt, and the electrode potentials of their ions — E*
for the cation, E' for the anion :
«H^+i>og^=^"+^'2l
where n and n' respectively denote the valencies of cation and anion, and the
solubility /S is expressed in gram-equivalents per litre. It is here assumed that the
free energy of the reaction is equal to the total energy change. F. Dolezalek also
calculated the solubility of gases in Uquids on the assumption that Raoult's law is
valid. J. H. Hildebrand deduced the following expression for the solubility N of
a solid at the absolute temperature T :
where njN denotes the solubihty of the compound expressed in terms of the
molecular fraction — n representing the number of gram-molecules of the solute in
the solution, and N the total number of gram-molecules of solvent and solute ;
A denotes the heat of fusion per gram-molecule assumed to be independent of the
temperature ; Ty^^ the absolute melting temperature of the solute. It follows
SOLUTIONS 511
from this conclusion that the solubility of a solid is smaller the greater the heat
of fusion, and the higher the melting point over T°.
D. Tyrer assumes that the solubility of a given substance depends not only upon the
temperature and nature of the solvent, but also on the mass of the solvent contained in
unit volume of the solution. The solubility of a substance in a given solvent is always
diminished when the solvent is diluted with a liquid in which the given solute is insoluble.
On this assumption he deduces the relation, Sn=za{V/v)n — b, which also contains three
constants, n, a, and b. V represents the total volume of the solution and v the specific volume
of the solute. Sufficient data have not been published to establish this relation.
The influence of pressure on solubility. — The effect of pressure on solubility in
condensed systems — liquids and solids — is relatively small — one per cent, per
1000 atm. — when contrasted with the effect of temperature, and it may be either
positive or negative. Pressure has but a slight influence on binary condensed
systems generally. The most accurate work on the effect of pressure on solubility
is that by E. Cohen and co-workers ^ on the solubility of sodium chloride and
mannite ; when at 24*05°, it was found :
Pressure .... 1 250 500 1000 1300 atm.
Solubility .... 26-41 26-60 26-76 27-02 27-20 per cent.
and H. F. Sill's work on sodium chloride and barium hydroxide, Ba(0H)o.8H20,
where it was found, for the latter, at 25° :
Pressure ....... 1 25 490 atm.
Solubility 8-299 8-790 9-366 per cent.
In 1862, K. Moller stated that that pressure must exercise an influence on the
solubility of a salt ; and in 1863, H. C. Sorby i^ made some remarks on the
subject. The solubility of a salt is increased by pressure if, during solution, a
contraction occurs ; and conversely, the solubility of a salt is decreased by
pressure if an expansion occurs during solution. For example, the percentage
changes in the volumes of solid sodium and ammonium chlorides over their
volumes in a saturated solution are respectively +13*57 and — 15*78 ; the
percentages changes in solubilities per atmosphere increase of pressure are re-
spectively +0*00419 and —0*00638, when the + signs denote increases, and the
— signs decreases. In 1870, C. M. Guldberg deduced a general expression for the
change of solubility >S which occurs when the pressure changes by an amount dp.
This is usually expressed in the form :
d log S 8v 1 dS_8v^
dp -RT' ""^'S'dp'RT • • • • (2)
These expressions follow directly from Clapeyron's equation. The observed results
are in agreement with these formulae when hv denotes the change in volume which
occurs during the solution of the solid, and dS/dp, the pressure coefficient of the
solubility — that is, the change of solubility which occurs when the pressure changes
one unit. F. Braun made a special study of the subject in 1870, and this work
has crystallized in the statement : The solubility of a salt will increase with
pressure if the solution occupies a less volume than the sum of the volumes of its
constituent parts ; while the solubility will diminish if the solution occupies a greater
volume than the sum of the volumes of its constituent parts. This is but a
specialized form of the so-called generalization of G. Kobin in 1879 : At constant
temperature there is one definite pressure at which a system will be in equilibrium ;
on raising the pressure, the reaction will take place in that direction which is
produced with a decrease in volume ; while if the pressure be reduced, the reaction
will proceed in that direction which has the greater volume. This, again, is a
special case of J. H. van't Hofi's law of mobile equilibrium ; which in turn is a
special case of the principle of least action, foreshadowed in a vague sort of way by
512 INOKGANIC AND THEORETICAL CHEMISTRY
Maupertius in 1747 — all natural changes take place in such a way that the existing
state of things will suffer the least possible change.
By division of the expression (1) for the relation between the temperature
coefficient of the solubility, dS/dT, and the heat of solution Q by the above expression,
(2) for the relation between the pressure coefficient of solubility, dS/dp, and volume
change dVy it follows that
where V2--V1 represents the change in volume, in c.c, which occurs when a gram-
molecule of the solid is dissolved at the temperature T in an unlimited quantity of
the saturated solution ; Q represents the heat of solution under these conditions.
The term dp may be taken to represent the increase in pressure necessary to cause
one gram more of the solute to pass into solution and dT the increase in temperature
necessary to produce that result ; or dS/dT, the temperature coefficient of the
solubility represents the change in solubility per degree change of temperature,
and ds/dp the pressure coefficient of the solubility. Values of Q for barium hydroxide
calculated from this equation agree well with the observed.
Since chemical equilibrium n is determined only by the relative concentration
of the different kinds of molecules concerned in the reaction, the equilibrium can
be altered by pressure only by changing the relative concentration of the substance
concerned in the reaction ; but the compressibility of liquids and solids is small, and
differences in the compressibility of the components in a reaction must therefore
be very small. Consequently, the effect produced by changes of pressure on
chemical equilibrium in condensed systems must be small. When one of the
components is a gas, the case is different because gases are highly compressible,
and their reactivity is almost proportional to the pressure. A compound involving
a volatile component will not be formed in a reaction unless the concentration or
partial pressure exceeds a certain limiting value which is mainly dependent on the
temperature. For instance, liquid water will not be formed at 200° if the pressure
is less than 15 atm., and at 300° if the pressure be less than 100 atm. Calcium
hydroxide in an atmosphere of steam at 550° and one atm. pressure, does not
dissociate into water and calcium oxide, but at 750° a pressure of 15 atms. is required
to prevent dissociation.
According to G. Tammann,!* if a solvent and a solution be subjected to a certain pressure
p, it is sufficient to raise the pressure on the solvent by a certain amount of 8p in order
that it may behave like the solution with respect to volume, temperature, and pressure.
The extra pressure 8p required to make the coefficient of thermal expansion of the solvent,
or the coefficient of compressibility of the solvent, equal to that of the solution under the
standard pressure, depends upon the concentration and nature of the solute. G. Tammann
explains the phenomenon by assuming that internal pressure is raised by the solution of a
substance in the solvent, so that the solvent requires an additional external pressure to
compensate the extra internal pressure of the solution. Under these circumstances the
equations of state of solution and solvent are the same.
Transition temperatures. — Some solubility curves exhibit irregularities at
certain temperatures. The solubility curve may change its direction, as calcium
sulphate does at 35°, and barium butyrate at 45°. The solubility curve
of sodium sulphate is a very trite illustration, but none the less instructive on
that account. It is shown in Fig. 2.1^ The solubility of sodium sulphate, said
J. L. Gay Lussac, follows une marche trhs singulihe for the solubility of the salt,
Na2S04.10H20, increases rapidly with rise of temperature, as shown by the slope of
the curve EO, Fig. 2. There is an abrupt change in the direction of the solubility
curve at 32*383°, 0, Fig. 2. Above that temperature the solubility decreases
with rise of temperature. This, said J. L. Gay Lussac in 1819, is the second
example of a body whose solubility decreases with a rise of temperature, for
J. Dalton had previously shown that lime behaves in a similar manner.
SOLUTIONS
51^
The break — foint de rebroussement — in the solubility curve of sodium sulphate,
the first of its kind, was discovered by J. L. Gay Lussac in 1819, and in 1839 he recog-
nized that the breaks in the solubility curves of some substances can be accounted
for by assuming that at this point it is no longer the same substance which dissolves
further. In 1840, H. Kopp showed that the solubility curves above and below
the point de rebroussement are two distinct curves representing the solubility of two
different substances. The one curve below the transition point can be represented
by the formula /Si=5-02+0-30594^— O-OOO41O02_|_o-O(X)9977^3 ; and the other
by /S2=58-50— 0-27783^+0-0(X)6900^2_j_0-000(X)49802^3. At the transition
point 81=82, and 6 then becomes 32'93°. The observed value is a little lower than
this, viz. 32'383°. At the transition temperature, adds H. Kopp, the crystallized
sodium sulphate passes into the anhydrous salt. Consequently, the curve of
increasing solubility of temperature below 32*383° represents the solubility of
curve of the decahydrate, Na2S04.10H20 ; and the curve of decreasing solubility
with rise of temperature represents the solubility curve of the anhydrous salt,
Na2S04. The decahydrate, at 32-383°, is transformed into the anhydrous salt.
The decahydrate is not stable above 32*383° ; the ,
anhydrous salt is not stable below 33°. This tem-
perature is called the transition temperature or
transition point, and the change is symbolized :
o2*383°
Na2SO4.10H26^Na2SO4-fl0H2O
The solubility curves, it will be observed, represent
the conditions of equilibrium between the solvent and
salt. It makes no difference whether we start with the
anhydrous sulphate or the decahydrate. When in
equilibrium, the solution in contact with the solid will
contain the amounts of sodium sulphate — Na2S04 —
indicated by the solubility curves, Fig. 2. The
saturated solutions, when in equilibrium, have the
same concentration and are identical in every way.
We cannot continue the observation of the solubility
of the decahydrate beyond 32*383°, because it im-
mediately splits up either into a less hydrated form —
e.g. Na2S04.7H20— or the anhydrous form, Na2S04.
The solubility curve of the heptahydrate meets the
solubiHty curve of the anhydrous sulphate in the region of instability ; the
transition point from the heptahydrate to the anhydrous salt is 34°, or
34°
Na2S04.7H20^Na2S04+7H20
The so-called eutectic points E and E2 will be discussed later, but since the trans-
formation of the anhydrous salt into the hydrate takes an appreciable time, it is
possible to measure the approximate solubility of the anhydrous salt below 32*8.
This is indicated by the dotted line in the diagram. In saturated solutions of
hydrates, a definite hydrate is in dynamic equilibrium with the solution ; if the
hydrate changes as shown by E. Demar9ay's study (1883) of the hydrates of thorium
sulphate, the maximum amount of a salt which can enter into solution depends
on its temperature and on its state of hydration ; the solubilities of the different
hydrates of a salt are different, and at the transition temperature, there is a break
in the continuity of the solubility curve. H. W. B. Koozeboom's studies of the
hydrates of a number of salts show that the solubility curves of the different
hydrates of a salt indicate the limits of their stability.
The solubilities of the two sodium sulphates — anhydrous and decahydrate —
are quite different. If the solid decahydrate were in contact with a saturated
Fig. 2. — Solubility Curve of
Sodium Sulphate.
VOL. I.
2 L
514 INORGANIC AND THEORETICAL CHEMISTRY
solution at 20°, and some of the anhydrous sulphate were added to the solution,
some of the latter would dissolve and be deposited later as the decahydrate.
The final result would be a transformation, through the medium of the solution,
of the anhydrous salt into the decahydrate. Although 100 c.c. of water at
0° can only dissolve about 5'0 orrams of the decahydrate, the same quantity of
water can dissolve much more of the anhydrous sulphate. The general result of
a multitude of experiments is to show that salts which crystallize in two or more
difEerent forms with difierent amounts of combined water, have different solubilities ;
and at certain temperatures a solution may be saturated with either of two different
hydrates, e.g. Na2S04.10H20, or Na2S04.7H20 ; it is therefore necessary to specify
which sodium sulphate is in question when reference is made to a saturated solution
of sodium sulphate. Of two hydrates that containing the less water is usually the
more soluble at any temperature below the transition temperature — H. le ChateUer's
rule. For instance, sodium sulphate forms the hydrates, Na2S04.7H20 and
Na2SO4.10H2O, and 100 grams of a saturated solution of the former at 10° has 23"1
grams of the former and 8*3 grams of the latter. The rule is not general ; the
hydrates of manganous sulphate do not fit the rule.
The solubiUty curve of anhydrous rhombic sodium sulphate progresses from
O3 into the metastable region. The solubiHty curve is at first retrograde —
decreasing with rise of temperature — and it then becomes normal — increasing
with rise of temperature. A. Smits explains the retrograde solubility curve of
rhombic sodium sulphate by assuming a retrogression of the degree of hydration
of the salt in solution with a rising temperature. At the transition point, 234°,
the rhombic crystals of sodium sulphate pass into the monoclinic form :
234°
Na2S04rhombic^^Na2SO4nionoclinic
The solubility of anhydrous monoclinic sodium sulphate is wholly retrograde, and
at the critical temperature (365°) the concentration of the solution is so small that
the critical temperature is virtually the same as that of water. A. Smits assumes
that the strongly retrograde solubility of monoclinic sodium sulphate indicates that
the latent heat of liquefaction of this salt is much less than that of the rhombic
salt. In the diagram, the concentration near the point (7 is on a much enlarged
scale in order to make the relations clear, for the curve up to C represents the
solubility of sodium sulphate in the vapour phase, and hence this curve virtually
coincides with the H2O axis. At the critical temperature of the solution, the
liquid and vapour perhaps have the same composition, and the two curves join up
with one another.
Is a heterogeneous solution to be regarded as a phase ?— In heterogeneous
solutions there are an infinite number of phases because every different degree of
concentration can be regarded as a phase. The pha§p rule is concerned with con-
ditions of equilibrium, and a heterogeneous solution is not in equihbrium because
there is a tendency to diffusion. Hence, the phase rule is not needed to determine
if such a solution is in equilibrium. If sulphur be placed in contact with iron,
it might be said that, neglecting vapour, there are two components, and two phases,
and therefore the system is univariant. Hence, sulphur and iron will not interact
when heated. It will be noticed, however, that the mixture of sulphur and iron is
not a system in equilibrium ; the two elements are not phases of a prior system, or
molten ferrous sulphide, FeS, on cooling would separate into particles of free sulphur
and free iron. Consequently, the phase rule does not apply.
Is a solution to be regarded as a one-phase or as a two-phase system ?— The
decrease in the solubiHty of a substance with rise of temperature is due to the
solute changing its nature thus, the diminishing solubiHty of sodium sulphate,
Na2S04.10H20, above 33° is referred to the passage of the decahydrate into the
anhydrous salt, Na2S04 ; with calcium hydroxide, Ca(0H)2, too, the change is
usually attributed to the transformation of some hydroxide into oxide, CaO. In
SOLUTIONS 616
general, a turning point in the solubility curve shows that the solid phase in the
saturated solution is changing. From this it follows that the molecules of a substance
in solution may retain their individuality and that they can undergo changes in.
the solution similar to those they suffer when heated alone. H. C. Jones and
J. S. Guy 14 showed that water which is combined with salts in solution is far more
transparent than pure water ; and J. E. L. Holmes and H. C. Jones, that the rate
of saponification of methyl acetate or formate is likewise faster with combined than
it is with free water.
While a solution in equilibrium can be said to have the same composition in
all its parts, so that it cannot be separated by mechanical or physical operations
into different individual parts, yet, according to the molecular theory, there must
be a limit to the subdivision beyond which the solution can no longer be regarded
as homogeneous. Consequently, there is no clearly defined line of demarcation
between heterogeneous and homogeneous mixtures. A so-called homogeneous
solution, for instance, can sometimes be separated into its component parts by
certain membranes, just as a mixture of gases can sometimes be separated into its
constituent parts by atmolysis. A homogeneous solution, or a mixture of gases,
however, is considered to be a homogeneous one-phase system because diffusion
maintains one uniform concentration throughout its mass.
References.
1 R. Boyle, The Sceptical Chymist, Oxford, 1661 ; A. Baume, Chymie experimentale et raison^e
Paris, 1773 ; H. Boerhaave, Elementa Chemice, Lugduni Batavorum, 1732.
2 J. Kunckel, Vollstdndiges Laboratorium chymicum, Berlin, 475. 1767.
3 J. L. Gay Lussac, Ann. Chim. Phys., (2), 11. 298, 1819 ; A. Etard, ib., (7), 65. 344, 1898 ;
Compt. Rend., 98. 993, 1884 ; E. Cohen and E. H. Biichner, Proc. Akad. Amsterdam, 3. 561,
1901 ; R. Abegg, Zeit. phys. Chem., 15. 241, 1894 ; V. Rothmund, ib., 26. 433, 1898.
* J. N. Legrand, Ann. Chim. Phys., (2), 58. 428, 1835 ; J. L. Gay Lussac, ib., (2), 11. 298, 1819.
« J. H. van't HofE, Zeit. phys. Chem., 5. 323, 1890 ; F. Wald, ib., 24. 648, 1897 ; W. Ostwald,
Lehrbuch der allgemeinen Chemie, Leipzig, 1. 606, 1903 ; The Fundamental Principles of Chemistry,
London, 1909 ; C. L. Berthollet, Essai de statique chimique, Paris, 1803.
« N. le Febure, Traicte de la chymie, Paris, 1. 381, 1660.
7 W. H. WoUaston, Phil. Trans., 103. 51, 1813 ; W. Ostwald, Zeit. phys. Chem., 34. 965,
1900 ; G. Hulett, ib., 37. 385, 1901 ; 47. 357, 1904 ; G. A. Hulett and L. Allen, Journ. Amer. Chem.
Soc, 24. 667, 1902 ; E. G. Pierce, Chem. Eng., 24. 62, 1906 ; H. B. Weiser, Journ. Phys. Chem.,
21. 314, 1917 ; P. Curie, Bull. Soc. Min., 8. 145, 1885 ; J. W. Gibbs, Scientific Papers, London,
1. 315, 1906 ; J. J. Thomson, Applications of Dynamics to Physics a^id Chemistry, London, 251,
1888 ; W. J. Jones, Zeit. phys. Chem., 82. 448, 1912 ; Ann. Physik, (4), 41. 441, 1913 ; W. J.
Jones and J. R. Partington, Phil. Mag., (6), 29. 35, 1915 ; W. Herz, Zeit. Elektrochem., 21. 373.
1915 ; 23. 23, 1917 ; E. Podszus, Zeit. phys. Chem., 92. 227, 1917 ; P. P. von Weimam, Grundziige
der Dispersoidchemie, Dresden, 119, 1911 ; E. Divers, Journ. Chem. Soc, 23. 359, 1870.
8 R. T. Hardmann and J. R. Partington, Journ. Chem. Soc, 99. 1769, 1911 ; D. Tyrer, ib.,
97. 631, 1778, 1910 ; Jour7i. Phys. Chem., 16. 69, 1912 ; S. Horiba, Mem. Coll. Science, Kyoto, 2.
], 1917 ; J. D. Hildebrand, Journ. Amer. Chem. Soc, 39. 1452, 2297, 1917 ; F. Dolezalek, Zeit.
phys. Chem., 64. 727, 1908 ; 71. 191, 1910 ; G. Bodlander, ib., 27. 55, 1898.
** E. Cohen, K. Inouye, and C. Euwen, Zeit. phys. Chem., 75. 257, 1910 ; E. Cohen and
L. R. Smnige,i6., 67. 432, 1909; H. F. Sill, Journ. Amer. Chem. Soc, 38. 2632, 1916; E. Cohen
and W. Schut, Piezochemie Kondensierter Systeme, Leipzig, 1919.
10 H. C. Sorby, Proc Roy. Soc, 12. 538, 1863; Phil. Mag., (4), 27. 145, 1864; F. Braun,
Wied. Ann., 30. 250, 1887 ; Zeit. phys. Chem., 1. 259, 1887 ; J. J. van Laar, ib., 15. 457, 1893 ;
18. 276, 1895 ; E. F. von Stackelberg, ib., 20. 337, 1896 ; C. M. Guldberg, Forh. Viden. Selskabet
Kristiania, 35, 1870 ; Ostwald's Klassiker, 139, 1903 ; J. J. Thomson, Applications of Dynamics
to Physics and Chemistry, London, 247, 1888 ; M. Planck, Wied. Ann., 32. 495, 1893 ; Vorlesungen
Hber Thermodynamik, Leipzig, 218, 1897; G. Robin, Bull. Soc Philomath., (7), 4. 24, 1879;
K. Moller, Pogg. Ann., 117. 386, 1862 ; P. A. Favre, Compt. Bend., 51. 827, 1027, 1860; E. Cohen
and A. L. T. Moesveld, Zeit. phys. Chem., 93. 385, 1919.
^^ J. Johnston, Journ. Franklin Inst., 183. 1, 1918.
12 G. Tammann, Ueber die Beziehungen zwischen den inner en Krdften und Eigenschaften der
Losungen, Hamburg, 1907.
13 T. W. Richards and R.C.Wells, Zeit. phys. Chem., 43. 455, 1903; L. C. de Coppet, ib.,22,
239, 1897; Ann. Chim. Phys., (4), 25. 539, 1872; H. Lowel, ib., (3), 49. 50, 1857; J. L. Gay
Lussac, ib., (2), 11. 312, 1819; Earl Berkeley, P/it7. Trans., 203. A, 209, 1904; W. A. Tilden
and W. A. Shenstone, ib., 175. 28, 1884; A. Etard, Compt. Revd., liS, S54, ISdl ; A. Smits
516
INOKGANIC AND THEORETICAL CHEMISTRY
and J. P. Wuite, Proc. Acad. Amsterdam, 12. 244, 1909; A. Smits, ih., 12. 227, 1909;
W. Meyerhoflfer, Journ,Phys. Chem., 8. 571, 1904; Zeit. phys. Chem.,^2, 501, 1903; W. Ostwald,
i6., 42. 503, 1903; J. L. Gay Lussac, Ann. Chim. Phys., (2), 11. 313, 1819; (2), 70. 402, 1839;
H. Kopp, Liebig's Ann., 34. 260, 1840; E. Demar9ay, Compt. Rend., 69. 1800, 1883; H. W. B.
Roozeboom, Die Heterogemn GleichgewicUe, Braunschweig, 1901-11; J. P. Wuite, Zeit. phys.
Chem., 86. 349, 1914.
1* H. C. Jones and J. S. Guy, The Absorption Spectra of Solutions, Washington, 1913:
J. E. L. Holmes and H. C. Jones, Journ. Amer. Chem. Soc, 38. 105, 1916; H. C. Jones,
E. J. Schaffer, and M. G. Paulus, Phys. Zeit., 15. 447, 1914.
§ 2. The Freezing of Solutions
Proof or disproof of the existence of many compounds must be sought in the physical
properties and in the behaviour of mixtures at different temperatures.- — L. W. Andrews
(1907).
The curve OB, the ^alt line. Fig. 3, represents the solubility of sodium chloride
at temperatures ranging from —23° to +40° ; the observation cannot be continued
below —23°, because the whole mass freezes at or above that temperature ; the
upward curve would probably stop only at the melting point of sodium chloride,
801°, if it were not for the volatilization of the
water. Hence, to determine the solubility, the
pressure would have to be very great at this high
temperature to prevent the water leaving the salt.
The freezing temperature of a solution is generally
lower than that of the pure solvent. More than a
century ago, C. Blagden (1788) i cited a number of
observations which led him to the belief that the
lowering of the freezing point is proportional to the
amount of substance in solution. In his own
words : The effect of a salt is to depress the
freezing point in the simple ratio of its proportion
to water. This generalization is sometimes called
Blagden's Law. The freezing point of an aqueous
solution of sodium chloride, that is, the temperature
at which ice begins to separate, is gradually
reduced by the continued addition of small quantities of sodium chloride, and
reaches its lowest value, —23°, when the solution has nearly 23"5, say 24, per cent.
of sodium chloride ; further additions of the salt raise the temperature at which
the soUd is deposited. SoHd sodium chloride, not ice, separates from the solution.
F. Guthrie's measurements (1875) ^ of the relation between the freezing point
and the concentration of aqueous solutions of sodium chloride are shown
graphically by the ice line, AO, Fig. 9, Cap. IX.
Impurities included in crystals. — It has long been known, even as far back as
Aristotle's day, that drinkable water could be obtained from frozen sea-water ; and
that if an aqueous solution of salt be gradually cooled, comparatively pure ice first
separates from the solution. The work of F. KiidorfE (1861) and of J. Fritzsche
(1863) 3 on the freezing of coloured solutions clearly established this fact. Thus,
magnesium cyanoplatinate forms a colourless solution from which colourless ice
separates, whereas, if solid magnesium cyanoplatinate separated, the colour would
be intensely red.
Faraday's experiment. — Water coloured with sulphindigotic acid is placed in a test-tube
and immersed in a freezing mixture ; the water on freezing near the walls of the tube
drives the colouring matter to the axis of the tube. The coloured liquid is poured away, and
when the cavity is rinsed out, a plug of transparent colourless ice is obtained. The trace of
salt which is generally found in the ice which separates from a salt solution is merely the
mother liquid or solute which is mechanically entangled in the crystals of ice.
FiQ 3. — -Solubility and Freezing
Curves of Sodium Chloride :
Water Solutions.
SOLUTIONS 517
The freezing curves of binary mixtures which do not form compounds.— A
solution of sodium chloride in water may be taken as an example. It may be
assumed that these substances are but partially soluble in the liquid state and
insoluble in the solid state. Imagine a 5 per cent, solution of salt subjected to a
gradually diminishing temperature. Start at 0°. When the temperature reaches,
say, —3*4° ice separates from the solution. The mother hquid remaining has
therefore more than 5 per cent, of salt in solution ; as the temperature falls, more
ice separates. The further concentration of the mother liquid and the separation
of ice continue until the mother liquid has about 23*6 per cent, of salt, when the
whole remaining liquid freezes en hloc at —23°. The solid now consists of crystals
of ice embedded in a matrix of ice and salt. Quite an analogous sequence of changes
occurs if solutions containing more than 23*6 per cent, of salt be gradually cooled.
This time, however, instead of pure ice, pure salt separates until the residual liquid
has 23'6 per cent, of salt. The whole solidifies en masse at —23°. If the cooling
solution has just 23*6 per cent, of salt, neither ice nor salt separates, until the tempera-
ture has fallen to —23°, when the whole freezes to a soUd mass. No other mixture
of water and salt freezes at a lower temperature than this. Hence a solution
containing 23*6 per cent, of salt is called a eutectic mixture or simply a eutectic ;
—23° is the eutectic temperature ; and the general phenomenon is called eutexia —
from the Greek ev, easily, and rrJAcw, I melt. Hence eutectic means " that which
is easily melted." The word cvttjkto^ was used by Aristotle (Problemata, 1. 50)
in the sense of easily soluble or digestible. The nature of the cooling liquid, or of
the solid of any given composition, is also shown by the shaded areas in Fig. 3.
F. Guthrie used to think that this mixture — water with 23'6 per cent, of salt- — corre-
sponded with the formation of a definite compound of sodium chloride and water —
NaC1.10H20 — stable only at low temperatures. Hence his designation cryohydrate
for the alleged compound. A. Ponsot (1896) * called the substance a cryosel. The
term eutectic mixture is preferred in place of cryohydrate or cryosel. The eutectic
temperature, —1*2°, represented at E, Fig. 2, corresponds with the eutectic
mixture of 3'85 per cent, of Na2S04 in 100 grams of solution when the decahydrate,
Na2S04.10H20, is the stable phase ; and with the heptahydrate, Na2S04.7H20,
the eutectic mixture contains 12"7 per cent, of Na2S04, and the eutectic temperature
is —3*55°. The same type of curve is illustrated in Fig. 4 — two separate branches
meeting in a eutectic is characteristic of binary metal alloys which form neither
compounds nor solid solutions — e.g. alloys of tin and bismuth, tin and zinc, cadmium
and zinc, lead and antimony, etc.
Cryohydrates and eutectics.— We now know that Guthrie's cryohydrates are
nothing but mechanical mixtures of ice and salt. The one is entangled with the
other. The more salient characteristics of eutectics are : (1) They have a lower
melting point than mixtures with a greater or less quantity of one component ;
(2) They freeze at a constant temperature ; and (3) They have a constant composi-
tion. With these qualities, it is easy to understand how eutectic mixtures were
mistaken for true chemical compounds. No matter what the original composition
of the salt solution, the last fraction to solidify always has the same composition ;
and a constant melting point. Both these qualities are often stated to be charac-
teristics of true chemical compounds. The inference that eutectics or cryohydrates
of sodium chloride and numerous other salts are not chemical compounds is based
on the following evidence : (1) The heterogeneous structure is frequently apparent
under the microscope. The crystals of ice can often be seen lying in a matrix of
salt, especially if a coloured salt like potassium permanganate, copper sulphate, or
potassium dichromate be employed. Indeed, the eutectic sometimes forms definite
patterns, with iron and carbon, the eutectic consists of alternate bands of the two
components — lamellar eutectic ; with copper and aluminium, one component forms
globules embedded in a matrix of the other — globular eutectic ; with copper and
antimony, the one component appears like small polyhedral crystals arranged in
matrix of the other — polygonal eutectic. (2) Unlike true crystalline compounds,
518
INORGANIC AND THEORETICAL CHEMISTRY
the cryohydrates are generally opaque and ill-defined. (3) Alcohol may dissolve
the solvent, leaving behind a network of salt. (4) There are no special signs of
chemical change during the formation of the cryohydrate. (5) The physical
properties of the cryohydrate — e.g. heats of solution, specific gravities — are a mean
of those of the corresponding constituents. This is not usually a characteristic
of chemical combination. (6) The ratio of salt to solvent is not always in molecular
proportions. The agreement in some cases is merely a coincidence. (7) The
composition of a cryohydrate is different when the solidification takes place under
different pressures. Hence, added A. Ponsot (1896), the eutectics or " cryohydrates
of F. Guthrie are not chemical compounds, they are
mechanical mixtures of pure ice and the solid salt. The
salt may be anhydrous like potassium nitrate, KNO3,
or hydrated like copper sulphate, CUSO4.5H2O."
There is another kind of eutectic which is formed in a
soHdified, congealed, or supercooled solution, at the time of
solidification. In the case of a solidified solution ; to
distinguish this mixture from a true eutectic it is some-
times called a eutectoid. The only difference between a
^ .. „ '"""^ - eutectoid and a eutectic is that the former is formed after ^
Fig. 4.— Coohng Curves of ^^d the latter at the time of solidification. In the case of a
Bmary Mixtures of ^ and solidified solution of carbon in iron, the eutectic— called
B with a Eutectic. pearlite— is formed after the metal has become solid. The con-
glomerate or impure ice formed by the freezing of a solution
with less than 23*o per cent, of sodium chloride is called a hypo-eutectic because it contains
less than the eutectic percentage of the salt ; likewise also the conglomerate formed by the
freezing of a solution with more than 23-5 per cent, of salt is called a hyper-eutectic because
it contains more salt than the eutectic proportion. The hypo-eutectic is a mixture of the
eutectic with an excess of the frozen solvent ; the hyper-eutectic is a mixture of the
eutectic with an excess of solute.
Cooling and heating curves. — If a thermometer or thermocouple be placed
in a cooling solution, and the time be plotted against the temperature, three
main types of cooling curve may be obtained : (1) Pure liquids show a break
in the continuity of the curve at the freezing point
corresponding with the evolution of heat — latent
heat of solidification — middle curve. Fig. 4. (2)
Mixed liquids— hinsLTy alloys and solutions — show two
breaks in the continuity of the curve : (a) when the
solvent begins to separate and there is a change in
the direction of the cooling curve, B, and (b) when
the eutectic freezes en masse, Fig. 4. (3) Eutectic
mixtures have a cooling curve with one break corre-
sponding with the evolution of heat when the whole
mass solidifies (20 per cent. A, Fig. 4). These phe-
nomena are reversed when the corresponding solids
are heated. The observation of the heat changes
which occur when a metal, alloy, or other substance
is cooled from an elevated temperature or raised to a higher temperature is called
thermal analysis, and it has played a great part in studying the constitution of
metals and alloy s.^
The freezing curves of binary mixtures which form compounds. — T. P. van der
Goot (1911)6 found that when a mixture of sulphuryl chloride, SO2CI2, melting at
— 54"1°, andof sulphur dioxide, SO2, melting at — 75"1°, be treated as in the case of
the above mixture of sodium chloride and water, a eutectic melting at —87 "3° is
obtained, while a mixture of sulphuryl chloride and chlorine — CI2, melting at
— l(X)-9° — furnishes a eutectic melting at — 107"5°. A mixture of sulphur dioxide
and chlorine furnishes two eutectics melting at —83" 7° and —107 "5° respectively
with a maximum point at — 541°, corresponding" with the formation of the compound
Fig. 5. — Freezing Curve of Mix-
tures of Chlorine and Sulphuric
Dioxide showing Two Eutec-
tics.
SOLUTIONS . 519
SO2CI2 from the components in question, S02+Cl2=S02Cl2, and the eutectic
at — 107*5° is characteristic of a mixture of sulphuryl chloride and chlorine. These
results are summarized in Fig. 5.
The raising of the melting or freezing point of one substance by the addition of
another often indicates that a compound is being formed. The freezing point of
zinc is depressed by addition of tin, bismuth, thalliimi, cadmium, lead, antimony,
magnesium, or aluminium ; and elevated by additions of silver, copper, gold, or
platinum. In general, when a pair of metals, minerals, or salts furnish a freezing
curve with a number of branches dependent on the number n of compounds formed,
the curve will have 2w+l branches and there will be n-\-l eutectics. E.g. alloys
of copper and antimony ; nickel and tin ; silver and aluminium ; zinc and anti-
mony ; lead and copper ; lead and aluminium ; bismuth and copper ; aluminium
and gold ; aqueous solutions of ammonia, nitric, hydrochloric, or sulphuiic acid ;
etc. If one or more of the compounds forms a solid solution with one of the other
metals, this would modify the character of the curve as indicated in the first type
of freezing curve.
J. P. Cooke (1855) 7 and N. S. Kurnakoif (1912) have concluded from a study
of alloys — zinc and antimony, in the former case ; and thallium and bismuth in
the latter— that there is a class of indefinite compounds which are not described
by the laws of definite and multiple proportions, and they support the view of
C. L. BerthoUet in his controversy with J. L. Proust :
The result of the different circumstances which modify chemical action is sometimes a
combination whose proportions are constant, and sometimes, on the contrary, the pro-
portions of the combinations which are formed are not fixed, but vary according to the
conditions vmder which they are formed.
N. S. Kurnakoff has pointed out that the composition of a compound is deter-
mined by the position of singular points on the curve representing physical
properties ; these points most frequently correspond with simple formulae, but
there are exceptions. For example, a maximum occurs on the freezing-point
curve of mixtures of thallium and bismuth with a mixture containing 62'8 atomic
per cent, of bismuth, but solid solutions at ordinary temperatures extend from
55 to 64 atomic per cent, of bismuth, and there is a cusp in the electrical conductivity
curve with 64 atomic per cent, of bismuth. N. S. Kurnakoff recommends naming
compounds whose composition does not change with changes in the equilibrium
conditions of a system, daltonides ; and compounds whose composition varies with
a variation in the conditions under which they are formed, berthollides. He
considers that solid solutions, brasses and bronzes, zeolites, metal ammines, etc.,
are representative berthollides. This is a direct attack on constant composition
as a test for chemical action, and if it were accepted as an arbitrary definition,
convention would return a different answer from that previously obtained for the
question : Are solutions chemical compounds ?
The freezing curves of binary mixtures which either form or do not form
continuous series of mixed crystals. — When two substances are reciprocally soluble
in all proportions and solidify to form solid homogeneous solutions^ — also called
mixed crystals — but not chemical compounds, a continuous curve will connect
the freezing points of the pure components. The properties of the mixtures will
vary in a continuous manner from one end of the series to the other. The freezing
points of all possible mixtures will be represented (i) by a straight (or almost
straight) line between the freezing points of the pure components as is the case
with mixtures of albite and anorthite. The pyrophosphates of manganese and
magnesium are miscible in all proportions and belong to Roozeboom's type I :
Manganese pyrophosphate . 100 75 50 25 0 per cent.
Magnesium pyrophosphate . 0 25 50 75 100 „
Melting point . . . 1196° 1242° 1286° 1340° 1383°
Refractive index (mean) . 1-70 1-67 1*65 1-63 1*60
520
INOKGANIC AND THEORETICAL CHEMISTRY
Other examples are lead bromide and iodide ; lead and stannous chlorides ; alloys
of gold and silver ; gold and platinum ; copper and nickel ; palladium and silver ;
palladium and gold ; palladium and copper ; and usually, with pairs of metals of
high melting points, but the further apart the melting points of the two com-
ponents the less the probability of this linear relation, (ii) By a continuous curve
which rises through a maximum, as is the case, for instance, with mixtures of
organic compoimds which show optical isomerism^ — e.g. d- and ?-carvoxime (J. H.
Adriani, 1900),® — but has not been otherwise verified, (iii) By a continuous
curve which drops down through a minimum as is the case with mixtures of
silver and cuprous sulphides — ^Ag2S— CugS ; copper and manganese sulphites,
CaSOa and MnSOs ; iron and vanadium ; manganese and nickel ; and copper
and gold. Mercuric bromide melts at 23'5° and the iodide melts at 255'4° ; the
system HgBr2— Hgl2 has a minimum point with 59 per cent, gram-molecules of
the bromide (41 of the iodide) and melts at 21 6*1°,
The dotted line. Fig. 6, shows the solidus curve —
W. Reinders (1899).
Consider the first case. The freezing and melting
point curves do not coincide, so that these two
curves divide the region, Fig. 6, into three parts ;
above the freezing-point curve AlbB — the liquidus
curve of H. B. Roozeboom — all possible mixtures
of platinum and gold are completely liquid ;
below the melting point curve AasB — the solidus
curve of H. B. Roozeboom — all possible mixtures
Platinum and Gold— Rooze- uquidus and soudus curves, the mixture is partly
boom's Type I. liquid and partly solid. If a molten mixture of 50
gram-atoms of each of the two metals at a tem-
perature represented by the point I be allowed to cool, the temperature of
the system will be represented by a point travelling down llss, and mixed
crystals, that is a solid solution of the two metals, will begin to separate
when the temperature drops to I. It is very unlikely that the reciprocal
solubility of the two metals will be equal in the solid and liquid states ; in most
cases, the solubility will be different, and be more complete in the liquid state so
that mixed crystals of an alloy richer in the less fusible metal and with a composition
represented by the point a will separate, and the mother liquid will have a composi-
1200*
25 50 75
Melting and Freezing Point Curves of
mixtures of Platinum & Cold.
0 50% Hg 1^
100 50% Hg Bp2
Fig. 7. — Roozeboom's Type II of
mixed Crystals of Carvoxime.
Fig. 8. — Roozeboom's Type III
of mixed Crystals.
tion richer in the more fusible metal and be represented by the point 6. As the
freezing continues, the composition of the mixed crystals which separate will be
represented by a point travelling along the line as, and the composition of the
mother liquid by a point travelling along the line Ih. The process of diffusion,
however, will tend to make the composition of the solid solution more and more
like that of the mother liquid. With complete diffusion, the solid solution at the
end of the process will be homogeneous in composition ; but since diffusion in a
solid is a very slow operation, in practice, diffusion will be incomplete, and the mass
will be more or less heterogeneous. If there were no diffusion, the final mass would
have a composition ranging from that represented by the point a (in the centre of
SOLUTIONS
521
f//^;Mixed.Crystak
100% A Composition 100% B
Fig. 9." — Diagrammatic —
Type IV.
the mass) to h (on the outside). By reheating the alloy (below its melting point)
diffusion may take place, and such a process — called annealing — is necessary to
make the solid mass approximate more and more to the homogeneous con-
dition. The alloy solidifies completely when the temperature has fallen to 5,
and is partly liquid and partly solid in the tempera-
ture interval Is. Hence, says H. M. Howe (1916),
the liquidus traces the history of the liquid or
mother liquor ; the solidus, the history of the
frozen or solidified part during freezing and melting.
In addition to the three types of mixed crystal
formation just considered, there are two others in
which the two substances are completely soluble
in the liquid state, but in one tjrpe (IV), the
liquidus curve shows a transition point, 0, Fig. 9,
and the solidus is compounded of two disconnected
curves Aa and Bh with a hiatus ah. The range of
composition of mixed crystals a and j3 is respec-
tively represented by the abscissae of Aa and Bh.
A magma of composition corresponding with the transition point 0 is in
equilibrium with the mixed crystals a or j8. Mercury and cadmium alloys
investigated by H. C. Bijl (1902) illustrate the type — mercury melts at — 38'8°,
cadmium at 320*8°. The transition point corresponding with 61 per cent, of
cadmium occurs at 188°, a corresponds with 61*7 per cent, of cadmium and h
with 65*2° per cent. Fig. 9 is based on this
example. Other examples are mixtures of en-
statite and diopside ; sodium and silver nitrates ;
and, according to Gr. Scarpa (1915), mixtures of
potassium hydroxide and chloride.
In the next type (V), the liquidus consists of
two curves meeting in a eutectic E, and the solidus
likewise consists of the two dotted curves A and
B, Fig. 10. Mixed crystals can exist only in the
range indicated. Fig. 10 is based on C.[Sandonnini's
work (1911) on mixtures of silver chloride melting
at 455°, and cuprous chloride melting at 422° ; the
eutectic is at 260°. Other examples are potassium ^
and thallium nitrates, orthoclase and albite, cuprous ^^^ '
and sodium chlorides, thaUium chloride and iodide, calcium and lithium .silicates
— CaSiOs and Li2Si03 ; mercuric chloride and iodide ; aluminium and zinc, gold
and nickel, etc.
To summarize the five types of mixed crystal formation, in which chemical
compounds are not formed :
Liquid
~ Solid ■ -'^-It'^"
Eutectic of Two
KJndsofCrystals. J • ^
100% A Composition 100^ B
Fig. 10 — Diagrammatic —
Liquid state.
Completely
soluble
Solid state.
Completely soluble, continuous series of
mixed crystals
Liquidus. Figure,
no max. or min. 6
maximum 7
minimum 8
Partially soluble limited range mixed ( transition point 9
crystals \ eutectic 10
Modifications of the two latter types are exhibited when the substances are only
partially soluble in the liquid state, and when they are partially or wholly insoluble
in the solid state. The case of salt and water, Fig. 3, illustrates the former ; and
copper and cuprous sulphide, iron and ferrous sulphide, and zinc and lead illustrate
the latter. Very complex curves may be obtained as a result of complications
introduced by the formation of chemical compounds which may or may not form
mixed crystals with one another or with the pure components ; the chemical com-
pounds formed may dissociate below the freezing temperature ; the solubilities in
522 INORGANIC AND THEORETICAL CHEMISTRY
the liquids and solid states may vary ; and transition points may appear in the
cooling solid.
References.
1 C. Blagden, Phil. Trans,, 78. 143, 277, 311, 1788; L. C. de Coppet, Ann. Chim. Phys.,
(4), 23. 366, 1871 ; (4), 25. 502, 1872 ; (4), 26. 98. 1872.
2 F. Guthrie, Phil. Mag., (4), 49. 1, 49, 354, 446, 1876; (5), 1. 1875; (5), 1. 49, 1876;
(5), 2. 211, 1876.
3 F. Riidorff, Pogg. Ann., 114. 63, 1861 ; 116. 55, 1862 ; 145. 599, 1871 ; J. Fritzsche, BuU.
Acad. St. Petersburg, 6. 385, 495, 1863.
* A. Ponsot, Ann. Chim. Phys., (7), 10. 79, 1897 ; Compt. Rend., 129. 98, 1899.
'^ R. Kremann, Ueber die Anwendung der thermische Analyse zum Nachweis chemischer Ver-
bindungen, Stuttgart, 1909.
« T. P. van der Goot, Zeit. phys. Chem., 84. 419, 1913.
' J. C. Cooke, Amer. Journ. Science, (2), 18. 229, 1854 ; (2), 20. 222, 1855 ; (2), 30. 194, 1860 ;
N. S. Kumakoflf, Bull. Acad. St. Petersburg, 321, 1914 ; N. S. Kumakoff and S. F. Schemtchuschny,
Journ. Russian Phys. Chem. Soc, 44. 1964, 1912 ; N. S. Kumakoflf, Zeit. phys. Chem., 88. 109,
1914.
8 J. H. Adriani, Zeit. phys. Chem., 33. 453, 1900; H. C. Bijl, ib., 41. 461, 1902; H. W. B.
Roozeboom, i6.,34. 451, 1900; Die heterogenen Gleichzewichte vom Standpunkte der Phasenlehre,
Braunschweig, 1901 ; W. Reinders, Proc. Acad. Amsterdam, 2. 146, 1899; G. Scarpa, Atti Accad.
Lincei, (6), 24. i, 738, 965, 1915; C. Sondonnmi, ib., (5), 20. i, 457, 1911; H. M. Howe, The
Metallography of Steel and Cast Iron, New York, 1916.
§ 3. The Solubility of Liquids in Liquids
Mixtures of liquids like benzene and water or carbon disulphide and water
are mutually insoluble, and they are said to be immiscible or non-miscihle. Each
liquid then behaves as if it were present alone, the vapour pressure of the mixture
corresponds with the sum of the saturation pressure of each. Each of the two
liquids may be partially soluble in the other as in the case of ether and water,
phenol and water, or aniline and water, and the liquids are then said to be partially
miscible ; or the two liquids may be miscible in all proportions — completely miscihle —
as in the case of alcohol and water. With partially miscible liquids, the composition
of each solution, per unit volume, is independent of the masses of the two components
— ^provided both are present. Thus, when ether is gradually added to water, a
solution of ether in water is formed, which becomes more and more concentrated.
When the aqueous solution is saturated any further addition of ether forms a
saturated solution of water in ether, and with further additions of ether, the ethereal
solution remains saturated with water until, finally, the ether has dissolved all the
water. Any further addition of ether simply dilutes the ethereal solution of water.
Consequently, partially miscible liquids have a sharply defined limiting surface
furnishing a heterogeneous two-phase solution ; with completely miscible liquids,
a homogeneous one-phase solution is formed.
When two liquid phases are present in a system — e.g. ether and water — the system
has the following characteristic properties : (1) The composition of each phase is
constant at any assigned temperature and pressure, and is independent of the
relative amounts of the two phases, and independent of the mode of preparation.
(2) The composition of each phase changes with temperature changes. (3) The
vapour pressure of both liquid phases are equal both as to the total pressure and
to the pressure of each constituent. If in the system ether and water, the water
is withdrawn continuously, the pressure will remain constant so long as both liquid
phases are present. When the water disappears, the pressure of the water vapour
begins to fall, and becomes zero simultaneously with the disappearance of the
water phase, and the residue will be a saturated solution of water in ether. At
any assigned temperature, the composition of the liquid will have one constant
value, but it will vary with variations of temperature.
In 1835, M. L. Frankenheim i noticed that while a variety of creosote and water
SOLUTIONS
523
are only partially miscible at 0°, at 22° the two liquids are mutually soluble in all
proportions ; and in 1857, D. Absaheff investigated the mutual solubility of a
number of pairs of liquids, and showed that the composition of partially miscible
liquids changes with a rise of temperature, so that the composition of the two
layers becomes more and more nearly alike until a point is reached at which the
liquids become completely miscible. A few exceptions have been encountered.
For example, for mixtures of water and ether E. A. Klobbie found that the
solubility of ether in water decreases with rising temperature, while the solubility
of water in ether increases, as indicated in Table I.
Table I. — Mutual Solubilities of Ether and Water.
Temperature.
Grams of ether per 100
grms. of aqueous solution.
Grams of water per 100
grms. of ethereal solution.
-4°
0°
10°
20°
30°
12-63
1217
9-02
6-48
5-04
0-92
100
112
1-23
1-33
This was further investigated by W. Alexejeff 2 between 1876 and 1885 ;
F. Guthrie in 1884, F. A. H. Schreinmakers in 1897, etc. W. Spring and S. RomanofE
observed that certain pairs of molten metals presented a similar phenomenon to
that observed by M. L. Frankenheim. For instance, mixtures of zinc and bismuth
which, below 800°, form two layers, like mixtures of ether and water at ordinary
temperatures, but are completely miscible above 900°. W. AlexejefE's data for the
mutual solubilities of aniline and water are indicated in Table II.
Table II. — Mutual Solubilities of Aniline and Water.
Temperature.
Grams of aniline per 100
grms. of aqueous solution.
20°
3-1
40°
3-3
60°
3-8
80°
5-5
100°
7-2
120°
9-1
140°
13-5
160°
24-9
167°
48-6
Grams of water per 100
grms. of aniline solution.
60
5-3
5-8
6-5
8-4
11-9
16-9
28-8
51-4
The temperature at which the two liquids become mutually soluble in all
proportions is called the critical solution temperature, and the corresponding
concentration of the solution, the critical concentration. Thus the critical solution
temperature of aniline and water is 167° ; and the critical concentration is a liquid
containing 48'6 grms. of water and 51*4 grms. of aniline. The analogy between
the critical solution temperature and the critical state of gases was pointed out by
0. Lehmann in 1888,3 and by 0. Masson in 1891, and it is illustrated by the curves.
Fig. 11. There has been some discussion as to whether the observed data lie on
one continuous curve, or form two curves which intersect one another at the critical
point.4
The determination of the critical solution temperature is effected by plotting
the solubility curves of the two liquids— J in B and J5 in ^ ; or by heating a mixture
of the two liquids until a homogeneous solution is obtained, and noting the tempera-
ture at which a turbidity appears. The blue opalescence is due to the separation
524
INORGANIC AND THEORETICAL CHEMISTRY
of minute drops of liquid ; and their appearance is evidence that the liquid is
saturated with respect to large drops, since E. Warburg ^ has shown that small
drops of liquid are more readily dissolved than large ones, just as Lord Kelvin has
shown that small drops of liquid appearing in vapour are absorbed by the larger
one owing to the greater vapour pressure of the former.
According to C. S. Hudson ^ nicotine and water are miscible in all proportions
at ordinary temperatures, but at temperatures exceeding 60°, the solution becomes
turbid owing to incomplete miscibihty ; at 210° the two liquids again become
completely miscible. The solubility curves of the binary system nicotine and
water thus appear to form a closed curve with an upper and a lower temperature
of complete miscibility. A similar closed curve has been observed with secondary
butyl alcohol and water. According to W. Dolgolenko, the lower temperature
limit of complete miscibility in the latter case is due to the presence of traces of
tertiary alcohol as an impurity ; if these be eliminated, the lower critical temperature
does not occur. The phenomenon of a closed solubility curve is taken to be
characteristic of a ternary, not a binary,
system. In the case of nicotine and water,
the third constituent is supposed to be
a hydrate of nicotine — such a system is
called a pseudoternary system by H. W. B.
Roozeboom.7 The hydration of nicotine
is evidenced by a considerable contraction
and evolution of heat when nicotine is
mixed with water.
The kinetic theory of the critical
temperatures, and the critical solution
temperature. — ^As the temperature of a
liquid is gradually raised, the average
kinetic energy of the molecules increases,
and a greater and greater proportion
of the molecules overcomes the cohesive
forces holding the particles of liquid
200*
160"-
120-
80
40
Critical SolutionTemperahirt
^--^-^^ 167" 1
f /^
^
i
c^\
-Is /'S'
^-K
/i
«\
■$.\
/ ^
0) \
/^
•sX
- /^
-i \
M
^ \
1^
- \
P
l\
"/■i
\ ■
N
Per cent Aniline
U 1
50
too
Fig. 11. — Reciprocal Solubility Curves of together, and escapes as vapour. The
Aniline and Water. thermal expansion of the liquid is due
to the increasing velocity of the vibrat-
ing molecules, which makes them behave as if they were actually repelled
from one another by a force which increases in magnitude as the tempera-
ture is raised. Consequently, as the temperature is raised, the repellent forces
become stronger and the cohesive forces weaker. When the two forces are just
balanced, the liquid can expand indefinitely and the surface separating liquid
and vapour vanishes. The properties of liquid and gas are then the same. This
is the critical temperature ; and the critical pressure has been taken to represent
the magnitude of the quasi-repulsive force between the molecules at the critical
temperature. When two liquids are in contact, molecules pass from one to the
other by a process which is analogous with vaporization, except that the vibratory
motions are slower owing to frictional forces. When the average speed of the mole-
cules is great enough just to balance the physical cohesive forces between the molecules
of the liquid, the surface of separation between the two liquids vanishes, and the
system becomes homogeneous. Hence, in both cases, the critical temperature is
due to the balancing of the cohesive forces of the liquid by the quasi-repulsive
force due to the increasing velocity of its particles.^
Assuming that the critical pressure of a gas is that at which the repellent forces
between the molecules are just equal to the forces of attraction at the critical
temperature, E. C. Bingham argues that the critical pressure of a gas is a measure
of the molecular attraction. Again, since Avogadro's rule applies very well for
gases whose particles are so far apart that the time during which they are within
SOLUTIONS
525
the range of one another's attraction is negligible in comparison with the time
the particles are independent. The closer the particles are together the more does
the intermolecular attraction predominate. If the deviations of a gas from
Avogadro's rule are a measure of molecular attraction, the attractive forces will
be inversely proportional to the molecular volume, and consequently, the pro-
duct of the molecular volume MjD and critical pressure po should be a constant.
E. C. Bingham has worked out Table III. in confirmation of these hypotheses.
lABLE III
• — Intermolecular Attraction and Solubility.
M
D
MID
Pe
pMID
b
Pjb
HaO
18
1
18
197
3500
0-00160
0-295
NH3 .
17
0-6089
28
114
3200
0-00162
0-285
HgS .
34
0-91
37-4
90
3370
0-00189
0-170
HCl .
36-4
0-835
43-6
86
3750
0-00173
0149
SO2 .
64-1
1-3769
46-5
78-9
3670
0-00249
0196
N2O .
44
0-758
52-6
75
3940
0-00189
0-141
CO2 .
44
0-8267
53-2
73
3800
0-0019
0-138
CSa .
76-1
1-2922
58-9
75
4400
0-0033
0-247
Cya . .
521
0-866
60-2
61-7
3710
0-0029
0-179
SUCI4.
260-8
2-28
114-4
36-9
4220
000733
0-271
GeCl^
214-3
1-887
113-6
38
4300
0-00663
0-255
(C^B.^)^0 .
76-1
0-7191
103-4
37-8
3890
0-00563
0-202
CHCI3
119-4
1-5039
79-4
54-8
4350
0-00445
0-244
CaHgOH .
46
0-7942
57-9
62-8
3640
0-00377
0-237
CCI4 .
153-8
1-5947
97-1
45
4370
0-00434
0-195
Roughly, the product of the critical pressure and molecular volume is a constant.
The molecular volumes should be observed at the critical temperature, but very
few data are available. The magnitude b of J, D. van der Waal's equation is
proportional to the volimie of the particles, and it has been compared with the
quotient M/D. The agreement is not so good on account of large experimental
errors.
Substances with a large molecular volume have a small intermolecular attraction
— e.g. ether and carbon tetrachloride — and hence such substances are more hkely
to be miscible than if one is replaced by a substance like water with a small mole-
cular volume and a large intermolecular attraction. Two substances with a
small intermolecular attraction mix readily ; no two substances are miscible when
their molecular volumes difEer very much, since molecular volumes are inversely
as molecular attractions. V. Rothmund gives Table IV. indicating the order of
solubilitv of a number of substances in water :
Table IV.
-The Relation between Molecular Volume, Dielectbic Constant, and
Solubility.
Molecular volumes.
Dielectric constants.
Water. H2O
18
81
Formic acid, H.COOH .
37-7
67
Methyl alcohol, CH3OH
39-5
32-5
Acetone, CH3CO.CH3
72-6
20-7
Aldehyde, CH3.COH
55-6
21-1
Propionitrile, CgHgCN .
68-8
27-7
Phenol, CeHgOH .
90
9-7
Benzaldehyde, CgHg.COH
100
18
Ether, (C2H6)20 .
103
4-36
Carbon tetrachloride, CCI4
97-1
1-00
Carbon disulphide, CSg •
68-9
2-64
Hexane, CgH^
129-9
1-86
526 INORGANIC AND THEORETICAL CHEMISTRY
Pairs of sub3tanoes near each other in the list are completely miscible, and the
mutual solubility is less, the greater the distance of the two liquids apart. With
three or four exceptions, the order of the molecular volumes (E. C. Bingham) and
the order of the dielectric constants (V. Rothmund) agree fairly well with the order
of solubility. So also the work of A. ChristofE shows that the solubility of a gas
increases as the cohesion (or surface tension) of the solvent decreases, and that
with mixtures of sulphuric acid and water, the solubility of the gases reaches a
minimum as the surface tension attains a maximum. I. Traube has made analogous
observations by showing that the more a solute decreases (or increases) the surface
tension of a solvent, the less (or greater) the force with which it is attracted by the
liquid.
The analogy between the critical solution temperature and the critical tempera-
ture led W. Alexejefi to try if L. Cailletet and E. Mathias' rule for the relation
between temperature and the density of a liquid and of its saturated vapour — a
straight line is obtained by joining the mid-points of the ordinates lying between
two branches of the curve — ^is applicable to the mid-points of the line between the
two branches of the curve representing the solubilities of one liquid in the other,
and he did obtain an approximation to a straight line. W. Rothmund attributed
the irregular deviations from a straight line to experimental errors. In cases where
accurate data are available the law of rectilinear diameters applies for pairs of
liquids with an upper critical solution temperature, so that if Cj and C2 be the two
concentrations at the same temperature 6, and A and B are constants, J(Ci-|-C2)
=sA-\-Bdj and the concentrations at any temperature 6 can be calculated from
one another when the constants A and B are known. When there is a lower critical
solution temperature, the line is not always straight — but then the lower solution
temperature is regarded as a proof that there is some disturbing factor such as the
presence of a third component as an impurity or as a compound of the two liquids,
or a polymeric modification of one or both of the liquids.
J. Holmes ^ has advanced the hypothesis that the intermiscibility of liquids is
a function of the molecular volume which is independent of chemical constitution,
and which therefore ofEers a means of ascertaining the molecular weights of liquids
comparable perhaps with the methods which are based on Avogadro's hypothesis
for the gaseous state. The molecular volume in question, referred to water unity,
is not that deduced from the regular formula. Carbon disulphide and ethyl alcohol,
for instance, have the same molecular volume from this point of view, but one
liquid is insoluble in water while the other liquid is miscible in all proportions. If
allowance be made for the degree of association of the molecules in the liquid state,
it is found that an increase in molecular volume is attended hy a decrease in miscihility
with water. Assuming that the liquid molecule consists of a nucleus of great density
and an outer envelope or sphere of influence, then in the absence of chemical change,
the more nearly equal the radii of the molecular spheres, the greater the miscibility
of the two substances. With spheres of different sizes, then, as the ratio of the
radius of the larger to that of the smaller sphere increases, the dimensions of the
interspaces also necessarily change, and a point is reached when the close packing
of spheres is a maximum. In an equimolecular mixture this occurs when the
radii are as 1'618 to 1. It is suggested that when these conditions are fulfilled the
liquids are on the border line between complete as compared with partial miscibility,
and that so long as this or a smaller ratio exists, the liquids will be mutually miscible.
When, however, the ratio becomes greater than this value, a complex is formed
which will allow of the same close-packing, but results in the formation of two
layers of liquid, in each of which the distribution of molecules depends on the
further change in this ratio until it reaches 2"414, when, the smaller spheres being
able to pass completely through the interspaces between the larger spheres, the
liquids should be now wholly immiscible.
SOLUTIONS 527
Bbfbbences.
1 M. L. Frankenheim, Cohesion, Breslau, 199, 1835 ; D. AbsahefF, Bull. Soc. Imp., Moscou,
30. ii, 271, 1857.
2 W. Alexejeff, Ber., 8. 275, 1875 ; 9. 442, 1810, 1876 ; 10. 410, 708, 1877 ; 12. 2172, 1879 ;
15. 352, 1882 ; 16. 2273, 1883 ; 17. 38, 193, 599, 1884 ; Journ. prakt. Chem., (2), 25. 518, 1882 ;
Wied. Ann., 28. 305, 1886 ; F. Guthrie, Phil. Mag., (5), 18. 29, 499, 1884 ; F. A. H. Schreinmakers,
Zeit. phys. Chem., 23. 417, 1897 ; E. A. Klobbie, ib., 24. 615, 1897 ; V. Rothmund, ib., 26. 433,
492, 1897 ; W. Spring and S. Romanoff, Zeit. anorg. Chem., 13. 29, 1897.
3 0. Lehmann, Molekularphysik, Leipzig, 2. 208, 1888 ; 0. Masson, Nature, 43. 480, 689,
1891 ; Zeit. phys. Chem., 7. 500, 1891 ; W. Nemst, ib., 4. 150, 1889 ; W. Ostwald, Lehrbuch der
allgemeinen Chemie, Leipzig, 2. ii, 341, 1902.
* W. D. Bancroft, Journ. phys. Chem., 1. 143, 1896 ; A. Aignan and E. Dugas, Compt. Bend.,
123. 498, 1897 ; V. Rothmund, Zeit. phys. Chem., 26. 433, 1898.
6 E. Warburg, Wied. Ann., 28. 394, 1886 ; Lord Kelvin, Proc. Boy. Soc. Edin., 7. 63, 1870.
« C. S. Hudson, Zeit. phys. Chem., 47. 113, 1904 ; W. Dolgolenko, ib., 62. 499, 1908.
' H. W. B. Roozeboom, Die heterogenen Gleichgewichte, Braunschweig, 1901.
8 E. C. Bingham, Amer. Chem. Journ., 37. 549, 1907 ; 38. 91, 1907; A. Christoff, Zeit. phys.
Chem., 55. 622, 1906; V. Rothmund, ib., 26. 433, 1898; S. C. Bradford, Phil. Mag., (6), 38.
696, 1919.
9 J. Holmes, Journ. Chem. Soc, 89. 117, 1906 ; 103. 2147, 1913 ; 107. 1471, 1915 ; J. Holmes
and P. J. Sageman, ib., 91. 1608, 1907 ; 95. 1919, 1909.
§ 4. The Solubility o£ Gases in Liguids— Henry's Law
The absorption of gases by liquids began to attract attention towards the end of
the seventeenth century; andlater the subject was investigated by J. Priestley and
H. Cavendish. The solubility of gases in liquids depends upon the nature of the gas
and of the solvent, as well as upon the temperature and pressure of the system.
No common solubility has been observed, and we have no generalization of such
wide apphcabiUty as Boyle's and Charles' laws. Some gases are only sHghtly
soluble in water, others dissolve very copiously. One volume of water, at normal
temperature and pressure, will dissolve 1200 volumes of ammonia, and but 0"210
volume of hydrogen.
The changes in the volume of the solvent. — There is usually a small expansion
when gases are dissolved in water. The phenomenon was noticed by T. Bergmann ^
near the end of the eighteenth century ; and the change was investigated by a few
workers during subsequent frears. Thus, T. Thomson, J. J. Mackenzie, and E. L.
Nichols (1878) found that the expansion which occurred when water is saturated
with carbon dioxide is almost proportional to the absorption coefficient at the
specified temperature. K. Angstrom (1882) found the increments in the volume of
a liquid following the absorption of its own volume of gas :
Nj Air. CO O2 Ha COj
Volume increment . 0*00145 0-00140 0*00127 0*00115 0*000106 0*00125
The general results show that the specific gravity of the absorbed gas is nearly
proportional to its specific gravity in the free state ; and the less compressible the
gas the greater the volume increment when it is dissolved in water. Concentrated
aqueous solution of some of .the readily absorbed gases — e.g. hydrogen chloride- —
contract when diluted with water, showing that the volume increment on absorp-
tion cannot be proportional to the quantity of gas absorbed.
Two systems have been much employed for representing the solubility of gases in liquids.
R. Bunsen (1857) 2 applied the term absorption coefficient, ^, to the volume of a gas
reduced to 0° and 760 mm. which is absorbed by unit volume of liquid when the pressure of
the overlying gas on the liquid is 760 mm. and the temperature, at the time of absorption,
is d°. For instance, suppose 20 c.c. of a dry gas at 10° be confined over mercury in a tube
such that when the barometric pressure is 750 mm. the upper level of the mercury is 200 mm.
above the lower level in the trough ; further, let 2*5 c.c. of water be introduced into the tube,
and, after shaking, absorption reduces the volume of the gas to 18 c.c. when the barometer
remains at 750 mm., the height of the column of mercury in the tube is 204 mm. The
voKime of dry gas before absorption at 0" and 760 mm. is 20 x 550/760 x 273/283 = 13*96 c.c.
528 INORGANIC AND THEORETICAL CHEMISTRY
The pressure of aqueous vapour at 10*5° is 9-5 mm. The volume of dry gas at a pressure of
536-5 mm. remaining after absorption by the water and reduced to 0° and 760 mm. is
18 X 536-5/760 X 273/283-5 = 12-4 c.c. Hence, the volume of gas measured at 0° and 760
mm. which is absorbed by 2-5 c.c. of water under a pressure of 536-5 mm. is 13-96 less 12-24,
that is 1-72 c.c. Had the pressure at the time of absorption been 760 mm. then, by Hem-y's
law, the 2-5 c.c. of water would have absorbed 1-72x760/536-5 = 2-436 c.c. Hence,
1 c.c. of water absorbs 2-436/2-5=0*974 c.c. of gas; or ^=0*974. The effect of the slight
pressure due to the height of the column of water on the volume of the gas has been here
neglected. To put this result in general terms, if v^ be the initial volume of dry gas at 0°
and 760 mm. ; v^, the volume of the dry gas remaining after absorption by V volumes of
liquid ; and p, the observed pressure of the gas remaining after the absorption, the
Coefficient of absorption, j3= "„ ' . —
W. Ostwald (1888) represented the solubility of a gas, S, as the volume of gas absorbed
by unit volume of the liquid all at the temperature of the experiment. Consequently, if
Si volumes of gas are absorbed at a pressure p, and temperature 6, by V volumes of liquid,
then the solubility S =SJ F, and
Sf 1 Sf 27^
Coefficient of absorption, ^= y' \\ a J ^^ ^—'^ ' rp
where a denotes the coefficient of thermal expansion of the gas, approximately 0*00367 ;
T the absolute temperature ; and V is usually taken to be 100 c.c. Hence, the solubility,
Sy of a gas differs from Bunsen's absorption coefficient, ^, in that in the former case the
volume of the dissolved gas is not reduced to 0° and 760 mm. Again, \-\-ad times the
3oefficient of absorption ^ represents the solubility S of the gas, or, ^{\-\-ad)=S,
and ^=S/{l-\-ad), since both the solubility and the absorption coefficient are indepen-
dent of the pressure. Sometimes the solubility of a gas is expressed as the weight w of the
gas in grams which is absorbed by 100 grms. of pure solvent at the indicated temperature
and a total pressure made up of the partial pressure of the gas plus the vapour pressure of
the liquid at the temperature of the absorption. Thus for ethane at 10°, )3 = 0*0656 ;
iSf =0-0648 ; and t^ =0*0087. The concentration of the gases in a solvent can also be treated
as in the case of solids. For example, O. F. Tower (1906) found that 27*22 c.c. of 98 per cent.
sulphiu*ic acid at 18*5° and 714*6 mm. barometric pressure absorbed 0-58 c.c. of nitrogen.
The solubility of nitrogen in the acid at 18-5° is therefore 0-58127*22=0*0213 ; and the
absorption coefficient is 0*0213^(1+0*00367 x 18*5) =0*0199.
J. Dalton ascribed the absorption of a gas by a liquid as being due to the pene-
tration of the particles of the gas between the molecules of the liquid, just as
E. Swedenborg^ in 1721 said that the particles of a salt which dissolve in a liquid cannot
add to the bulk but only to the weight, because they occupy the spaces between the
particles of water. J. Dalton, however, said that the greatest difficulty attending
the mechanical hypothesis arises from the different gases observing different laws.
Why does water not admit to its bulk every kind of gas aUke ? The fact that absorp-
tion decreases as the temperature increases and as the intermolecular spaces increase
is opposed to the idea that absorption is due to a simple penetration of the gas mole-
cules between the molecules of the solvent.
The kinetic theory o! gaseous solution. — The kinetic theory of gases furnishes
a mental picture of the process of solution of a gas in water. Suppose that a gas-
free liquid be brought into a vessel containing a gas. The molecules of gas impinging
upon the surface of the liquid will be absorbed. The dissolved molecules move about
in the Hquid in all directions, a small number escape back into the gas above. As
the molecules of the gas crowd more and more in the liquid, the number of molecules
absorbed by the liquid becomes more and more nearly equal to the number which
escapes back into the superincumbent gas. If the pressure of the gas remains con-
stant, a time will come when the number of gas molecules which leave the liquid will
be equal to the number absorbed. The system is then in a state of dynamic equili-
brium resembling the equilibrium of a vapour in contact with its own liquid, and the
solution is saturated with the gas under the given conditions of temperature and
pressure. The solution of oxygen, for example, can be symbolized : 02gas^02soiution.
The surface of the liquid in contact with a dissolving gas must be very quickly
saturated with the gas, and the rate of absorption of a gas by a liquid at rest is
is really a measure of the rate of diffusion of the gas from the surface through the body
SOLUTIONS 529
of the liquid. The molecules of the liquid must have some attractive influence on
the molecules of the dissolved gas.
It is probable that the attractive forces between the molecules of the liquid and
the dissolving gas determine the solubility, otherwise we should expect the solubility
of certain groups of gases to be the same. True enough, as M. Kofler (1913) ^ has
shown, the critical temperature oftJie gas is a controlling factor because the solubilities
of different gases in a given solvent are approximately the same at corresponding
temperatures. M. Kofler assumed that the solubilities of gases in liquids is dependent
upon the magnitude of the intermolecular spaces, and that the latter in turn are
dependent on the compressibility of the liquid solvent ; if so, the greater the compressi-
biUty of a liquid the greater its solvent action on gases. Similarly, assuming that a
high dielectric constant is associated with small intermolecular spaces, the solu-
bility of a gas should be related with the dielectric constant of the liquid solvent.
On comparing these deductions with observations, it was found that some quali-
tative relations could be detected ; but the greater variations in the solubility of
different gases in the same liquid show that the properties of the gases themselves play
a most important part. A. Kitzel assumed that the solubility of a gas in a liquid
is proportional to the compressibility j3 of the solvent, and inversely propor-
tional to the change in volume S which accompanies the absorption of the gas ; or
S=kpi8, where k measures the solution pressure of the gas.
The influence of pressure on the solubility of gases. — With gases which are
not very soluble in liquids, the greater the pressure, the more soluble the gas ; that
is, the greater the pressure, the greater the concentration of the gas in the solution.
W. Henry (1803) ^ discovered an important relation between the pressure and the
solubility of a gas. A specified quantity of liquid under a total pressure, p, of
2 atmospheres, holds in solution twice as much gas by weight, w, as under a total
pressure of 1 atmosphere. Otherwise expressed, zv/p=Wilpi= . . . ; or wlp=a,
constant, or, the weight of gas absorbed by a specified volume of liquid is directly
proportional to the pressure. According to Boyle's law, the concentration of a
gas, or the amount of v in an enclosed space, is proportional to the pressure or
^i;=constant ; combining these two relations, w;/t;= constant. This means that
a gas will distribute itself so that its concentration in the liquid is proportional to
that in the space above — and this occurs whatever be the superincumbent pressure.
One volume of a gas at atmospheric pressure will contract to half a volume at a pres-
sure of two atmospheres. Under a pressure of 2 atmospheres, a saturated solution
of gas holds twice as much gas in solution as it did under a pressure of 1 atmosphere,
but two volumes of gas at atmospheric pressure occupy but one volume at a pressure
of 2 atmospheres. Hence follows Henry's law : under equal circumstances of
temperature, water takes up in all cases the same volume of the condensed gas
as it would if the gas were under ordinary pressure. That is to say,
S aSi S
— = — = . . .; or == Constant
P Vl P
The proportion of gas and liquid do not matter so long as a sufficient quantity of
each is present to allow the measurements to be made. The law thus describes the
behaviour of the less soluble gases very well — carbon monoxide, nitrogen, hydrogen,
oxygen — but not the more soluble gases like ammonia, hydrogen, chloride, sulphur
dioxide. The deviation is not very great with carbon dioxide,^ though it is appreci-
able :
Pressure, p .
. 1
5
10
15
20
25
30 atms.
Solubility, S
. 1-80
8-65
1603
21-95
26-65
30-55
33-74
Constant, S/p
. 1-80
1-73
1-60
1-46
1-33
1-22
1-12
The graph, Fig. 12, represents the observed values of p and S ; and the dotted
curve in the same diagram represents what the graph would have been had carbon
dioxide behaved as described by Henry's law. The value of Sjp is not therefore quite
VOL. I. 2 m
530
INORGANIC AND THEORETICAL CHEMISTRY
constant. If the volume of absorbed gas is referred to the volume of the solution
instead of the solvent alone, the constant works out better. The failure arises from
the fact that carbon dioxide reacts chemically with water. For the very soluble
gases SilTp = constant, where i is a constant characteristic of the gas.
Henry's law refers to gases which do not act chemically on the solvent. When
carbon £oxide dissolves in water, one portion enters into combination to produce
a new substance — carbonic acid — while the other portion dissolves in the physical
sense as carbon dioxide. The latter portion alone comes within the province of
Henry's law. The condition of the carbon dioxide which reacts with the water
is represented by C02+H20=H2C03. With a rise of temperature the
equilibrium is displaced from right to left, corresponding with the fact that the
higher the temperature of observation, the more nearly does Henry's law approximate
to the truth. Thus, H. E. Roscoe and W. D. Dittmar (1869) found that the law
applies very well for sulphur dioxide at 40°, and for ammonia near 100°. At 100°,
also, the solubility of carbon dioxide is proportional to the pressure. If the partial
pressure of the hydrate CO2.H2O is proportional to the concentration of the dis-
solved carbon dioxide, the presence of the hydrate has no influence on the law of absorp-
tion 7 provided there is no change in the gaseous molecule, resulting in the formation
of molecules of the tjrpe (C02)n-H20. Suppose the initial value of n is unity.
If Si be the concentration of the gas CO2 ; S2, that of the water ; and S, that of the
hydrate, then the condition of equilibrium in the solution C02.H20^C02+H20
is, according to Gulberg and
Waage's law, SiS2=KiS. If the
solubility of the gas is not great
the solution will be dilute, and the
water will be in so great an excess
that its concentration ' will be
virtually constant, and therefore
30
20
10
.
rn/n
n
rat
n
^0
0
y
7^
pe
^
V"
^
y
,-.
-'-
'"
^
x
-'■
.^
-^
_J
10
20
30
vols. CO2 per cc
Fig. 12.— Solubility Curve of COg in Water.
Si=K2S, where K2=KiS2. From
Henry's rule, the partial pressure
p of the gas will be p=K^Si,
and therefore by substitution for
Si, p=K2K^S ; or, substituting
the constant K=K2K^, it follows
that p=KS, or the concentration of the hydrated molecules is proportional to the
pressure of the gas, just as is the case with the unhydrated molecules. Consequently,
assuming that the partial pressure of the hydrate is proportional to the concentra-
tion S of the dissolved gas, ^/^=: constant, just as would be the case if the carbon
dioxide were all dissolved in the form of CO2, and none as CO2.H2O.
Henry's law also assumes that the molecular weight of the dissolved gas is the
same in solution and in the gaseous state. If ^the gas A be polymerized in solution
so that 2Aga3=A23oiution ; then, if Si be the concentration of the free gas, and S the
concentration of the dissolved gas, by the law of mass action, Si^=kiSy where ki
is a constant. Accordingly, Si=\/kiS, and Henry's law assumes the form p=k\/S,
where k is constant, p the pressure of the gas, and S the solubility — all expressed
in proper units. Analogous remarks apply if the gas in solution is polymerized to
a higher degree, say n; or if the gas is depolymerized or decomposed.
Since the concentration of a substance is understood to refer to the quantity of
substance in unit volume, Henry's law means that in a closed vessel, containing gas
and liquid, the gas will distribute itself so that its concentration in the liquid is
proportional to that in the superincumbent space. Hence it may be inferred that
i! a gas obeys Henry's law, it will have the same molecular weight in solution
and in the gaseous condition. Henry's law is therefore to be regarded as a link
connecting the molecular weight of gaseous and dissolved substances with one unit
of measurement. If 32 grams of oxygen depress the freezing point to the same extent
as 342*2 grams of cane sugar, it would be inferred that the molecular weights of
SOLUTIONS 531
oxygen and cane sugar are related as 32 : 342*2 ; and since oxygen gas has a mole-
cular weight of 32, it is assumed that cane sugar if it could be vaporized, and if its
gas obeyed Henry's law, would have a molecular weight of 342*2, because, as indi-
cated above, the molecular weights of a substance in solution and in the gaseous
state are assumed to be the same.
It might be emphasized, in passing, that when a gas is dissolved in a liquid at a
given temperature, the ratio between the concentration of the gas in the liquid and
in the space above is always the same. Thus, Henry's law is a law of distribution
for gases because it describes the way a gas distributes itself between the solvent and
the space above. Henry's law also describes the condition of equilibrium of a gas
whose molecules are physically and chemically independent of each other, and of
the solvent.
Example.- — Show that the absorption coefficient is independent of pressure. If S
volumes of gas at a pressure p are absorbed by unit volume of liquid, S^ vols, by Henry's
law will be absorbed at 760 mm. such that SiP = 160S. Again, if S^ volumes of gas at a
pressure p and temperature 6, become Sq volumes at 760 mm. and 0° ; by Boyle's and
Charles' la,w8, S^p = l 60S o{l-{-ae). Substituting iov S^ from the preceding expression and
solving for Sq, it follows that So=S/{l-\-ae), which is independent of p. If F volumes
of liquid have been treated, aS'o='S'/F(14-«^), which is Bunsen's coefficient of absorption, j8.
Solids and liquids dissolve in a vacuum or in a dilute indifferent gas in accord with
their vapour pressure,8 but if the indifferent gas be strongly compressed, say at 100
atm., a specific solvent action appears. Thus, compressed oxygen has a greater
solvent action on bromine vapour than oxygen under reduced pressure, and at 300
atm. pressure, the colour of the vapour is six or seven times as dense as under atmo-
spheric pressure ; while compressed hydrogen'has but a smaller solvent action. Iodine
imparts an intense violet colour to methane under 300 atm. pressure. Camphor
and paraffin likewise dissolve in compressed methane or ethylene gas, and on removing
the pressure, the iodine, camphor, or paraffin are deposited as crystals on the walls
of the vessel.
The applicability of Henry's law for solvents other than water was proved by
R. Bunsen^ for alcohol; byM. Woukoloff, for chloroform and carbon disulphide ;
and for petroleum by S. Gniewosz and A. Walfisz. The absorption coefficients of
a gas in different solvents are not proportional to one another. In order to test the
applicability of Henry's law, W. Sander examined the solubility of carbon dioxide
in water and in a number of organic liquids at temperatures between 20° and 100°
and at pressures between 20 and 170 kgrms. per sq. cm. He found that Henry's
law is the more nearly followed the higher the temperature ; at the lower tempera-
tures, the solubility of the gas in alcohol, benzene, chloro-, bromo-, and nitro-
benzene, and toluene increases faster with increasing pressure than corresponds with
Henry's rule, and with ethyl ether, ethyl acetate, and water, the increase is slower
than Henry's law requires. The law is more nearly followed when the volume of
gas absorbed is referred not to the volume of the solvent (coefficient of absorption)
but to the volume of the solution (Ostwald's solubility). With carbon dioxide in
ether, and ether vapour in carbon dioxide, Henry's law is not approximately valid
in the neighbourhood of the critical point. 0. Sackur and 0. Stern have likewise
examined the effect of pressure on solutions of this same gas in methyl and ethyl
alcohols and acetates between —59° and 78°, and between 50 and 700 mm.
pressure.
Henry's law and the kinetic theory. — The gas is in equilibrium with its own
solution when the number of molecules which escape from the solution is the same
as those which are captured by the solution in a given time. By, say, doubling the
pressure the molecular concentration will be doubled, the gas molecules will be
crowded more closely together, and the rate at which the solution captures the
molecules will be increased twofold for the new state of equilibrium. Similarly the
rate of escape will be doubled. Hence variations o! pressure do not alter
the relative number of molecules per unit volume of solution and of gas ; and the
532 INORGANIC AND THEORETICAL CHEMISTRY
volume of gas dissolved will be independent of the pressure on the gas, while the
weight of gas dissolved will be directly proportional to the pressure.
It will be observed that in the relation showing the influence of pressure on the solubility
{d log S)/dp=BvlRT, the magnitude 8v may be interpreted to mean the change in volume
which occurs during the process of solution. The volume of the gas is so great in relation
to the solution that the volume v of the gas can be substituted for 8v, and assuming Boyle's
law is applicable pv=RTf and substituting for v
dlogS _l
dp p
which, on integration, furnishes S==kpy that is, Henry's law.
The influence of temperature on the solubility o! gases. — The solubility of a gas
in a liquid is very sensitive to changes of temperature. The higher the temperature,
the less the solubility of the gas. R. Bunsen's measurements agree with the
assumption that the absorption coefficient of hydrogen in water and of oxygen in
alcohol are not affected by changes of temperature between 0° and 20° ; but
W. Timofejeff found that R. Bunsen's results were not confirmed by a more sensitive
method of measurement.io The solubihty curve of helium is not much affected by
changes of temperature up to 50°, but what little effect there is seems to indicate that
the solubihty of the gas increases as the temperature rises from 25° to 50°. Hydro-
gen was once supposed to behave in a similar way, between 0° and 25°, but later,
more careful measurements show that the solubility decreases steadily from 0*0214
at 0° to 001 71 at 26°. The solubility of carbon dioxide in nitrobenzene is nearly
the same at 100° as it is at 60°.
A. Imhof represented the solubility, S, of a gas (litres of gas in a litre of water) by the
expression :
^ = e27-4. orr = 27-4 1og6f
where T denotes the absolute boiling temperature in degrees reckoned from the point where
5=1 and log 6'=0, namely, —100°. T is positive if higher than —100°, and negative
if below —100°. The results have less than a 2 per cent, error with all the gases tried
excepting methane, where the error is 5*3 per cent., and hydrogen and helium where the
errors are respectively 16*1 and 18*9 per cent. For oxygen, <S = 0'049 ; T=— 82*66°
(observed -82-5°) ; for acetylene, 5 = 1-73 ; and r = — 15-02 (observed + 16'0).
The influence of the surface tension of the solvents has been previously discussed.
References.
1 T. Bergmann, Opuscula physica et chemica, Holmiae, 1. 9, 1779 ; H. Deicke, Pogg. Ann.,
119. 156, 1863 ; T. Thomson, J. J. Mackenzie, and E. L. Nichols, Ann. Chim. Phys., (5),
3. 134, 1878 ; W. Ostwald, Lehrbuch der allgemeinen Chemie, Leipzig, 1. 632, 1903 ; K. Angstrom,
Wied. Ann., 15. 297, 1882 ; E. L. Nichols and A. W. Wheeler, Phil. Mag., (5), 11. 113, 1881.
2 R. Bunsen, Liehig's Ann., 93. 1, 1855 ; W. Ostwald, Lehrbuch der allgemeinen Chemie,
Leipzig, 1. 616, 1903 ; 0. F. Tower, Zeit. anorg. Chem., 50. 382, 1906.
3 E. Swedenborg, Prodromus principorum rerum naturalium sive novorum testaminum chymiam
et physicam experimentalem geometrice explicandi, Amsterdam, 51, 1721 ; J. Dalton, A New
System of Chemical Philosophy, Manchester, 1. 56, 1808.
« M. Kofler, Sitzber. Akad. Wien, 122. 1, 1913 ; A. Ritzel, Zeit. phys. Chem., 60. 319, 1907.
6 W. Henry, Phil. Trans., 29. 274, 1803 ; R. Bunsen, Liebig's Ann., 93. 1, 1856 ; J. W. Doyer,
Zeit. phys. Chem., 6. 481, 1890.
« N. Khanikoff and V. Longuinine, Ann. Chim. Phys., (4), 11. 412, 1867 ; S. von Wroblewsky,
Wied. Ann., 18. 302, 1883; M. WoukolofF, Compt. Rend., 108. 674, 1889; 109. 61, 1889;
L. Carius, Liebig's Ann., 93. 33, 1855 ; H. E. Roscoeand W. Dittmar, Jowm. Chem. Soc, 13. 128,
1860 ; T. H. Sims, ib., 14. 1, 1861 ; Liebig's Ann., 118. 345, 1861 ; F. Schonfeld, ib., 95. 1, 1854 ;
W. M. Watts, Liebig's Ann. Suppl, 3. 227, 1865.
' W. Ostwald, Lehrbuch der allgemeinen Chemie, Leipzig, 2. ii, 608, 1902 ; W. D. Bancroft,
Journ. Phys. Chem., 3. 551, 1899 ; The Phase Rule, Ithaca, 5, 1897.
» P. Villard, Journ. Phys., (3), 5. 453, 1896 ; Chem. News, 78. 297, 309, 1898.
» R. Bunsen, Liebig's Ann., 93. 10, 1855 ; M. Woukoloff, Compt. Rend., 108. 674, 1889 ; 109.
61, 1889 ; S. Gniewosz and A. Walfisz, Zeit. phys. Chem., 1. 70, 1887 ; W. Sander, ib., 78. 513,
1912 ; 0. Sackur and O. Stem, Zeit. Elektrochem., 18. 641, 1912.
10 R. Bunsen, Liebig's Ann., 93. 10, 1855 ; Ber., 22. 1439, 1889 ; W. Timofejeff, Zeit. phys.
Chem., 6. 141, 1890 ; A. Imhof, ib., 91. 124, 431, 1916.
SOLUTIONS 533
§ 5. The Solubility o! Mixed Gases in Liquids— Dalton's Law
When a mixture of two gases is exposed to the action of a solvent, the quantity
of each gas dissolved by the liquid depends upon the amount and the solubility of
each gas present. The amount of each gas determines its partial pressure, and
since the partial pressure of each gas is independent of the others, it follows that
when a mixture of gases is exposed to the action of a solvent, and no chemical
action intervenes, the amount of each gas which is dissolved by the solvent is
proportional to the partial pressure of the gas. Each gas behaves as if the others
were absent. This is called Dalton's law, after its discovery by J. Dalton,i 1805 ;
it is obviously a special case of Henry's law.
J. Dalton's idea was that the gases dissolved in water retain their elasticity
or repulsive power among their own particular molecules the same in the water as out
of it, the intervening water having no other influence in this respect than a mere
vacuum. The idea that the solvent water is wholly passive cannot now be main-
tained. J. Dalton further showed that each gas is retained in water by the pressure
of gas of its own kind incumbent on the surface ; abstractedly considered, no other
gas with which it may be mixed has any permanent influence. J. Dalton also
had wrong ideas of the numerical relation between the gas dissolved and that
incumbent on the liquid. R. Bunsen 2 studied the solubiHty of mixed gases, and, in
his Ueher das Gesetz der Gasabsorptionj pointed out :
Let V vols, of a mixture of gases at a pressure P, containing, per unit volume, v^
volumes of a gas A, v^ vols, of a gas B, Vg vols, of a gas C, . . . with the respective coeffi-
cients of absorption jSi, ^2> i^a, . . . be agitated with V vols, of a liquid, so that there remains
u vols, of the gaseous mixture containing w^, u^, u^,... respectively of the gases A, B, C, . . .
per unit volume at a pressure P^. The temperature remains constant at 0°. The mixture
contains v^v vols, of the gas A at a pressure P, or v^vPjlQO vols, at a pressure 760 mm. This
volume of gas is divided into x^ vols, of gas which remain unabsorbed and yj vols, which are
absorbed by the v vols, of liquid ; but from the law of absorption, unit volume of liquid
absorbs ^^ vols, of gas at 760 mm. pressure, or V vols, of liquid absorb ^iVpiflQO vols, at a
partial pressure p^. The gas Ai, however, expands from x^ to up J7 60 vols, when admixed
with the other gases, so that the quantity absorbed by V vols, of liquid in virtue of the
partial pressure piVx^/u is 2/1, or yi=PiVxJu; or the volimae of the component A is
Xi-\-fiiVxJu=VivP:160 ; or Xy=v^vPJ760{l-\-$iVlu) ; and generally.
760
OTTy '"MPT)
The volume of residual gases remaining when a gaseous mixture of volume v=v^-\-v^
-\- . . . has been exposed to v vols, of a solvent is w = (Wi+W2+ . . ., where
aJa
.' " x^+x.,+ . . .'
J. Dalton believed that the influences of temperature on the amounts of various
gases dissolving in the same liquid is proportional to the influence of each as
separately, so that the composition of the gas dissolved by a specified liquid acting on
a mixture of gases is independent of the temperature. This statement is not strictly
accurate.
When air containing, say, 79 volumes of nitrogen (neglect the argon and rare
gases) and 21 volumes of oxygen, and 0"04 volume of carbon dioxide, is skaken up
with water, the amount of each gas absorbed by the water can be approximately
computed in the following manner : The relative solubilities are : nitrogen, 002 ;
oxygen, 0*04 ; and carbon dioxide, 1'79. The partial pressure of each gas is propor-
tional to the relative amount of that gas present in a given volume of air. If the
pressure of air be just one atmosphere, the partial pressure of the nitrogen will be
proportional to 0"79xl ; of oxygen, 0'21xl ; and of carbon dioxide, 0'0004xl.
Hence the relative amounts of these gases absorbed by the water will be : nitrogen,
0-79x0-02=0-0158; oxygen, 0-21 X 0*04 =0*0082 ; and carbon dioxide,
534
INORGANIC AND THEORETICAL CHEMISTRY
0'0004xl*79=000072. Hence 1- c.c. of water dissolves 0*0158 c.c. of nitrogen;
0*0082 c.c. of oxygen ; and 0*00072 c.c. of carbon dioxide. The composition of the
dissolved gases, if removed from the solution by boiling, or exposure to a vacuum, will
be : nitrogen, 63'9 per cent. ; oxygen, 33*2 per cent. ; carbon dioxide, 2*9 per cent.
The relatively large solubility of the carbon dioxide of the atmosphere is counter-
balanced by its low partial pressure, otherwise we might expect a heavy rainstorm
to remove a great part of the carbon dioxide from the surrounding air.
J. T. A. Mallet (1869) ^ has a proposal to separate oxygen from atmospheric air
freed from carbon dioxide, which is based on the different solubilities of the oxygen
and nitrogen. If the carbon dioxide be removed by passing the air through an
aqueous solution of sodium hydroxide, the oxygen and nitrogen in the remaining
gases after the first absorption will be nearly in the proportion : nitrogen 65'7 per
cent . , and oxygen 34*3 per cent. If this mixture be driven from the water by boiling,
and the mixture again treated with air-free water, a gaseous mixture containing
49 per cent, of oxygen is obtained ; and after the eighth absorption, a gas containing
97 per cent, of oxygen results. The relative proportions of oxygen and nitrogen
in air obtained from water after successive absorptions is
Table V.-
-Effect of Successive Absorptions by Water on the Composition
OF Air.
Number of absorptions.
0
1
2
3
4
5
6
7
8
Nitrogen
Oxygen
79
21
66-7
33-3
62-5
47-5
37-5
62-5
25-0
750
25-0
85-0
9-0
91-0
5-0
95-0
2-7
97-3
The method is not practicable though it is an interesting application of Henry's
and Dalton's laws.
Examples. — (1) The solubility of hydrogen is 0'02 and of oxygen 0*04. Show that
13"3 c.c. of each of these gases is dissolved by 1000 c.c. of water from an electrolytic mixture
of hydrogen and oxygen.
(2) If 10 c.c. of an aqueous solution of carbon dioxide saturated at 0° is introduced into
a vessel already containing 10 c.c. of carbon dioxide all at atmospheric pressure, show that
8*69 c.c. of carbon dioxide will remain in solution. D. I. Mendelceff (1868) found 10 c.c. of a
saturated solution at 0° contain 18 c.c. of carbon dioxide, and if x denotes the number of c.c.
which remain in solution, 18— a; will represent the number of c.c. expelled and 28— a;
will be present in the atmosphere. Hence, the partial pressure of the dissolved carbon
dioxide is (18— rc)/(28— a;). When the solution is at atmospheric pressure it contains
18 c.c. of carbon dioxide, and when the partial pressure is (18— x)/(28— a;), it contains
18(18— a;)/(28—aj)=a; c.c. of carbon dioxide, when a: = 8*69 c.c.
The effect of saline solutions on the solubility o! gases.— The behaviour of
gases towards salt solutions first attracted the attention of physiologists owing to
its bearing on the absorption of gases by the blood. Thus J. S. F. Pagenstecher,*
R. F. Marchand (1846), J. von Liebig (1851), L. Meyer (1857), and E. Fernet (1858),
examined the solubiUty of carbon dioxide in solutions of sodium phosphate. The
solubility of a gas is lowered by the dissolution of a salt which does not act chemically
on the gas. Thus, F. M. Raoult (1873) ^ found that the solubiHty of ammonia in
aqueous solutions of potassium hydroxide decreased as the proportion of alkali
increased from 72 with solution containing llj per cent. K2O, to 495, with
solutions containing 25^ per cent, of K2O. J. Setschenoff (1889) found that the
relation between the quantity of salt x and the absorption coefficient is given very
nearly by the formula :
Absorption coefRcient=pe x
where j3 represents the absorption coefficient of the gas for water, ^ is a specific
constant dependent upon the nature of the dissolved salt, and e is the base of
k
/
\
/
s
\,
/
X,
1 1 1 1
Composition of So/vent
, SOLUTIONS 535
natural logaritlims. Equivalent solutions of similar salts of the same acid absorb
nearly the same quantities of gas. For instance, J. Setschenofli (1875) found that
with calcium, strontium, and barium nitrates the absorption coefficients were
respectively 0*923, 0-916, and 0'922. The efiect of mixing another liquid with the
water resembles that obtained by the dissolution of a
salt ; thus J. S. Setschenoff (1875) found the absorption ^
coefficients of mixtures of water and sulphuric acid to | o9
be less than for either water or sulphuric acid alone, as | ^ g
illustrated in Fig. 13. Other physical properties of
mixtures — viscosity, electrical conductivity, etc. — 1°^
change in a similar manner. 0. Miiller (1889) obtained b gg
similar results with mixtures of alcohol and water. If ^
the salt is acted on chemically by the gas, as is the case °'^ioo eo eo 4o 20 ohjo
, , T -1 • T 1 J • 1 J.- 0 20 40 60 80 lOOHSO,
when carbon dioxide is dissolved in aqueous solutions
of borax, sodium carbonate, or sodium phosphate, the ^Jjioxide in Mixtures of Sul-
portions of gas held chemically by the salt is almost phuric Acid and Water,
independent of pressure, while the other portion follows
Henry's law. The decrease is supposed to be due to the fixation of some of the
solvent by the molecules or the ions, or both molecules and ions of the dissolved
salt. On this assumption, J. C. Philip ^ calculated the degree of hydration of the
salt from the decrease in the solubility of the gas from the formula :
, . a— 6100— c Ml
Degree of hydra tion = - :r=r
^ ^ a c M
where Mi and M respectively denote the molecular weights of salt and water ; c
the per cent, of salt in solution ; a denotes the number of c.c. of oxygen dissolved by
a litre of water; h the number of c.c. of oxygen in 1000 grams of water in the solution
calculated from h=aj{I)—C), where D denotes the density of the solution, and C
the number of grams of water per c.c. of solution. For example, with potassium
chloride, bromide, and iodide, C. G. McArthur finds :
KCl KBr KI
Concentration
. IN-
2JV-
IN~
2^-
IN~
2N~
Specific gravity .
1-0086
1086
1-017
1-150
1-027
1-230
Oxygen c.c. per litre .
5-30
3-21
5-52
3-37
5-49
3-77
Degree of hydration
. 160
10-4
7-8
8-9
8-9
6-4
The hydration data with salts whose ionization is small were found to give results
consistent with the degree of hydration calculated by other methods.
References.
1 J. B<on, Mem. Manchester Lit. Phil. Soc, (2), 1. 271, 1805 ; A New System of Chemical
Philosophy, Manchester, 1. 197, 1808.
2 R. Bunsen, Liebig's Ann., 93. 1, 1854; Phil. Mag., (4), 11. 116, 181, 1855; W. Dittmar,
Chemical Arithmetic, Glasgow, 1890.
3 J. T. A. Mallet, Dingier s Journ., 199. 112, 1871.
* J. S. F. Pagenstecher, Buchner's Repert., 22. 318, 1840 ; R. F. Marchand, Journ. prakt.
Chem., (1), 37. 321, 1846 ; J. von Liebig, Liebig's Ann., 79. 112, 1851 ; E. Fernet, Compt. Bend.,
46. 620, 1858 ; L. Meyer, Pogg. Ann., 102. 299, 1857 ; L. Meyer and R. Heidenhain, Liebig's
Ann. Suppl, 2. 157, 1863 ; W. Ostwald, Solutions, London, 44, 1891.
« 0. duller, Wied. Ann., 37. 24, 1889 ; 0. Lubarsch, ib., 37. 524, 1889 ; J. Setschenoff, Mem,
Acad. St. Petersburg, 22. 102, 1875 ; Mem. Soc. Nat. Moscou, 15. 6, 1889 ; F. M. Raoult, Ann.
Chim. Phys., (5), 1. 262, 1874 ; C. G. McArthur, Journ. Phys. Chem., 20. 495, 1916 ; P. Steiner,
Wied. Ann., 52. 275, 1894 ; V. Gordon, Zeit. phys. Chem., 18. 1, 1895 ; H. John, ib., 18. 8, 1895 ;
W. A. Roth, ib., 24. 114, 1897 ; L. Braun, ib., 33. 721, 1900; W. Knopp, ib., 48. 97, 1904 ;
G. Hufner, ib., 57. 6] 1, 1907 ; G. Geffcken, ib., 49. 257, 1904.
« J. G. Philip, Jonrn. Chem. Soc, 91. 711, 1907 ; Trans. Faraday Soc, 3. 140, 1907 ; W. R.
Bousfield and T. M. Lowry, ib., 3. 123, 1907 ; C. G. McArthur, Journ. Phys. Chem., 20. 495,
1916 ; G. McP. Smith, Journ. Amer. Chem. Soc, 37. 722, 1915.
536 INORGANIC AND THEORETICAL CHEMISTRY
§ 6. Diffusion in Gases and in Liquids
If a very small quantity of a salt be dissolved in a great quantity of water, the particles
of the salt will not sink to the bottom though they be heavier in specific gravity than the
water, but they will evenly diffuse themselves into all the water so as to make it as saline
at the top as at the bottom. Does not this imply that the parts of the salt recede from one
another, and endeavour to expand themselves and get as far asunder as the quantity of
water in which they float will allow ? And does not this endeavour imply that they have a
repulsive force by which they fly from one another, or at least, that they attract the water
more strongly than they do one another ?■ — Isaac Newton (1675).
Let a large crystal of a coloured salt — say copper sulphate or potassium dichro-
mate — be placed at the bottom of a tall glass cylinder, and the remainder of the jar
be filled with water. A coloured salt is chosen because the movements of the
resulting solution can be readily seen. Let the jar stand where it will not be disturbed
by evaporation, agitation, etc. The surface of separation between the solid and
solvent will be gradually obliterated ; in time, the coloured salt will diffuse uniformly
throughout the whole body of liquid. Similarly, if a solution of one concentration
be in contact with a solution of another concentration, the dissolved substance
passes from the region of greater to the region of lesser concentration, until the
concentration is uniform throughout the whole mass of liquid provided the tem-
perature is everywhere the same.^
The phenomenon of diffusion was known to Isaac Newton (1695), to C. Berthollet
(1803), and to F. Parrot (1815). The last-named attributed the action to a special
force which he called Affinitdt erster Art, eine neu aufgedeckte Naturkraft — affinity of
the first degree, a newly discovered natural force — which M. L. Frankenheim (1835)
styled diffusion. F. Parrot said : All miscible liquids show a tendency to wander
one into the other when they are brought into contact, and this process continues
until the liquids are perfectly evenly distributed. W. Nernst attributed the driving
force to osmotic pressure, a phenomenon about to be described. T. Graham first
obtained quantitative data about the speed of diffusion of different salts, and
A. Fick then developed a theory of the process based on the hypothesis that tJie quan-
tity of a salt which diffuses through a given area in a given time is proportional to the
difference between the concentration of two vertical and parallel planes indefinitely close
to OTte another ; or, the amount of solute dm which will pass in a given time dt between
two parallel planes unit distance apart and of unit sectional area, is proportional
to the difference in concentration, Ci—C^, on the two sides of that section —
Fick's law of diffusion — or dm=k{Ci—C2)dt, where k is the coefficient or constant
of diffusion. The hypothesis was tested by many investigators — F. Beilstein (1856),
T. Simmler and H. Wild (1857), F. Hoppe-Seyler (1867), E. Voit (1867), and
A. Johannisjanz (1877), but the results, as J. Stefan (1878) showed, were not of a suffi-
cient degree of accuracy. Then followed the work of H. F. Weber, W. Seitz (1898),
J. Schuhmeister (1879), J. H. Long (1880), R. Lenz (1882), J. D. R. Scheffer (1881),
and P. deHeen(1884). The general results have established the validity of A. Fick's
law based on the theory of the conduction of heat, and that just as the magnitude
of the heat conduction decreases slowly with rise of temperature, so does the diffusion
decrease as the concentration increases. The mathematical theory has been dis-
cussed by W. Seitz, F. Niemoller, E. Voit, T. Simmler and H. Wild, J. Stefan,
0. Wiedeburg, 0. Wiener, P. G. Tait, H. F. Weber, F. Neimbrodt, J. Trovert, etc.
The speeds of diffusion of many salts have been investigated by T. Graham,
J. D. R. Scheffer, L. W. Oeholm, J. Trovert, F. Heimbrodt, W. Seitz, 0. .Wiede-
burg, W. Kawali, etc. T. Graham observed great differences in the rates of
diffusion of two classes of substances — what he called crystalloids diffused rapidly,
colloids slowly.
HCl NaCl Cane sugar. MgSO. Albumen. Caramel.
Velocity of diffusion . . 1 2-3 7 7 49 98 units
J. H. Long noted a parallelism between the velocity of diffusion and the electrical
conductivity. The diffusion of mixtures of salts has been investigated by T. Graham,
SOLUTIONS 537
J. C. G. de Marignac, and F. Kiidorff. Each salt seems to diffuse independently
of other accompanying salts. T, Graham, P. de Heen, H. F. Weber, W. Seitz,
L. W. Oeholm, and others have investigated the influence of temperature on
the speed diffusion, and found it to increase rapidly with a rise of temperature.
W.Nernst found that with dilute solutions of neutral salts, the coefficient of diffusion
k at e° is A:i8{ 1+0-026(0— 18)}, and for acids and bases ki^{\-\-0'02^e—\^)}.
T. Graham, E. Detlefsen, H. de Vries, L. Chabry, P. Nell, H. Bechhold and
J. Liegler, J. Hausmann, S. Leduc, N. Pringsheim, and F. Voigtlander studied the rate
of diffusion in agar-agar jelly and found the process similar to that which occurs
with water. E. E. Liesegang found that the diffusion of silver nitrate in a tube
of gelatine containing ammonium chromate furnishes a series of rings or laminae —
Liesegang's rings— at right angles to the axis of the tube. W. Ostwald suggested
that it is a supersaturation phenomenon. F. Kohler found that if the ammonium
chromate be too concentrated or too dilute the rings are not well developed ; and
if the gelatine contains the silver salt, rhythmic 'precipitation does not occur.
H. W. Morse and G. W. Pierce obtained a similar result with lead nitrate diffusing
into gelatine and sodium sulphate.
There seems to be some force at work driving the molecules of the solute upwards
against the force of gravity. From the kinetic theory, it is inferred that the mole-
cules of the liquid are in perpetual motion in all directions ; and that the protracted
time occupied by the diffusion of the molecules of the dissolved salt in the liquid is
due to the close packing of the molecules of the liquid, such that the free progress of
the molecules of the dissolved salt in the solvent is greatly impeded. It can be
shown from the kinetic theory that the potential energy of the molecules of a mixture
of gases is diminished by diffusion, and in consequence the phenomenon is due to the
tendency of the molecules of the mixing gases to follow the dynamical principle :
the position of stable equilibrium is the position of minimum potential energy ;
diffusion is motion towards a state of stable equilibrium.
The analogy between the dissolution of a substance in a solvent, and vaporiza-
tion, has been emphasized by R. Hooke (1664), by J. L. Gay Lussac (1839),^ by
B. Bizio (1845), and by A. Rosenstiehl (1870). A substance in solution was regarded
as an elastic vapour, and the difference between the dissolved substance and a gas
was said to arise from the circumstance that *' a gas does not need the presence of
the molecules of a solvent, and of their affinity to sustain it in the occupied space."
In 1873 A. Horstmann developed a thermodynamic theory of equilibrium between
gaseous substances, and showed that the same laws applied for substances in solution.
The idea gradually grew into chemistry, and proved singularly fruitful in the work of
J: H. van't Hoff (1886), who widely extended the analogy between the physical and
chemical behaviour of substances in dilute solution, and in the gaseous state.
Just as the molecules of a gas in a closed vessel are disseminated in a relatively
large space, so are the molecules of a solid in solution scattered in a relatively large
.volume of solvent. It is true that the molecules of the salt in solution could not
occupy the space if the solvent were absent, otherwise the analogy between a sub-
stance dissolved in a solvent and a gas scattered in space would be very close. Argu-
ments from analogy are notoriously treacherous ; and whatever conclusions might
be inferred from a closer study of the analogy between the process of solution and
gaseous diffusion, the fact that the molecules of the dissolved substance are co-
mingled with the solvent, and that the molecules of the gas are not associated with
such an agent, must be constantly borne in mind. As G. F. Fitzgerald 3 has said :
" The dynamical condition of molecules in solution is essentially and utterly
different from that of the molecules of a gas."
The rate of solution of a solid in a solvent depends on the surface area, and on
the amount of the solid already present in solution. This latter was suspected by
C. L. Berthollet* in 1803, and established by the experiments of A. A. Noyes and
W. R. Whitney. They showed that the rate of solution of a solid is proportional to
the difference between the concentration of the film in immediate contact with the
538 INORGANIC AND THEORETICAL CHEMISTRY
solid and with the more dilute layers. Consequently, it follows that the solution
of a solid involves two processes : (i) The reaction between the solvent and solid ;
(ii) The rate of diffusion of the solute away from the solid. If the speed of the latter
process predominates, the observed rate of solution will not depend merely on the
amount of solid already on solution ; whereas if the speed of the former predominates,
the observed rate of solution will be proportional to the concentration of the solution
in conformity with the observations of A. A. Noyes and W. R. Whitney — see crystals.
References.
1 T. Newton, Opticks, London, 1695 ; C. L. BerthoUet, Essai de statique chimique, Paris, 1803 ;
F. Parrot, OilberVa Ann., 51. 318, 1815 ; M. L. Frankenheim, Die Lehre von der Kohdsion, Breslau,
1835 ; W. Nemst, Zeit. phys. Chem., 2. 613, 1888 ; T. Graham, Phil. Trans., 140. 1, 805, 1850 ;
141. 483, 1851 ; 151. 188, 1861 ; A. Fick, Fogg. Ann., 94. 59, 1855 ; Phil. Mag., (4), 10. 30, 1855 ;
R. E. Liesegang, Chemische Fernewirkung, Dusseldorf, 1896 ; Chemische Reaktionen in
Oallerten, Dusseldorf, 1898 ; H. de Vries, Maandhlad Naturw., 11. 118, 1884 ; F. Beilstein, Liehig's
Ann., 99. 165, 1856 ; A. Lieben, ib., 101, 77, 1857 ; A. Fick, ib., 102, 97, 1857 ; E. Detlefsen,
Zeit. phys. Unierrichts, 2. 249, 1885 ; A. Weinhold, Zeit. chem. Unterrichts, 1. 262, 1888 ;
P. Nell, Ann. Physik, (4), 18. 323, 1905 ; G. Quincke, ib., (4), 11. 447, 1903 ; F. Heimbrodt, ib., (4),
13. 1028, 1904 ; T. Simmler and H. Wild, Fogg. Ann., 100. 217, 1857 ; E. Voit, ib., 130. 227,
393, 1867 ; N. Umoflf, Journ. Russian Phys. Chem. Soc, 23. 335, 1891 ; T. Martini, Nuovo CimerUo,
(3), 9. 156, 1882 ; Atti Jst. Veneto, (6), 6. 16, 1889 ; (6), 7. 17, 1889 ; L. Marini, Rend. Accad.
Lincei, (5), 4. 135, 1895 ; F. Hoppe-Seyler, Medicinischchemische Untersuchungen, Berlin, 1. 1,
1867 ; E. L. R. Beez, Zeit. Math. Phys., 4. 212, 1859 ; 7. 227, 1862 ; 10. 358, 1865 ; R. Lenz, Mem.
Acad. St. Petersburg, (7), 30. 9, 1882 ; P. de Heen, Bull. Acad. Belgique, (3), 8. 219, 1884 ; F. Riidorff ,
Ber., 21. 4, 3044, 1888 ; L. Marchlewsky, ib., 26. 983, 1893 ; J. J. R. Schefter, ib., 15. 788, 1882 ;
16. 1903, 1883 ; Zeit. phys. Chem., 2. 390, 1888 ; J. Hausmann, ib., 40. 110, 1904 ; A. Johannisjanz,
Wied. Ann., 2. 24, 1877 ; H. F. Weber, ib., 7. 469, 536, 1879 ; W. Sietz, ib., 64. 759, 1898 ; H. L.
Long, ib., 9. 613, 1880 ; B. von Tietzen-Hennig, ib., 35. 467, 1888 ; F. Niemoller, ib., 47. 694,
1892 ; 0. Wiedeburg, ib., 41. 675, 1890 ; 0. Wiener, ib., 49. 143, 1893 ; S. von Wroblewsky, ib.,
13. 606, 1881 ; W. Kowalki, ib., 52. 302, 1894 ; F. Wohler, Zeit. Kolloid, 19. 65, 1916 ; J. Stefan,
Sitzber. Akad. Wien, 78. 957, 1878 ; 79. 161, 603, 1879 ; J. Schuhmeister, ib., 79. 603, 1879 ;
E. Lenssen, Journ. prakt. Chem., (1), 85. 416, 1862 ; J. C. Graham, Zeit. phys. Chem., 50. 257, 1904 ;
F. Voigtlander, ib., 3. 316, 1889 ; H. Bechhold and J. Ziegler, ib., 56. 105, 1906 ; W. Ostwald,
ib., 22. 302, 1897 ; 23. 365, 1898 ; H. W. Morse and G. W. Pierce, ib., 43. 589, 1903 ; N. Pringsheim,
ib., 17. 473, 1895 ; M. W. Beyerinck, ib., 3. 110, 1889 ; L W.' Oeholm, ib., 50. 307, 1904 ; P. G.
Tait, Trans. Roy. Soc. Edin., 30. 551, 1883 ; S. Leduc, Compt. Rend., 132. 1500, 1901 ; 139. 986,
1904 ; A. P. Dubrunfaut, ib., 66. 354, 1868 ; D. Calugareanu and V. Henri, Compt. Rend. Soc.
Biol, 112, 1901 ; L. Chabry, Journ. Phys., (2), 7. 114, 1888; J. Trovert, Ann. Chim. Phys.,
(7), 26. 366, 1902 ; J. C. G. de Marignac, ib., (5), 2. 546, 1874.
2 J. L Gay Lussac, Ann. Chim. Phys., (2), 70. 407, 1839 ; B. Bizio, Mem. 1st. Veneto,
9. 79, 1860; M. Belleti, Atti 1st. Veneto, (7), 6. 679, 1895; A. Rosenstiehl, Compt. Rend.,
70. 617, 1870 ; A, Horstmann, Ber., 2. 137, 1869 ; 14. 1242, 1881 ; Liebig's Ann. Suppl, 8. 112,
1872 ; Liebig's Ann., 170. 192, 1873 ; Ostwald' s Klassiker, 137, 1903.
3 G. F. Fitzgerald, Journ. Chem. Soc, 69. 885," 1896.
* C. L. BerthoUet, Essai de statique chimique, Paris, 1. 65, 1803 ; L. Bruner and S. Tolloczko,
Zeit. phys. Chem., 35. 283, 1900 ; Zeit. anorg. Chem., 28. 314, 1901 ; 35. 23, 1903 ; 37. 455, 1903 ;
K. Drucker, ib., 29. 459, 1902 ; F. Novak, ib., 47. 421, 1905 ; P. de Heen, Bull. Acad. Belgique,
(3), 23. 235, 1892; A. A. Noyes and W. R. Whitney, Journ. Amer. Chem. Soc, 19. 930, 1897.
§ 7. Solution Pressure — Osmotic Pressure
Just as a small quantity of water is able to dissolve a quantity of salt which can diffuse
itself through a large quantity of water, so a quantity of air which can expand and diffuse
itself through a large space may be contained within a small compass.- — R. Hooke (1664).
It has been shown that if the diffusion of gases be resisted by placing a permeable
partition between two gases, a pressure will be exerted upon the partition. It is
easy to show that the particles of a dissolved substance exert a similar pressure
when a partition is placed between the solution and solvent so that the partition
offers no obstacle to the free circulation of the molecules of the solvent, but resists
the free passage of the molecules of the dissolved substance.
A piece of wet bladder is stretched and wired over the head of a wide thistle -headed
funnel with a stem about 10 cm. long. When nearly dry, the bladder is removed and the
SOLUTIONS 539
hot fiinnel is smeared about the rim with marine glue. The bladder is immediately wired
securely in position. The thistle -headed funnel is nearly filled with a concentrated solution
of cane sugar and joined by means of pressure tubing or a rubber stopper with a piece of
capillary tubing of ^ mm. bore bent S-shaped as indicated in Fig. 14. The fimnel is immersed
in a jar of water. The level of the index of coloured water in the capillary tube is marked
with gummed paper, and the apparatus is allowed to stand over night. In the morning the
liquid in the capillary will have risen about 10 cm. Water has obviously passed from the
beaker through the membrane into the sugar solution.
The passage of water through a membrane in this manner is called osmosis —
from the Greek wa/xos, a push. If the osmosis be inwards, towards the solution,
H. Dutrochet's term endosmosis can be used ; if outwards, exosmosis. The mem-
brane permeable to the solvent, impermeable to the dissolved substance, is called
a semipermeable membrane. The extra hydrostatic pressure exerted upon the
membrane by the sugar solution was styled, by W. F. P. Pfeffer (1877), " the osmotic
pressure of the sugar solution." Solutions with the same osmotic pressure are said
to be iso-osmotic or isotonic.
Experiments on osmosis were made by Abbe NoUet (1748).i He showed that if
the opening of a glass vessel containing alcohol be tightly covered with a bladder and
inverted in water, the contents of the vessel increase so that the bladder sometimes
bursts. F. Parrot next studied the phenomenon in 1803, and N. W. Fischer in 1822.
F. Parrot saw the important bearing of this subject on phenomena
or processes which occur in the living organism. Then K. J. H.
Dutrochet took up the subject in 1826 and subsequent years. The
greatest interest centred about the changes of level which occurred
when two different liquids separated by an animal membrane were
kept in contact. G. Magnus (1827), E. B. Jerichau (1825), E.
Briicke (1842) tried to develop a theory of the process ; K. Vierordt
(1845-8), P. Jolly (1849), J. von Liebig (1848), C. Ludwig (1849),
A. Fick (1854), and T. Graham (1861) investigated the subject of
osmosis through animal membranes.
The action is curious. In the ordinary nature of things the
sugar would diffuse into the solvent until the whole system had one
uniform concentration. The membrane retards this. If the sugar ,,
cannot get to the solvent, the solvent goes to the sugar — a case of tration of Os-
Mahomet and the mountain. Molecules of sugar and molecules motic Pressure,
of water attempt to pass through the membrane ; the way is open
for the molecules of water, but not for the molecules of sugar. Water can pass
freely both ways. The extra pressure on the solution side of the membrane — the
solution pressure — is supposed to be due to the bombarding of the membrane by
the molecules of sugar. Equilibrium occurs when the number of molecules of
water passing downwards through the membrane is equal to the number passing
in the opposite direction. The resulting pressure is the solution pressure or the
osmotic pressure of the solution.
Let us be perfectly clear about this or we may be led into error. The/ac^ observed
is that the osmotic pressure is the excess of the hydrostatic pressure on the solution
side of a semipermeable membrane over the pressure on the solvent side. The
hypothesis here suggested — often styled J. H. van't Hoff's kinetic theory of solutions
(1886) — is that this pressure is due to the bombarding of the semipermeable membrane
by the dissolved molecules trying to diffuse into the solvent and make solvent and
solution one uniform concentration. The hypothesis was developed in a very im-
portant memoir : The role of osmotic pressure in the analogy between solutions and
gases (1887). ^ The hypothesis has served as a stimulus to much valuable work ; there
are, however, other possible explanations of the phenomenon. The merits of rival
hypotheses cannot be settled by'symposia although discussion may bring fundamental
issues into relief. Harsh experience alone can shatter or estabHsh this interesting
analogy — for comparaison n'est pas raison.
Imagine the experiment arranged a little differently. Suppose the aqueous
540 INORGANIC AND THEORETICAL CHEMISTRY
solution of sugar in the lower part of a cylinder, Fig. 15, to be separated from the
pure solvent in the upper part of the cylinder by a semipermeable membrane Ay
so fitted that it can slide freely up and down the cylinder. The upward osmotic
pressure of the solution will naturally force the piston upwards, and a weight, P,
equivalent to the osmotic pressure of the solution, will be required to keep the semi-
permeable membrane in one fixed position.
Many hypotheses have been suggested to explain the function of the membrane
in osmotic phenomenon, ranging between the purely physical conception which refers
the effect to the passage of the liquid through capillary pores, and the purely chemical
conception of a combination between the membrane and the liquid passing through.
M. Traube (1867), S. U. Pickering (1891), and W. Sutherland (1907) considered the
semipermeable membrane acted as a kind of sieve which allowed the passage of the
molecules of the solvent, but obstructed the passage of the supposed larger mole-
cules of the solute. This hypothesis is now abandoned, for no attempt to distinguish
between true pore diffusion occurring through capillary openings and the so-called
true endosmosis occurring through smaller molecular interstices, has proved success-
ful ; and even in the case where collodion membranes and porcelain plates serve
as partitions, S. L. Bigelow (1907) found that the same laws described the passage
of liquids through both ; there is no experimental evidence clearly distinguishing
between the passage of a liquid through capillaries and through molecular interstices.
According to the solution hypothesis, a substance will pass through a membrane
only if it is soluble therein. According to this hypothesis, if two miscible liquids,
A and B, are separated by a membrane, and the membrane has
r^ p the power to absorb or dissolve only one of them, say A, this
■=: L liquid will be dissolved on one side of the membrane and given
^ up on the other, and if the liquid B is in a closed cell, an hydro-
static pressure will be there developed. The magnitude of this
pressure will depend on the relative attractions or solubility of
A and B in the membrane. If A is soluble and B insoluble or
Soi\fent sparingly soluble, the membrane will be saturated with A on one
A side and supersaturated on the other, and there will be a transfer
Solution, of solvent through the membrane until hydrostatic pressure is
developed sufficient to check the flow. Hints of this hypothesis
Fig. 15.— Osmotic were given by T. Graham, but M. I'Hermite (1855) published
Pressure. the first clear statement of a possible development of osmotic
pressure by a selective action of the membrane, and he gave the
three-Hquid experiment — with chloroform, water, and ether — with the express idea
of demonstrating that a substance which passes through the membrane dissolves
in that membrane. Accordingly, argued M. I'Hermite, there must be a relation
between solution and chemical union ; osmotic phenomena are not the result of a
special force, but rather the effect of forces of affinity similar to those acting in
solutions. L. Kahlenberg (1906) also has sought for evidence in support of the
solution theory of osmosis.
The following is A. C. Brown's modification of M. I'Hermite's three-liquid layers to
illustrate the development of osmotic pressure by the solvent action of the membrane. A
concentrated solution of calcium nitrate is saturated with phenol and the mixture poured
into a tall narrow cylinder. The excess of phenol rises and floats upon the surface of the
calcium nitrate solution. The phenol should not be in larger excess than is required to give
a layer a few millimetres thick. Distilled water saturated with phenol is carefully poured
above the two layers of liquid in the cylinder. The water floats on the surface of the phenol.
The water on both sides of the phenol can traverse the partition of phenol, but the calcium
nitrate cannot pass through. Hence the layer of phenol is a semipermeable membrane.
Mark the level of the layer of phenol in the cylinder by means of a piece of gummed paper.
If the upward motion of the layer of phenol be marked from day to day, it will be found to
rise higher and higher, and finally surmount the rest of the liquid in the cylinder.
Osmotic phenomena can be obtained by continuous and by discontinuous or
porous films. With continuous films it is necessary for the solvent but not for the
=far5
■n
SOLUTIONS 541
solute to dissolve in the membrane ; with porous films it is necessary for the pure
solvent to be adsorbed by pores so small that only the solvent not the solute can
pass through. Benzene, toluene, and pyridine were found by L. Kahlenberg to pass
through a rubber membrane while water does not. Hence rubber probably acts as
a semipermeable membrane to the three first-named liquids, because these liquids
dissolve in the rubber.
W. Ramsay (1894) illustrates the production of an osmotic pressure in solutions
by the following analogy illustrating what has been termed the osmotic pressure of
gases.
A palladium vessel at 250° to 350° is filled, at atmospheric pressure, with nitrogen gas
or with some gas not absorbed by the warm palladium. This vessel is immersed in hydrogen
at a given pressure ; hydrogen gas diffuses through the metal membrane until the
increase of pressure inside the vessel is nearly equal to the outside pressure. In one
experiment, this increase was equivalent to 733 mm. of mercury, which is " regarded as the
osmqtic pressure of nitrogen dissolved in hydrogen." The excess pressure is independent
of the concentration of the hydrogen molecules, for the pressure of the hydrogen is the
same on both sides of the septum. The (osmotic) pressure of the nitrogen is produced by
the bombardment of the nitrogen molecules on the walls of the vessel, while the osmosis of
the solvent hydrogen is possible in virtue of its faculty of dissolving in the metal membrane
under conditions where the solute nitrogen is insoluble.
In ordinary or positive osmosis the direction of flow of the solvent, water, is from
the less towards the more concentrated solution ; in some cases the direction of
flow is from the more to the less concentrated solution ; the phenomenon is then
styled negative or reversed os^nosis. H. Dutrochet first described osmosis with
inorganic membranes, and T. Graham attributed the phenomenon to chemical
interaction between the salt and the membrane. F. E. Bertel, P. Girard, and
H. Freundlich attribute the anomalous efiect to the electrical endosmose ; the flow of
liquid is brought about by a difference in electrical potential, the two ends of the
capillary pores in the membrane becoming oppositely charged. Potential differences
of this kind were shown to exist in animal cells by M. Oker-Blom and W. Ostwald ;
in frog's muscle by A. Briinings ; in vegetable skins by M. Loeb and R. Beutner ; in lung
tissue by R. S. Lillie and P. Girard ; in copper ferrocyanide membranes by R. Beutner ;
and in clay by A. Briinings. According to W. D. Bancroft, the sign of the electric
charge on the membrane is dependent on the absorption of anions or cations.
J. Perrin ascribed the polarization to contact electrification being dependent on the
preponderance of H'-ions or OH'-ions. F. E. Bartel also showed that the appearance
of negative osmosis is dependent on the pore diameter, for the phenomenon occurs
with solutions of magnesium chloride only when the pore diameters are less than
0'4jLt. J. Mathieu found negative adsorption occurs with a number of dilute solutions
when adsorbed by porous plates, membranes, or capillary tubes, such that the
liquid adsorbed by the capillary tubes from iV-solutions was often only .^N ; and
he suggests that if the capillary were fine enough only pure water would be adsorbed.
Summing up the literature on the subject, W. D. Bancroft says : (1) Osmotic phe-
nomena may occur with a porous diaphragm provided we have very marked negative
adsorption and provided the diameter of the pores is so small that the adsorbed
films fill practically the whole of the pores. (2) A porous diaphragm will act as a
semipermeable membrane in case there is no measurable adsorption of the solute
and in case the adsorbed films fill the pores completely. (3) In the usual case of a
semipermeable diaphragm, we do not have a porous diaphragm and the semiper-
meability is due to the fact that the solvent dissolves in the diaphragm while the
solute does not to any appreciable extent under the conditions of the experiment.
(4) A liquid is not to be considered as a porous substance and solubility does not
depend on porosity. Again, A. M. C. Chanoz found that when the two sides of the
membrane differ, as with a skin, differences in the osmosis are obtained depending
on whether a given side of the membrane is in contact with solution A or solution B.
These differences disappear, of course, when the two sides of the membrane are
542 INORaANIC AND THEORETICAL CHEMISTRY
alike, as with parchment paper. It seems probable that the behaviour of the
membrane depends largely on its greater or less permeability.
Animal membranes are objectionable when exact measurements are required,
because to a certain extent the results depend upon the nature of the membrane,
which is not strong enough to withstand the great pressures developed by osmosis ;
and, most serious of all, the membrane is not quite semipermeable, so that an
appreciable amount of, say, sugar does actually pass through. It would therefore
be as profitable to measure the pressure of a gas in a leaking vessel as to try to measure
the osmotic pressure of a solution with a membrane which allows part of the dissolved
substance to pass through. We therefore fall back on artificially prepared mem-
branes. No artificial membrane has been so successful as a film of copper ferro-
cyanide deposited between the inner and outer walls of a porous earthenware pot —
prepared by M. Traube,^ and described in 1867 in his Experimente zur Theorie der
Zellenbildung und Endosmose. The film is made by steeping a clean porous pot in
an aqueous solution of potassium ferrocyanide, rinsing in water, and then* sub-
merging the pot in an aqueous solution of copper sulphate, and subsequently washing
out the soluble salts. The deposition of the copper is symbolized by the equation :
2CuS04+K4reCy6=Cu2FeCy6+2K2S04. The porous pot with its semipermeable
membrane is fitted with a suitable manometer to indicate the pressure. In 1877,
W. F. P. PfefEer made some measurements with cells prepared in this manner.
The apparatus was immersed in a large bath of water to maintain the temperature
constant during the experiment. Analogous experiments were made by H, de
Vries (1878), G. Tammann (1888), P. Walden (1892), etc. Earl of Berkeley and
E. G. J. Hartley (1904) placed a solution of sugar in a porous earthenware pot with
a semipermeable membrane of cupric ferrocyanide, and surrounded the pot with
water. The pressure on the solution was increased until it was just sufficient to
prevent the passage of water into or out of the cell through the septum of the ferro-
cyanide. H. N. Morse (1901-9) employed an apparatus similar to that of W. F. P.
Pfefier, but he improved the quality of the membrane by depositing the cupric
ferrocyanide in the pot electrolytically ; and also improved the joints between the
cell and the manometer ; and the manometer itself.
References.
1 Abbe Nollet, Hist. Acad. Sciences, 101, 1748 ; Lecons de physique experimentale, Amsterdam,
1754 ; R. J. H. Dutrochet, Ann. Chim. Phys., (2), 35. 393, 1827 ; (2), 37. 191, 1828 ; (2), 49. 411,
1832 ; (2), 51. 159, 1832 ; Memoires pour servir a Vhistoire anat. et physiol. der vegetaux et des
animaux, Paris, 1837 ; U agent immediat du mouvement vital, Paris, 1826 ; K. Vierordt, Pogg.
Ann., 73. 519, 1848 ; E. Briicke, ib., 58. 77, 1843 ; P. Jolly, ib., 78. 261, 1849 ; C. Ludwig, ib.,
78. 307, 1849 ; A. Pick, ib., 94. 59, 1855j G. Magnus, ib., 10. 160, 1827 ; E. B. Jerichau, t6., 34. 613,
1835 ; J. Lie big, Ucber einige Ursachen der Saftbewegung in tierischen Organismus, Braunschweig,
1848 ; W. F. P. Pfeffer, Osmotische Untersuchungen, Leipzig, 1877 ; T. Graham, Phil. Trans., 151.
183. 1861.
2 H. Dutrochet, Ann. Chim. Phys., f2), 60. 337, 1835; G. Flusion, ib., (8), 13. 480, 1908;
T. Graham, Phil. Trans., 144. 177, 1854 ; P. Girard, Compt Rend., 146. 927, 1908 ; 148. 1047,
1186, 1909 ; 150. 1446, 1910 ; 153. 401, 1911 ; F. S. Bartel, Journ. Amer. Chem. Soc, 36. 646,
1914 ; 38. 1029, 1916 ; S. L. Bigelow, ib., 29. 1576, 1907 ; 31. 1194, 1909 ; H. FreundUch, Kolloid.
Zeit., 18. 11, 1916; M. Oker-Blom, Pfiuger's Arch., 48. 191, 1901 ; A. Brunings, ib., 84. 241,
1903 ; 117. 409, 1907 ; R. S. Lillie, Amer. Journ. Physiol, 28. 194, 1911 ; P. Girard, Rev. Gen.
Science, 20. 694,1909 ; R. Beutner, Journ. Phys. Chem.,n.SU, 1913 ; S. L. Bigelow,ib., 15. 659,
1911 ; 16. 318, 1912; W. D. Bancroft, ib., 16. 312, 1912; 21. 441, 1917; J. Mathieu, Ann.
Physik, (4), 9. 340, 1902 ; F. Trouton, B. A. Rep., 84. 288, 1914 ; W. Ostwald, Zeit. phys. Chem.,
6. 71, 1890 ; M. Loeb and R. Beutner, Science, 34. 866, 1906 ; J. Perrin, Journ. Chim. Phys.,
2. 601, 1904 ; W. Ramsay, Phil. Mag., (5), 38. 206, 1894 ; W. Sutherland, ib., (5), 44. 493, 1897 ;
S. U. Pickering, Ber., 24. 3629, 1891 ; M. Traube, Oesammelte Abhandlumj, Berlin, 200, 213,
1899; M. I'Hermite, Ann. Chim. Phys., (3), 43. 420, 1855; Compt. Rend., 39. 1177, 1854;
L. Kahlenberg, Journ. Phys. Chem., 10. 141, 1906 ; Trans. Faraday Soc.,Z. 23, 1907 ; J. H. van't
Hoff, Arch. Nierl., 20. 239, 1886 ; Zeit. phys. Chem., 1. 481, 1887 ; Phil. Mag., (5), 26. 81, 1888;
Harper's Scientific Memoirs, 4. 11, 1899; F. Tinker, Nahire, 97. 122, 1916; A. M. C. Chanoz,
Recherches ezperimentales sur les contacts liquides, Paris, 1 906.
' M. Traube, Archiv. Anat. Physiol. Wiss. Medizin., 87, 129, 1867 ; Oesammelte Abhandlungen,
Berlin, 200, 213, 1899 ; W. F. P. Pfeffer, Osmotische Untersuchungen, Leipzig, 1877 ; H. N. Morse
SOLUTIONS
543
and co-workers, Amer. Chem. Journ., 26. 80, 1901 ; 34. 1, 1905 ; 36. 39, 1906 ; 37. 324, 425,
588, 1907 ; 38. 175, 1907 ; 39. 667, 1908 ; 40. 194, 1908 ; 41. 257, 1909 ; Earl of Berkeley
and E. G. J. Hartley, Proc. Boy. Soc, 73. A, 436, 1904 ; Phil. Trans., 206. A. 481, 1906 ; Earl of
Berkeley, E. G. J. Hartley, and C. V. Burton, ib., 209. A 177, 1908 ; 218. A, 295, 1919 ; H. de
Vries, Arch. Neerl, 13. 344, 1878 ; Zeit. phys. Chem., 2. 415, 1888 ; P. Walden, ib., 10. 619, 1892 ;
G. Tammann, M&m. Acad. St. Petersburg, 35. 169, 1887 ; Wied. Ann., 34. 299, 1888.
Manometer. -^
§ 8. The Osmotic Pressure of Dilute Solutions and the Gas Laws
Every formula obtained by the application of thermodynamical considerations alone
to a mixture or solution remains the same, no matter what assumptions be made regarding
the molecular condition of the substances. Consequently, thermodynamics alone cannot
decide whether solution is attended by a chemical change in the molecular state of the dis-
solved substance or otherwise.- — P. Duhem (1894).
J. H. van't Hofi's kinetic theory of osmotic pressure (1887) i emphasizes the
analogy between the process of vaporization and the process
of solution. In a solution the dissolved substance is dis-
tributed throughout the whole bulk of the solvent, and the
solvent plays the part of so much space. The vapour pressure
of a liquid in space will thus be represented by the osmotic
pressure of a solution. In the words of A. Eosenstiehl, the
osmotic pressure is analogous to the elastic force of vapours.
Just as the closed space above a liquid becomes saturated
with vapour, so does a solvent in contact with the solute form a
saturated solution. An increase of temperature augments the
vapour pressure of a liquid, and also the osmotic pressure of a
solution.
I. The relation between osmotic pressure and the concentra-
tion of the solution — Boyle's law. — W. Pfeffer in his Osmotische
Untersuchungen (Leipzig, 1877) obtained some data with the
apparatus which J. H. van't Hoff (1887) utilized, with remarkable
cleverness, in developing what he called " the role of osmotic
pressure in the analogy between solutions and gases." The ex-
perimental data showed that the osmotic pressure is very nearly
proportional to the concentration of the solution; otherwise
expressed, the osmotic pressure appears to depend upon the j, ,g _Meas re-
degree of crowding of the molecules of the dissolved substance, mentof Osmotic
Instead of repeating Pfeffer's measurements, a selection from Pressure,
some later determinations with solutions of glucose (sugar) by
H. N. Morse (1907) can be quoted (temperature nearly 0°, rounding off the decimals
to the nearest tenth of a unit) :
Concentration
. 0-1
0-2
0-3
0-4
0-5
0-6
1-0
Osmotic pressure .
. 2-4
4-7
7-0
9-3
11-7
141
23-7 atm
Equivalent gas pressure .
. 2-2
4-5
6-7
8-9
IM
13-4
22-3 atm
In dealing with the concentration of solutions, it wiU be well to adopt the same unit
of comparison as that employed in dealing with gases, i.e. the molecular weight of
the solute expressed in grams per 22*3 Utres of solution at normal temperature and
pressure. H. N. Morse found that his direct measurements of osmotic pressure
came out best when referred to a constant volume of the solvent, not to the volume
of the solution.
Assume that a gram-molecule of glucose (180) were it a gas would occupy 22*3 litres.
Hence, O'l gram-molecule will occupy 2*23 litres. By choosing the concentration so that in
Boyle's relation, PF = constant, a solution containing a molecular weight expressed in grams,
per 22'3 litres, has a concentration of 22*3 units when P = l, we get from Boyle's law
P-f-C' = 22-3. The concentration, it will be remembered, is inversely proportional to the
volume. Hence for a concentration 0-1, we get P = 2-23, for 0=0-2, p=4-46, etc.
544 INORGANIC AND THEORETICAL CHEMISTRY
The " equivalent gas pressure " is here calculated on the assumption that a
" sugar gas " obeying Boyle's law really exists. The results are plotted in Fig. 17.
The deviation of the osmotic pressure curve from the dotted curve emphasizes the
fact that the deviations of the osmotic from the equivalent " gas pressures " grow
larger with increasing concentrations, and hence exact proportionality occurs
on& when the solutions are very dilute. For dilute solutions, the osmotic
pressure is nearly proportional to the concentration, or, as W. Ostwald puts it,
" the osmotic pressure of a sugar solution has the same value as the pressure the
sugar would exert if it were contained, as a gas, in the volume occupied by the
solution — of course assuming Avogadro's rule." This is another way of saying
that the relation between the osmotic pressure of a solution and its concentration
has the same form as Boyle's law for gases.
The analogy does not work out so well for concentrated solutions as with dilute
solutions — ^possibly owing to the disturbing effects of overcrowding produced by :
(1) molecular attraction between the molecules of the dissolved substance ; (2) the
volumes of the molecules themselves. The two effects for gases were discussed
when dealing with Boyle's law for gases. J. D. van der Waals' corrections for the
gas equation pv—RT, involves the introduction of terms for the mutual attraction
of like molecules and for the space occupied by the molecules, and the corrected
^^^^^^^^^^^^^^ equation takes the form {p-{-alv^)(v—h)=RT, and
Z4 iiiiniiniiiiiiiiiiiiniiiiiiiMi iiii im ^y. regarding v in the equation pv=RT as the
immimiii[[i[[lllllllllii|[ni[l[[[l volumc of the solvcut uot of the solution, H. N.
Morse really corrected the equation for the space
/6 IIIIIIIIIIIIIIIIIIIBttttlllllltffllllllll^^ occupied by the molecules of the solute as J. D.
van der Waals' did for gases. And (3) the mutual
Q, /2 |||||l|l||||||||||l^?Sffi?i"Sllllllll^ attraction between the molecules of the solute and
solvent. On account of the enormous number of
molecules of the solvent which are present, each
molecule of the solute is probably completely
surrounded by molecules of the solvent, and the
resultant of all the forces due to the solvent, act-
ing upon each molecule of the solute, is zero.
Co ntr hi "^^^ velocity of the solute molecules impinging on
Fig. 17.— Osm^o^tic Pressure and ^^^ semipermeable membrane is not affected pro-
Concentration. vided the solution is so dilute that the difference
in the concentration of the molecules of the
solvent on the two sides of the membrane is negligibly small. 0. Stern
diminishes the factor a of J. D. van der Waals' equation by a factor aj2(a^o— a?),
expressing the attraction between the molecules of solvent and solute which,
so far as osmotic pressure is concerned, acts in the opposite direction to the
attraction ai between the molecules of the solute itself, for this attraction pulls the
molecules of the solute away from the solvent. The term Xq—x represents the
difference between the concentration of the solvent outside the membrane and
in the solution itself. The term 6 of J. D. van der Waals' equation is also increased
by a factor hi2(xQ—x), because the repulsive force 612 between the molecules of
solvent and solute which makes the solute behave as if the molecular volume bi of
the solute is smaller than it really is. 0. Stern's equation is then ;
There are thus four constants in the equation, and since the new constants have
to be evaluated from the experimental data, better agreement is to be expected
than with an equation including two constants. If two miscible liquids with critical
states not very far removed from one another be under investigation, ai2 and &22
are of the same order of magnitude as % and bi. The difference between the con-
centrations of the pure solvent and that which it has in solution is nearly identical
SOLUTIONS
545
with the concentration of the solute, then {xq—x)Iv will be nearly 1/^), the concen-
tration of the solute ; the term Xq—x then cancels out, so also do the terms involving
a and h, and the gas in solution will then obey the ideal gas law more nearly than
it does in the gaseous state, as 0. Stern found to be the case with solutions of carbon
dioxide in methyl and ethyl alcohols.
2. The relation between osmotic pressure and temperature. — Charles' law. —
W. F. P. Pfeffer's measurements on the influence of temperature also showed that
the osmotic pressure is proportional to the absolute temperature, which means that
the relation between the osmotic pressure and temperature of a given solution has
a formal analogy with Charles' law for gases. In illustration, some results by
H. N. Morse (1911) for unit concentration may be quoted :
Temperature .
0°
5°
10°
15°
20°
25°
Osmotic pressure
24-8
25-3
25-7
26-2
26-6
27-0 atm
Equivalent gas pressure .
22-2
23 0
23-4
23-8
24-2
24-5 „
5" to" /5"
Temperature.,
— Osmotic Pressure
Temperature.
20''- 25'
and
The " equivalent gas pressure " is here calculated on the assumption that a
" sugar gas " obeying Charles' law really exists. These numbers are plotted in
Fig. 18, and the graphs show the proportionaHty between osmotic pressure, P,
and temperature ; P/T=constant. The space between the two curves represents
the deviation of the observed osmotic pressure, from the pressure calculated on the
assumption that the dissolved substance behaves as if it were a gas. W. F. Magie
has also studied the relation between osmotic pressure and temperature.
An experiment due to C. Ludwig (1856) and investigated by C. Soret (1881) ^
— Soret's phenomenon — may be cited in illustration of the applicability of the
gas laws to dilute solutions. If a solution be
kept at one uniform temperature, it will in time
become homogeneous ; on the contrary, C. Soret
showed that if the two ends of a tube containing
a homogeneous solution be kept at difierent
temperatures, the concentration of the solution
at the cooler end will increase, and decrease at
the warmer end. The warmer solution becomes Fig. 18.
more dilute because the osmotic pressure of the
warm solution is greater than that of a cold solu-
tion ; and conversely. E quilibrium will be established when the osmotic pressure in all
parts of the solution is the same. If the warm end of the tube be 50° hotter than
the cold end, then, if Charles' law applies to solutions, the cold solution should in-
crease in concentration 273rd more than the warm solution per degree difference
of temperature, and hence the colder solution should be ^V^ more concentrated
than the warm one. N. M. Hopkins (1905) claims to have observed a difference of
14*03 per cent, in the density of a dilute solution of copper sulphate in a tube 80° at
the one end and 20° at the other when the theoretical difference by Charles' law
was 14:3 per cent. The phenomenon also appears to be connected with an observa-
tion made in 1799, by N. Leblanc, to the effect that if crystals of a salt are placed
some at the upper and some at the lower part of a cylinder containing a saturated
solution of the same salt, the lower crystals grow larger at the expense of the upper ;
and likewise also the upper portion of the crystals at the bottom of a liquid decreases
while the lower portion increases.
3. Avogadro's hypothesis applied to solutions. — If P denotes the osmotic
pressure of a solution, and F the volume containing one gram-molecule of the
solute, PV=RT (where R is a constant), and for a solution of volume V con-
taining n gram-molecules of the solute PV=nRT. The volume F is to be
regarded as the molecular volume of the solution only when, as J. H. van't Hoff
(1887) said : "the volume occupied by the molecules of the solute is negligible in
comparison with the volume of the solution." Hence, at constant temperature,
n molecules of a solute in unit volume of a dilute solution (for which F=l) have the
VOL. I. 2 N
546 INORGANIC AND THEORETICAL CHEMISTRY
same osmotic pressure ; otherwise expressed, equal volumes of solutions con-
taining the same number of solute molecules have the same osmotic pressure ;
and conversely, solutions, at the same temperature and the same osmotic
pressure contain the same number of molecules of the dissolved substance per unit
volume. There is a striking resemblance between this assumption and Avogadro's
hypothesis for gases, and it harmonizes with a number of facts. The principle
can be applied to measure the molecular weight of substances in solution, for the
term " number of molecules " is used in the same sense as the term is used in
stating Avogadro's hypothesis for gases : Equal volumes of all gases at the same
temperature and pressure contain the same number of molecules ; or, conversely,
at any assigned temperature the pressure of a gas depends on the number of mole-
cules and not on their kind. Hence van't Hoff's hypothesis assumes that the
osmotic pressure and related properties — vapour pressure, freezing point, and boiling
point — of dilute solutions (1) depend upon the number of molecules of solute dissolved
in unit volume of the solution, and are independent of (2) the chemical nature of the
solvent and solute, and (3) of the relations between solvent and solute. In contradis-
tinction to additive properties like the specific gravity of mixtures, colligative
properties depend merely on the relative nimiber of molecules present and not on
the kind of molecules — e.g. the osmotic pressure, freezing and boiling points of
solutions.
Examples." — (1) An aqueous solution of 1 '0047 grams of orthoboric acid per litre at
0° has an osmotic pressure of 27*3 cm. of mercury. What is the molecular weight of the
acid ? Since one gram-molecule of a substance in the gaseous state occupies 22-3 litres
at 0° and 760 mm., we have here to find what weight of substance will occupy 22*3 litres
at 0° and 760 mm., given 1-0047 gram occupy 1 litre at 0° and 273 mm. pressure. Obviously,
1-0047 gram will occupy 0-361 litre at 0° and 760 mm. ; and if 0-361 htre weighs 1*0047
gram, 22-3 litres will weigh 62 grams at the same temperatiu-e and pressure. Hence the
molecular weight of the given acid is 62 ; this agrees with the formula B(0H)8 for orthoboric
acid.
(2) A two per cent, solution of cane sugar has an osmotic pressiire of 1016 mm. at 15° ;
what is the molecular weight of cane sugar ? 100 c.c. at 1016 mm. pressure becomes 0-126
litres at 760 mm. pressure and 0°, and 0*126 litre corresponds with 2 grams of cane sugar.
Hence 22-3 litres will have 355 grams at the same temperature and pressure. The mole-
cular weight of cane sugar therefore approximates 355. The true number is 342 for
In^ place of using the equation PV=RT, K. Jellinck^ used J. D. van der Waals*
equation in his study of osmotic pressure from the kinetic point of view, and
F. Tinker used C. Dieterici's equation.
There are so many experimental difficulties involved in the direct measurement
of osmotic pressure that the method is rarely, if ever, employed directly for mole-
cular weight determinations. As in the kinetic theory of gases, it can be shown,
with the above assumptions, that the kinetic energy of the solute in dilute solutions is
equal to that of a gas at the same temperature and pressure ; and that with the same
average kinetic energy, the number of impacts depends only on the concentration,
and is independent of the presence of the solvent. As a corollary, too, it follows
that the mutual exchange of energy at each colHsion, when equilibrium is estab-
lished between the solvent and solute, will make the average molecular kinetic
energy of solvent and solute the same. Hence, said W. Ostwald (1890),^ the
kinetic energy of the molecules of a liquid is the same as that of the molecules of
a gas at the same temperature and pressure.
E. W. Washburn (1915) ^ has drawn attention to the fact that the term osmotic pressure
is loosely employed to designate three quite different ideas : (1) The osmotic pressure of a
solution is really a physical quantity and not a real pressure, and is the difference in the
pressure which must be established upon solution and pure solvent in order to make the
tendency of the solvent to escape as vapour the same for both ; it is the difference of pressure
necessary to prevent osmosis through a perfect semipermeable membrane. For dilute
solutions, the osmotic pressure at the limit is equal to CRT, where C denotes the concen-
tration of the solute. As the concentration increases, the osmotic pressure increase towards
infinity, as illustrated in Fig. 19. (2) In virtue of unordered heat motions, the molecules
SOLUTIONS
547
of a solute in a solution may be considered as exerting a certain pressure, called the thermal
pressure. For dilute solutions the thermal pressure will be equal to CRT, but as the con-
centration C increases, the thermal pressure increases towards a large but finite limit, as
illustrated in Fig. 19. (3) The partial pressure exerted by the molecules of the solute in
a solution against a membrane permeable only to the solvent is called the diffusion
pressure. For dilute solutions the diffusion pressure is equal to CRT, and as the concentra-
tion C increases, the diffusion pressure increases to a finite definite limit whose value depends
upon the temperature, pressure, and the attractive forces extended on the molecules of the
solute, in the interior of the liquid. See Fig. 19.
The effect of the heat of dilution on osmotic pressure. — Describing osmosis
in the language of free energy, the osmosis is attributed to the difference which
exists between the free energy of the solvent and solution ; and diffusion is
an effect of the free energy driving the solvent from the region where the free
energy is greatest to the solution where the free energy is least. The process of
diffusion continues until the free energy has fallen to the value characteristic of a
solution with one uniform composition. When solution and solvent are separated
by a semipermeable membrane, the solvent will travel into the solution until the
free energy of both is the same. The solute cannot travel through the membrane
to the solvent, and therefore the system can never have one uniform composition ;
for equilibrium, however, the free energy of solution and solvent must be the same.
1-0
^ 0-8
t? 07
^ 0-3
J 0-1
0
/
"f
^
—
c
■""
—J — hW
To oo_J
tl
P
[^
^^'''
UOI^
4 .^
>f:.
?^
^ss^r.
h
m
1
yl^
f
i
7/
i
//
^
f Pressure -
-^
0 I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20
Fig. 19. — Washburn's Illustration of Solution Pressure.
The free energy of the solvent is the greater, and the free energy of the solution
can be augmented only by increasing the pressure on the solution ; consequently,
solvent will diffuse through the membrane to the solution until the hydrostatic
pressure — osmotic pressure — required to make the free energy of solvent and solution
the same, is attained. The change in the free energy of an isothermal reversible
process is given by the expression W =^Q'\-T{d'W jdT), where Q is the heat evolved
when the solution is diluted. The diminution in the free energy which occurs
when a gram-molecule of the solvent passes reversibly and isothermally into a large
volume of solution through a semipermeable membrane is therefore
VY^Q-^T
d(PV)
dT
where P is the osmotic pressure and F is the increase in volume. This relation
represents the temperature coefficient of the osmotic pressure. If Q, the heat of
dilution, is zero, the preceding expression reduces, on integration, to PF/ J=constant,
and hence PV=RT is true only when the heat of dilution is zero. For all but
extremely dilute solutions, Q is not constant, and the osmotic pressure cannot be
calculated from PV—RT, but rather from PV=RT-}-Q. Before the above equation
can be integrated, it is necessary to know how Q varies with temperature (say,
Q=A'\-BT-\~ . . .). When molecular weights are calculated from the osmotic
548 INORGANIC AND THEORETICAL CHEMISTRY
pressure or related phenomena, neglect of this factor — heat of dilution — may give
quite erroneous results.
In his work on the thermodynamics of osmotic pressure, J. H. van't Hoff
assumed that the solutions were so dilute that no thermal change occurred on
further dilution. W. D. Bancroft (1905) ^ has shown that the osmotic pressure
is abnormally high when heat is evolved on dilution, and similarly the lowering
of the freezing point of such a solution will not be so great as when the heat of
dilution is zero. In the case of the metals of the alkalies and alkaline earths
dissolved in mercury, molecular weights equal to half the atomic weights are obtained
by formulae in which the heat of dilution is assumed to be zero ; and with sulphuric
acid, values ranging from 57*7 to 11"7 when the concentration of the solution ranges
from 5'6 to 68'5 per cent, respectively — the anomaly of a decreasing molecular
weight with increasing concentration disappears if the heats of dilution are included
in the computation.
References.
1 J. H. van't HofF, Arch, Nierl, 20. 239, 1886 ; Zeit. phys. Chem., 1. 481, 1887 ; Phil. Mag.,
(5), 26. 81, 1888; Harper's Scientific Memoirs, 4. 11, 1899 ; A. Rosenstiehl, Compt. Rend., 70.
617,1870; 152. 1305,1911; W. F. Magie, P%5. Rev., (1), 35. 272, 1912; (2), 10. 64,1917;
H. N. Morse, Amer. Ghent. Journ., 26. 80, 1901 ; 28. 1, 1902 ; 29. 137, 1903 ; 32. 93, 1904 ; 34.
39, 1905 ; 37. 324, 425, 558, 1907 ; 38. 175, 1907 ; 39. 667, 1908 ; 40. 194, 266, 325, 1908 ;
41. 92, 557, 1909 ; 45. 91, 237, 283, 517, 554, 1911 ; 48. 29, 1912 ; 0. Stem, Zeit. phys. Chem.,
81. 441, 1912.
2 C. Ludwig, Sitzher. Akad. Wien, 20. 539, 1866 ; C. Soret, Arch. Sciences Geneve, (3), 2. 48,
1879 ; Ann. Ghim. Phys., (5), 22. 293, 1881 ; jST. M. Hopkins, Experimental Electrochernistry,
London, 28, 1905 ; N. Leblanc, Journ. Phys., 33. 376, 1800 ; C. L. Berthollet, Essai de
statique chimique, Paris, 1. 49, 1803 ; L. Gmelin, Handbook of Chemistry, London, 1. 112, 1848 ;
S. Arrhenius, Oef. Svensk. Forh., 61, 1894 ; R. Abegg, Zeit. phys. Chem., 26. 161, 1898.
3 K. JelUnck, Zeit. phys. Chem., 92. 169, 1917 ; F. Tinker, Phil. Mag., (6), 33. 428, 1917.
* W. Ostwald, Lehrbuch der allgemeinen Ghemie, Leipzig, 1. 699, 1903 ; P. Fireman, Journ.
Phys. Chem., 6. 636, 1902.
* E. W. Washburn, An Introduction to the Principles of Physical Chemistry, New York, 1915.
8 W. D. Bancroft, Journ. Phys. Chem., 9. 216, 1905 ; J. E. Trevor, ib., 10. 400, 1906 ; 12.
141, 1908 ; T. Ewan, Zeit. phys. Chem., 14. 409, 1894 ; 31. 23, 1899 ; H. L. Callendar, Proc.
Ray. Soc, 80. A, 466, 1908 ; A. Gouy and C. Chaperon, Ann. Chim. Phys., (6), 12. 1384, 1887 ;
J. H. van't Hoff, Etudes de dynamique chimique, Amsterdam, 187, 1884; F. Tinker, Phil. Mag.,
(6), 33. 428, 1917.
§ 9. The Relation between the Vapour Pressure o£ a Solution and the
Molecular Weight o£ the Solute
The mutual action of two liquids is observable when a mixture of say alcohol and ether
is subject to experiment in the vacuum of a barometer column, for the mixture depresses
the colunm less than either component alone.— C. L. Berthollet (1803).
M. Faraday ^ knew, in 1822, that the vapour pressure of a solution is lower
than the vapour pressure of the pure solvent ; and C. L. Berthollet's experiment
in 1803 shows that at a given temperature the vapour pressure of a solution of
ether in alcohol is less than that of either ether or alcohol alone. A. Wiillner
discovered the important fact experimentally, in 1858, that the lowering of
the vapour pressure of a solution is proportional to the quantity of
substance in solution provided that the dissolved substance is non-volatile.
This is sometimes called Wiillner's law. A. Wiillner worked with aqueous
solutions, but the abnormal behaviour of such solutions as a result of the
extraordinary properties of water prevented him decisively demonstrating
the generalization. F. M. Raoult (1887) worked with non-aqueous solutions
and succeeded better than his predecessors. The phenomenon can be illustrated
by introducing about 2 c.c. of water, 2 c.c. of a 2 per cent, solution of potassium
iodide, and 2 c.c. of a 4 per cent, solution of the same salt into the Torricellian
SOLUTIONS
549
vacuum of each of three barometer tubes mounted within a hot jacket. The more
concentrated solutions will depress the mercury most, the less concentrated solution
will depress the mercury more than water alone, but less than the more concentrated
solution. Some observed values 2 of the difference between the vapour pressures
of water and of solutions containing w grams of potassium iodide, KI, in 100 grams
of water are : ^
10-33
30-71
54-75
71-54
111-14
134-93
169-14
200-25
15-6
47-6
910
21-7
191-3
231-6
283-4
321-3
W. W. Reed (1913) has an experiment to show the lowering of the vapour pressure of a
solution. Three similar thermometers are arranged like the two in a wet-and-dry bulb
hygrometer. The wick of one bulb dips in the given solution. The thermometer with the
wick dipping in the pure solvent reads lower than the one with its wick dipping in the
solution, and both read lower than the dry-bulb thermometer.
Suppose a solution A, Fig. 20, confined in a long-stemmed tube, as illustrated
in the diagram, be separated by a semipermeable membrane M from th^ pure
solvent. Let all be confined in a closed vessel. Osmotic pressure will force the
solution to rise in the narrow tube to a height h, until the whole system is in equili-
brium. Let ^j,. denote the vapour pressure of the solution in
the narrow tube, and p the vapour pressure of the pure
solvent in the outer vessel. The vapour pressure of the solu-
tion at the surface in the narrow tube must be equal to the -j;
vapour pressure of the solvent at the same level, otherwise
distillation would take place either to or from the surface of
the liquid in the narrow tube. In either case, there would be
a constant flow of liquid respectively to or from the vessel A
through the semipermeable membrane in order that h may have
a constant value. Otherwise expressed, perpetual motion would
occur. By the law of excluded perpetual motion this is not
possible, hence the vapour pressure of solution and solvent at ^—
the upper level of the solution in the narrow tube must be
the same. The vapour pressure of the solvent at the level a |EE^^
will be equal to the vapour pressure of the solvent at the lower {rSc/ven/-.-
level b, less the pressure of a column of vapour of height h per Fig. 20. — Dia-
unit area. Hence, the hydrostatic pressure of the liquid column grammatic.
h measures the osmotic pressure ; and the hydrostatic pressure
of the vapour column h, measures the difference in the vapour pressure of solution
and solvent. Since the height h is determined by the osmotic pressure, which,
in turn, is determined by the concentration of the solution, there must be a
simple proportionality between the osmotic pressure or concentration of the
solution and the lowering of the vapour pressure (p—jps), for the vapour of
the solution will be in equilibrium with the vapour of the solvent at such a height
h above the surface of the solvent, that the hydrostatic (osmotic) pressure of the
column of liquid will make the vapour pressure of the solution equal to that of
the solvent, so that the height h measures the lowering of the vapour pressure,
p—psr and also the osmotic pressure, P.
If the tube h has unit sectional area, p—ps ' P= Weight of column of vapour :
Weight of equal column of liquid ; or, if w and W respectively denote the weights
of columns of solution and vapour of height ^, and unit sectional area, then,
p—ps : P=W : w. Let one gram-molecule M of vapour occupy a volume v, and
let V denote the volume of the solution also containing a gram-molecule of the
solute, then pv=^FV for dilute solutions, and hence, v=VPIp, or the volume
occupied by a gram-molecule of the vapour is P/p times the volume of the solution
containing a gram-molecule of the solute. A volume v of the vapour containing a
gram-molecule o'f the solvent will weigh M grams, or Tf =ilf . If n gram-molecules
of the solute are dissolved in N gram-molecules of the solvent, a volume V of the
550 INORGANIC AND THEORETICAL CHEMISTRY
solution containing one gram-molecule of the solute will weigh MNJn grams, and a
volume V or VP/p of the solution will weigh w=PMN/np grams. Substituting
these values of w and W in the above proportion, and reducing the resulting expres-
sion to its simplest terms, there remains :
p '^N
In concentrated solutions when n, the number of molecules of the solute, is equal
to N, the number of molecules of the solvent, n=N, and ps/p is zero, which is im-
possible, because the vapour pressure of a concentrated solution will always have
some numerical value. F. M. Raoult therefore changed the preceding expression
to
P-Vs__ n
p N+n ^^
where n denotes the number of gram-molecules of the solute, and N the total
number of gram-molecules of solvent. In words, the relative lowering of the vapour
pressure of a dilute solution is proportional to the relative number of molecules
of the solute and solvent. F. M. Raoult 3 found this rule valid for dilute solu-
tions— Raoult 's vapour pressure law — and for more concentrated solutions
{jp—ps)lp=^hn{(N-\-n) represented the results more exactly.
Instead of starting from J. H. van't HofE's rule applicable to dilute solutions,
G. N. Lewis * argues that the results are incompatible with solutions of finite con-
centration, and he prefers to start from Raoult's law, which is the only law of
dilute solutions which ever holds in concentrated solutions. He therefore defines
a perfect solution as one which follows this rule : At a constant temperature the
vapour pressure of the solvent is proportional to its molecular fraction n/(N-\-n).
Thus, a solution which contains n=0'l gram-molecule of the solute and iV=0'9
gram-molecule of solvent, w/(iV-f-w)=0*9 ; and the vapour pressure of the solvent
p should be nine-tenths of the vapour pressure in the pure state. It is assumed
that concentrated solutions which deviate very much from this rule are those in
which solvent and solute form complex compounds either with themselves or with
one another. From this relation it is then possible to deduce expressions for the
osmotic pressure and related properties of solutions by the regular methods of
thermodynamics. It will be observed that Raoult's rule reduces to Henry's law
when appUed to gases when the concentration is expressed as a molecular fraction
n/(N-\-n). Henry's law is symbolized p=Knl(N-\-n), where the constant K is
the vapour pressure of the pure gas, for p=K, if the amount N of the solvent
is zero.
The relation between the relative lowering of the vapour pressure and the
osmotic pressure. — A relation between the osmotic pressure and the vapour pres-
sure was deduced by J. H. van't Hoff in 1886, and the demonstration was improved
in 1889 by S. Arrhenius.^ Many other modifications have been suggested. Let p
denote the vapour pressure of water, and ps that of the solution. S. Arrhenius
found the osmotic pressure, P, to be
P=^-^ log ^
where u is the specific volume of the solution, s the specific volume of the vapour
of the solvent at the pressure p. W. Spens obtained a somewhat similar expression
with u representing the increment in volume of a large mass of solution when unit
mass of solvent is added ; and Earl of Berkeley and E. G. J. Hartley with u repre-
senting the specific volume of the solvent. A. W. Porter, J. E. Trevor, J. J. van
Laar, H. Boldingh, G. N. Lewis, E. W. Washburn, and H. L. Callendar have
also deduced expressions for the relation between the vapour pressure and osmotic
pressure of solutions.
SOLUTIONS
551
In the demonstration that {p—p8)lp=i^lN, the molecular weight of the solvent
is assumed to be the same in vapour and in solution. If n represents the number
of molecules of the solute of molecular weight m, then n=w/m, where w denotes
the weight of the n molecules ; similarly, N=WIM, where W denotes the weight
of N molecules of the solvent. Substitute these values of n and N in (p—p8)lp
=n/N, and solve for m ; there remains m=wMplW{p~ps)' Let If =100 grams,
then for any given solvent the constant MpjW can be represented by h. Con-
sequently, if w denotes the weight of substance in grams dissolved in 100 grams of
solvent, and if p' denotes the resultant lowering in the vapour pressure of the
solvent p—pg—p\ the molecular weight of the solute in dilute solution is :
Molecular weight ::= A;
(2)
lilllllllllillllllillllllllllllllllllllll{
let ^t i li i^
fy^i H 1
p ^
=1 E^
4 li
i 1
-E r<
-E E-
where k is the so-called vapour pressure constant whose numerical value depends
upon the particular solvent used ; k therefore represents the diminution in the
vapour pressure which occurs on dissolving one
gram-molecule of the solute in 100 grams of the
solvent.
Examples.' — (1) A solution of iodine in ether showed
a difference of p' = l'4: cm. at 15° in the levels of the
mercury in the two legs of the differential manometer. The
solution contained w = l '139 grms. of iodine per 100 grms.
of ether. The value of h — the reduction in the vapour
pressure produced by one gram of the solute in 100 grms.
of solvent — -for ether at 15° is 260. Hence, from formula
(1), 260x7-139-h7-4=250. Theory gives l2 = 254.
(2) A solution containing the equivalent of 7-435 grms.
of sulphur per 100 grms. of carbon disulphide lowered
the vapour pressure 4-9 cm. The constant for carbon
disulphide at 15° is 171-5. Hence, the molecular weight of
sulphur under the conditions of the experiment is 260-2.
Theory for 83 = 256.
The method for determining the molecular weight
of a substance from direct measurements of the
lowering of the vapour pressure is of great theoretical
interest, but in practice it is seldom employed,
because some of the related properties of solutions
are more amenable to measurement — freezing point,
boiling point, etc.
The practical methods for determining the lowering
of the vapour pressure of a solvent are classed as :
(i) dynamical, or (ii) statical.e W. Ostwald's ^Y" "^Method 7f Ist^SL^in^'^Xir
namical process (1891) consists in determining the cular^eights '"^*^ ^^
lowering of the vapour pressure from the ratio of
the loss of weight of the solvent to the gain in weight of an absorption bulb when
a slow current of air is passed in order through solution, solvent, and absorption
bulb. The barometric vacuum process — Fig. 21 — is a statical process.
The following modification is one of the most convenient forms of the statical process :
Two small c.c. flasks are fitted to a differential manometer, as illustrated in Fig. 21. One
flask contains the pure solvent, and the other a known weight of the solute— say from 2 to 4
grms.- — and three-fourths filled with solvejit. Each flask is heated by placing it in a beaker
containing warm water so that the vapour of the solvent can sweep air out of the system
via the three-way cock G. When all the air is expelled, the stopcock is put in communica-
tion with the flask and both flasks allowed to cool to any desired temperature by placing
them in a suitable bath. The difference in the levels of the mercury in the two tubes
is measured. The reservoir of mercury is again adjusted until the level of the mercury in
the bulb is below the stopcock ; and the flask containing the solution is weighed. Other
manipulation details should be obvious. The subsequent procedure can best be illustrated
by example.
552 INORGANIC AND THEORETICAL CHEMISTRY
The methods used in hydrometry to determine the pressure of aqueous vapour
in the air, and the relation between this pressure and that of saturated water vapour
at the same temperature can be employed to measure the vapour pressures of
solutions and solvent (G. Guglielmo, 1901) ; vapour pressures can also be determined
from the rate of evaporation of solutions (H. Kronberg, 1893)7
The osmotic pressure of solutions of any concentration.^ — Just as the attempts to
adapt the relation pv=RT to gases has led to about fifty more or less unsatisfactory
modifications of the formula, so the attempt to adapt the gas equation pv — BT to con-
centrated solutions has led to quite a number of tentative equations some of which are
modelled after J. D. van der Waals' well-known gas equation. While the law holds good
for very dilute solutions in which there is no polymerization, dissociation, or formation of
complexes, yet it can be regarded only as a limiting law which is approached as the solution
becomes more and more dilute, and it results from an unknown general equation by
omitting certain terms which become negligibly small as dilution increases. Accordingly
E. W. Washburn (1910) * has attempted to reconstruct this equation in the following
manner. For dilute solutions, PVs—nRT, where Vg denotes the volume of the solution.
If N gram-molecules of the solvent have n gram-molecules of the solute in solution and
the molecular volume of the solvent is V, V8=NV ; consequently, NFV^nET. If
the composition of the solution be expressed as a gram-molecular fraction of the solute
such that m' represents the molecular fraction of the solute and N' that of the solvent,
n'-{-N' = l ; n=n'/{N'-\-n') ; and N=N'/{N' -{-n') ; hence, it follows that
V N'
This relation is true only at the limit to which the osmotic pressure approaches when the
concentration of the solution approaches zero. Hence, by differentiation, dw'-fdiV' = 0,
and dn' = —dN' ; accordingly
/RT\/-dN'\
which means that the addition of dN' molecules of a solute to a solvent raises the osmotic
pressure dP. If the molecular volume of the pure solvent be Vq, under a standard pressure,
and if a is the coefficient of the compressibility, F= Fo(l +aP), and integration of the above
equation furnishes
/aP^\
^+(-^j=-^ log(l-iV')
which resembles J. J. von Laar's equation, and represents the osmotic pressure of solutions
for all concentrations. The relations between osmotic pressure and vapour pressure, boiling
point, and freezing point can be derived from this equation in the usual way. For dilute
solutions, if a be very small, the solution can be assumed to be incompressible, without
committing sensible error ; and if N' be small log {l—N') will be virtually equivalent to
N' ; and consequently, P= —N'RT/V. A deduction from this equation has been tested,
with satisfactory results, by J. von Zawidzky (1900) for concentrations ranging from zero
to infinity for over a dozen different binary mixtures.
G. F. Fitzgerald (1896) ^ has pointed out that the kinetic theory of evaporation
describes the lowering of the vapour pressure of a solution in this manner : The
presence of non- volatile molecules of the solute at the surface of the solution hinders
the egress, but does not prevent, or possibly facihtates, the return of the volatile
molecules. The gas-analogy hypothesis of osmotic pressure assumes that the
presence of a body in solution produces no effect or the same effect on the ingress
or egress of the molecules of the solute, for the surface of a liquid with a non- volatile
solute is a perfect semipermeable membrane — water molecules can pass through
the surface freely, but the molecules of the solute cannot. It is a remarkable co-
incidence that with dilute solutions the osmotic pressure is roughly the same as that
which would be produced by the molecules of the solute if it were in the gaseous
state, but, as previously indicated, the dynamical theory of the two must be in-
trinsically different.
References.
1 M. Faraday, Ann. Chim. Phys., (2), 20. 324, 1822 ; J. W. le Grand, ib., (2), 53. 423, 1833 ;
(2), 59. 423, 1835; A. Wiillner, Pogg. Ann., 103. 529, 1858; 105. 85, 1858; 110. 564, 1860 ; W. W.
SOLUTIONS 553
Reed, Chem. News, 107. 64, 1913 ; F. M. Raoult, Ann. Chim. Phys., (6), 15. 275, 1888 ; Compt.
.Rend., 103. 1125, 1886; 104. 976, 1430, 1887; 107. 442, 1888; Tonometries Paris, 1900; G.
Tammann, Wied. Ann., 24. 523, 1885 ; 36. 692, 1889 ; R. Emden, ib., 31. 145, 1887 ; W. Ostwald,
Lehrbuch der allgemeinen Chemie, Leipzig, 1. 709, 1891 ; L, von Babo, Ueber die Spannkraft des
Wasserdamjifes in Salzlosungen, Freiburg, 1847 ; C. L. Berthollet, Essai de statique chimique,
Paris, 1803.
2 G. Tammann, Mem. Acad. St. Petersburg, (7), 35. 9, 1887.
3 F. M. Raoult, Compt. Bend., 103. 1125, 1886 ; 104. 976, 1430, 1887 ; Zeit. phys. Chem., 2.
353, 1888.
4 G. N. Lewis, Jonrn. Amer. Chem. Soc, 30. 668, 1908 ; E. W. Washburn, ib., 32. 653, 1910 ;
E. W. Washburn and J. W. Read, Proc. Nat. Acad., 1. 191, 1915.
5 S. Arrhenius, Zeit. phys. Chem., 3. 115, 1889; J. H. van't Hoff, ib., 1. 494, 1887; Earl of
Berkeley and E. G. J. Hartley, Proc. Roy. Soc, 77. A, 156, 1906 ; A. W. Porter, ib., 79. A, 519,
1907 ; 80. A, 457, 1908 ; W. Spens, ib., 77. A, 234, 1906 ; H. L. Callendar, ib., 80. A, 466, 1908 ;
Proc. Boy. Inst., 19. 485, 1911 ; A. W. Porter, Journ. Phys. Chem., 12. 404, 1908 ; J. E. Trevor,
ib., 10. 392, 1906 ; 12. 141, 1908 ; J. J. van Laar, Sechs Vortrdge uber das thermodynamische
Potential, Braunschweig, 1906 ; G. N. Lewis, Journ. Amer. Chem. Soc, 30. 668, 1908 ; E. W.
Washburn, ib., 32. 653, 1910; H. Boldingh, De Afwijkingen van de Watten voor verdunds
Oplossingen, Amsterdam, 1893.
^ H. C. Biddle, Amer. Chem. Journ., 29. 340, 1903 ; A. W. Menzies, Journ. Amer. Chem. Soc,
32. 1615, 1624, 1910 ; 36. 798, 1914 ; R. Wright, Proc Chem. Soc, 28. 96, 1912 ; W. Ostwald
and J. Walker, Zeit. phys. Chem., 2. 602, 1888.
7 H. Kronberg, Monatsh., 14. 24, 1893 ; G. Guglielmo, Atti Accad. Lincei, (5), 10. ii, 232,
1901.
* E. W. Washburn, Journ. Amer. Chem. Soc, 132. 653, 1910; J. von Zawidzky, Zeit. phys.
Chem., 35. 77, 1900.
» G. F. Fitzgerald, Journ. Chem. Soc, 69. 885, 1896.
§ 10. Distillation
It is rather curious that the solvent can usually be evaporated from a solution
of a non-volatile solid mthout any of the latter passing away. If the molecules
of the solute are moving freely among the molecules of the solvent, it is difficult
to understand clearly why the former do not escape from the surface of the solution
at a rate which is comparable with the escape of the molecules of the solvent.
According to the hydrate theory, the molecules of the solute are loaded with the
molecules of the solvent, and this prevents the former from moving fast enough to
escape from the attraction of the molecules of the liquid. Of course, most solids
are not volatile at the temperature the solutions are evaporated, and the act
of solution implies that the attraction of the molecules of the solute for one another
is less than the attraction of the molecules of the solvent for those of the solute.
It has been argued that the non-volatility of the original solid implies that its
molecules are retained by the attraction of the molecules of the solvent so as to
prevent the vaporization of the solute along with the free solvent. When a mixture
of two liquids is boiling in a retort, their joint vapour pressure will be equal to the
barometric pressure, and the boiling point of each liquid will be lower than its
boiling point under atmospheric pressure because the partial pressure of each liquid
must be lower than the total pressure. Consequently, each substance will behave
as if it were being distilled under a reduced pressure, and will volatilize at a lower
temperature than its boiling point under atmospheric pressure. In illustration,
ammonium chloride or boric acid volatilizes much more rapidly in a current of steam
than when alone, and hence these substances are carried off the rising vapour when
aqueous solutions are evaporated. J. L. Gay Lussac (1832) ^ showed that if fi and
di respectively denote the partial pressure and vapour density of the liquid, and
^2 and ^2 tbe corresponding constants for the other component of the mixture to be
distilled, then the relative quantities of the two liquids which distil over are related as
Quantity of substance 2 in distillate _p2<^2
Quantity of substance 1 in distillate ""pjCfj
554 INORGANIC AND THEORETICAL CHEMISTRY
If this ratio be large, the distillation of the required substance will be rapid, and
conversely, if this ratio be small.
Example. — A. Naumann (1877-79) applied Gay Lussac's law to mixtures of immiscible
liquids. He showed that a mixture of water ana nitrobenzene boils at 99° under atmo-
spheric pressure. The vapour pressure of water at this temperature is 733 mm. and its
vapour density is 18 ; while the vapour density of nitrobenzene is 123. Hence, the
vapour pressure of nitrobenzene at 99° is 760 — 733 = 27 mm. Here p, =733; rf, = 18;
P8=27 ; and ^2 = 123, and the proportion of water to nitrobenzene in the distillate is as
13194 : 3328, or nearly as 4 : 1.
The composition of the vapour formed by the evaporation of the mixture of
two liquids depends on the proportion in which the two components are contained
in the mixture, and on the vapour pressures of these components at the temperature
at which the evaporation occurs. The total vapour pressure of a mixture of com-
pletely miscible liquids depends upon (1) The relative solubilities of the vapours
in the unlike components of the liquid mixtures ; (2) on the relative attraction
between like and unlike molecules ; and consequently also on (3) the relative pro-
portions of the components of the mixture. If the attraction of the unlike mole-
cules of solvent and solute for one another be negligibly small, the two substances
will appear to be insoluble in one another ; consequently, the vapour of the one
liquid will not dissolve in the other liquid. The greater the attraction of the unUke
molecules for one another, the greater will be the solubility of the one in the other.
In the case of two immiscible liquids, each exerts its own characteristic vapour
pressure independently of the other, although, if one liquid be covered by a deep
layer of the other, the lower liquid may take some time to develop its equilibrium
pressure — unless the mixture is agitated so as to bring the heavy liquid to the
surface and thus facilitate vaporization. In the case of consolute liquids — i.e.
liquids which can be mixed in all proportions — if the mutual attraction of the
unlike molecules be not much in excess of that needed to produce complete
miscibility, the vapour pressure may be less than the sum of that of the com-
ponents, but greater than either one taken singly ; while if this attraction be
relatively large, the vapour pressure may be less than that of either component.
WTien the mutual attractions of the like and unlike molecules of two consolute
liquids are nearly the same the relation between the vapour pressure of the mixture
and its composition is comparatively simple, and can be represented by a straight
line.
C. L. Speyers (1900) 2 supposed that the total pressure P of a mixture of two
liquids, 1 and 2, is related to m, the percentage molecular composition, by lOOP
=mpi-{-{100—m)p2, where ^1 and p2 respectively denote the vapour pressures of the
two liquids. This expression resembles one previously obtained by F. Guthrie
(1884) when the percentage weights were employed in place of molecular propor-
tions. This is probably the case with closely related compounds — e.g., chloro- and
bromo-benzene — where admixture is attended neither by any measurable change
in volume, nor by any evolution or absorption of heat. Let ai2 represent the attrac-
tion of unlike molecules, and a^ and a^ the mutual attractions of like molecules
of the liquid 1 and 2. B. Galitzine (1890) assumed that the relation «i2='\/«i«2
will probably obtain when there is no appreciable alteration of temperature or
volume when the liquids are mixed in equimolecular proportions. C. L. Speyers'
rule does not apply to many mixtures of normal liquids — e.g. carbon tetrachloride
and benzene — but, according to J. D. van der Waals (1900), it should hold good
for liquids whose critical pressures are equal, and whose molecular attractions
agree with Galitzine's relation ai2=V«i«2* ^* Young (1902) tested this hypothesis
for mixtures of chloro- and bromo-benzene, and found that the differences between
the observed and calculated pressures were within the limits of experimental error.
Example." — The molecular proportion of a mixture of bromo-benzene mixed with
chloro-benzene is 50 percent., while the vapour pressure, pi, of the former is 526*25 mm.,
SOLUTIONS
555
and of the latter, p^, 992-30 mm. Hence, P={50 x326-25 + (100-50)992-30}-M00
= 759-4 mm. ; the observed value was 760 mm.
P. Duhem (1887),3 M. Margules (1895), and others have shown that the relation
between the composition and the vapour pressures of the components of a mixture
of two Hquids, 1 and 2, can be represented by an expression equivalent to
\—X X
where x and 1 —x represent the respective molecular proportions of the components
of the mixture, such that if % and n^ respectively denote the number of gram-
molecules of the hquids 1 and 2, in the mixture, a;=Wi divided by ni-\-n^, and
1— iC=W2 divided by ^1+^2- It is assumed that the molecular weights of each
liquid is the same in the gaseous and liquid state of aggregation, and that the
vapours obey the ideal gas laws. The above equation — sometimes called Duhem
and Margules' equation — can be translated into a number of difierent though
equivalent forms, e.g.
dpi/x \ dp2(l—x\
dx^PiJ dx\ P2 ^
When one liquid, say 2, is present in large excess, the partial pressure pi of the other
liquid will follow Henry's law approximately, so that pilx=conat3i,nt—d2)ildx,
or {dpildx){x/pi)=l, and graphically, the variations of x
and pi will be represented by a straight line ; it also
follows that the partial pressure curve of the other
substance will be represented by a curve of the same
type. The value of pjx may be expected to increase or
decrease continuously as x changes in value from zero
to unity when it becomes equal to the vapour pressure
of the pure liquid. Thus, for binary mixtures of con-
solute liquids the vapour pressure curve of the one com-
ponent will be represented by a straight line — Type
II, Fig. 22 — ^or by a curve showing a minimum — Type
I — or a maximum — Type III — value. There are no
sharp lines of demarcation between the three types ;
representatives with every degree of curvature between
the two extreme maximal and minimal curves are ^^^- 22. Diagrammatic
known, with the straight line as a special case. If ^^'^ourPri'sTre" ^ud
the two substances are not miscible in all proportions. Composition of Binary
the partial pressure curve assumes the form Type IV Mixtures,
shown in Fig. 22, and in the region of immiscibility,
when the solution separates into two layers corresponding in composition and
vapour pressure with the points a and h, each of the two layers has the same
partial pressure. More complicated variations of each of these types of curve are
known.
The total vapour pressure P of a binary mixture is, according to Dalton's partial
pressure law, P=Pi+P2, and hence, by difierentiation dpildx=dP/dx—dp2ldx.
Substituting this value of dpi/dx in equation (2), it follows that the variation of
the total pressure with composition will be represented by
dP
dx
dx\ V9 X /
and the vapour pressure P will be a maximum or a minimum when dPldx=0, as
is the case in the region between a and h, Fig. 22, or else when Pilp2=^IO-—^)- The
last condition is satisfied by mixtures with a constant boihng point ; and for
556 INORGANIC AND THEORETICAL CHEMISTRY
consolute liquids it also follows that there can be one and only one maximum
or minimum.
The vapour pressure curves of consolute liquids shown in Fig. 22, are related
with the general phenomena of distillation. The vapour pressure curve III with
a maximum vapour pressure corresponds with a curve with a minimum boiling
point, while the vapour pressure curve I with a minimum corresponds with a mixture
showing a maximum boiling point. In general, the distillation of a binary mixture
furnishes a distillate which is richer in the more volatile constituent, and a residue
in the retort which is richer in the less volatile constituent. There are, however,
three cases :
During distillation, the boiling point of the liquid steadily rises. This corresponds
with the curve II, Fig. 22, and is illustrated by mixtures of methyl alcohol and
water, or of liquid oxygen and nitrogen. The vapour pressure of such mixtures
steadily falls during the progress of the distillation, and the vapour pressures and
boiling points of all possible mixtures lie between those of the two single con-
stituents. The vapour must, therefore, be always richer than the liquid in the
more volatile constituent. By repeated distillation of the distillate, an almost
perfect separation of the two liquids can be effected into distillate (more volatile)
and residuum (less volatile constituent) — ^provided, of course, that their boiling
points be not too close together.
During distillation ^ the boiling point of the liquid rises to a maximum corresponding
with the minimum vapour pressure. This is typified by curve I, Fig. 22. H. E.
Roscoe (1861-2) showed that with aqueous solutions of hydrochloric, hydrobromic,
hydriodic, sulphuric, nitric, or perchloric acid, there is a certain mixture which has
a higher boiling point than any other mixture. The case of hydrochloric acid was
studied by A. Bineau much earlier — ^1838. He found that if hydrochloric acid with less
than about 20*24 per cent, of hydrogen chloride be distilled, water will accumulate
in the distillate until the liquid in the retort has 20*24 per cent, hydrogen chloride.
Such a mixture will distil over unchanged. If the concentration of the liquid being
distilled be greater than about 20*24 per cent, hydrogen chloride, acid will accumu-
late in the distillate until the residue has 20*24 per cent, hydrogen chloride, and
after that an acid of this composition will distil over unchanged. The boiling-
point curves of aqueous solutions of formic acid and of hydrazine also show
maxima corresponding with distillates of constant composition. The residue in
the retort will always have a composition corresponding with a minimum vapour
pressure, that is, with the least volatile mixture.
During distillation the boiling point of the liquid falls to a minimum corresponding
with the maximum vapour pressure. This corresponds with curve III, Fig. 22.
Examples occur with mixtures of propyl alcohol and water. A. A. Noyes and
R. R. Warfel (1901) * also found a mixture of 4*43 per cent, of water with ethyl
alcohol boils at 78*15°, while ethyl alcohol alone boils at 78*3° and water at 100°.
The particular liquid with a maximum vapour pressure has a minimum boiling
point. Whatever the concentration of the original liquid there is a tendency for
the more volatile mixture with the minimum boiling point to distil first.
In neither of these two latter cases can one component be made to accumulate
in distillate or in the residue in the retort so that a separation is as complete as
in the first case — e.g. aqueous solutions of ethyl alcohol can be obtained with no
more than about 96 per cent, of alcohol.
W. Ostwald (1904)* used the term hylotropic mixture {v\v, matter; rpoiros, form)
for a body which retains the same composition and the same properties when it changes
from one phase to another^ — e.g. when a substance changes its state of aggregation. J. Wade
and R. W. Merriman (1911) proposed the term azeotropic mixtures (a, not; C^w, to
boil ; rpoiros, form) instead of the phrase " mixture of maximum or minimum boiling
point." An azeotrope resembles a chemical individual in boiling without undergoing
a change in composition, but differs from it in losing this character when the pressure is
altered. If the composition is independent of the pressure and therefore of the temperature,
the hylotrope is a chemical individual. Hence, a chemical individual is a body which can
SOLUTIONS 557
form hylotropic phases within finite ranges of temperature or pressure. When these limits
are exceeded, and the substance begins to dissociate, it is a chemical compound ; and if
there are no known limits to the range of existence, the body is an element.
Refebences.
^ S. Young, Fractional Distillation, London, 1903 ; J. P. Kuenen, Theorie der Verdampfung und
Verflussigung von Gemischen und der fraktionierten Destination, Leipzig, 1906 ; J. L. Gay Lussac,
Ann. Chim. Phys., (2), 49. 392, 1832 ; A. Naumann, Ber., 10. 1421, 2015, 2099, 1877.
2 C. L. Speyers, Amer. Journ. Science, (4), 9. 341, 1900 ; B. Galitzine, Ueher das Dalton'sche
Oesetz, Strassburg, 1890 ; F. Guthrie, Phil. Mag., (5), 18. 495, 1884 ; S. Young, Journ. Chem.
Soc, 81. 768, 1902 ; 83. 68, 1903 ; S. Young and E. C. Fortey, ib., 83. 45, 1903.
8 P. Duhem, Ann. J^cole Norm. Sup., (3), 4. 9, 1887 ; (3), 6. 163, 1889 ; Trav. Mem. Lille,
3, 1894 ; Traite elementaire de mecanique chimique, Paris, 1899 ; M. Margules, Sitzber. Akad.
Wien, 104. 1243, 1895 ; M. A. Rosanoff, Journ. Franklin Inst., 172. 527, 1911.
* A. A. Noyes and R. R. Warfel, Journ. Amer. Chem. Soc, 23. 463, 1901.
^ W. Ostwald, The Fundamental Principles of Chemistry, London, 1909 ; Journ. Chem. Soc,
86. 506, 1904 ; J. Wade and R. W. Merriman, ib., 99. 984, 1911.
§ 11. Other Hypotheses explaining Osmosis
The substitution of analogy for fact is the bane of chemical philosophy ; the legitimate
use of analogy is to connect facts together and to guide to new experiments.- — ^H. Davy.
Vague similarities on certain properties are never sufficient to determine a person who
earnestly seeks for truth, and is not shackled by hypotheses.— T. Bergmann (1767).
After J. H. van't Hoff i had established the striking analogy between the
osmotic pressure of solutions of cane sugar in water, and the elastic pressure of gases,
he attacked the problem on the theoretical side, and proved that if the solution be
so dilute that the mutual attractions of the molecules of the solute, and the space
they occupy can be neglected, and with a perfectly semipermeable membrane, the
numerical value of the osmotic pressure must be equal to the elastic pressure the
solute would exert if it were in the gaseous conditions at the same temperature and
pressure. The method of proof depends on a reversible cycle of operations and the
second law of thermodynamics. Both J. H. van't Hoff (1890) and Lord Kayleigh
(1897) based their proofs on the applicability of Henry's law, and this was considered
by Lord Kelvin (1897) to be an objection. Accordingly, J. Larmor (1897) devised
a proof in which the laws of the solubihty of gases or possible changes in the state
of molecular aggregation during solution are not involved. The thermodynamic
explanation of osmotic phenomena makes no assumption as to the cause of osmosis,
or how the membrane does its work, and the validity of its deductions is quite
independent of all hypotheses as to the cause of the phenomena it seeks to investigate.
Consequently, it throws little or no light on the mechanism of the process, and the
quest for an explanation of osmotic phenomena proceeds independently of thermo-
dynamics. There are a number of explanatory hypotheses — both physical and
chemical — sur le tapu'. The chemical hypotheses will be considered later.
The gas-analogy hypothesis of osmotic pressure. — The laws associated with
the names of Boyle, Charles, Dalton, and Graham, and the hypothesis of
Avogadro, are but a few of the many striking analogies subsisting between the
behaviour of gases confined in a given space and substances in dilute solution.
We know enough about nature to believe that if two things are exactly alike, they
will behave alike under the same circumstances ; but when the things compared
are not quite similar, we must be prepared for discrepancies. Analogy is not proof.
Had Isaac Newton measured the refractory power of native cadmium sulphide —
greenockite — he would no doubt have said : " greenockite is probably an unctuous
substance coagulated," and he would have been wrong. As it happened, this
prognostication turned out all right with the diamond. The hypothesis that the
osmotic pressure of a dilute solution is produced by the bombardment of the
semi-permeable membrane by the dissolved molecules gives a very plausible
558 INORGANIC AND THEORETICAL CHEMISTRY
interpretation of the analogy between the behaviour of dissolved molecules, and the
molecules of a gas brought out by J. H. van't HofE in 1887, but the analogy appears
to break down so completely with more concentrated solutions that a number of
rival hypotheses have been advanced to explain the phenomena. The principle
of exhaustion compels us to investigate other hypotheses.
Solvent pressure hypothesis. — According to F. Tinker,^ osmotic pressure is
primarily a solution pressure and is not a pressure produced by the solute or
dissolved substance. He says : The tendency of a liquid to diffuse is measured
by its diffusion pressure, which may be defined as the bombardment pressure exerted
by the liquid molecules on either side of a plane of unit area placed anywhere within
the liquid. Since the absolute concentration of a solvent is reduced by the intro-
duction of a solute, it follows, therefore, that the diffusion pressure of the solvent in
a solution is always less than that in the pure solvent itself, and the osmotic pressure
of the solution will be proportional to the difference between the absolute concen-
tration of the solvent on the two sides of the membrane. It is also approximately
proportional to the concentration of the solute because the latter is itself approxi-
mately equal to the difference in solvent concentration on the two sides. If now
there be removed from the solution side all the solvent molecules and an equal
number from the pure solvent side, the residue of solvent molecules will diffuse
across the membrane as before while the solute molecules wiU bombard the mem-
brane. Moreover, the pressure of the solvent residue on the one side will be equal
to the pressure of the solute on the other, and both will be equal to the corresponding
gas pressure since the molecules are now at distances from one another comparable
to gas distances. Hence to prevent the residue of solvent from flowing across
the membrane, a hydrostatic pressure equal to the gas pressure will have to be
applied. The phenomenon of osmotic flow is therefore due to the residue or excess
of solvent molecules on the pure solvent side. The solute molecules play an indirect
part only, but they do cause a strain to be placed on the membrane which tends to
rupture it. The fundamental difference between osmotic phenomena in the gaseous
and solution states is that whereas the active molecules have a vacuum for a
medium in the case of gases they have a liquid for a medium in the case of solutions.
Vapour pressure h3rpothesis. — H. L. Calendar's hypothesis 3 (1908) is one
of the most satisfactory of the purely physical explanations of osmotic pressure,
and it is superior, in many respects, to the gas-analogy hypothesis. H. L.
Callendar's hypothesis has been tested with somewhat concentrated solutions,
and wherever data are available it has been eminently successful. Experiment
shows that the maximum vapour pressure of a solution can be altered in three
ways : (1) by altering the temperature ; (2) by varying the concentration of the
solution ; and (3) by altering the pressure under which the liquid itself is confined.
The effect of pressure on the freezing point of water (OiV, Fig. 9, Chapter IX)
is an application of the third principle. An objection might very properly be
raised to the third method of altering the vapour pressure of a liquid ; it has been
shown to be impossible to raise the pressure on a saturated vapour, without
causing some of it to liquefy. If a vertical cylinder, provided with a piston, con-
tains nothing but water — liquid and vapour, it is quite true that the descent of the
piston will result in the condensation of water vapour until all the vapour is liquefied,
and as long as water vapour is present the vapour pressure remains constant. On
the contrary, if air as well as water vapour be present, it is easy to see that the
volume of the air decreases, or the pressure of the air on the surface of the liquid
increases during the descent of the piston. The water vapour still supports its
own share of the total pressure up to its maximum vapour pressure, and not
quite so much water vapour as before will condense, consequently the liquid under
a considerable external pressure can exert a greater vapour pressure than the
maximum vapour pressure under atmospheric pressure.
It has been proved experimentally that the maximum vapour pressure Of
a solution under very great pressures is rather greater than the maximum
180
300
420
540
14-6
26-8
44 0
67-5
14-1
26-8
43-7
67-6
SOLUTIONS 559.
vapour pressure of the same solution under atmospheric pressures (see the
curve ON, Fig. 9, Cap. IX). Again, the vapour pressure of a solution is less than
the vapour pressure of the pure solvent, Fig. 21. Consequently, if the pressure
on a solution be sufficiently augmented, the pressure of its vapour can be made
equal to the vapour pressure of the pure solvent under atmospheric pressure.
This is the condition necessary in order that solution and solvent can exist side
by side in equilibrium. If the vapour pressure of the solution were less than that
of the pure solvent, the system would not be in equiUbrium, because vapour
would distil from the solvent into the solution until the vapour pressure of both
were the same. Conversely, when a solution under its own osmotic pressure and
the pure solvent are in equilibrium, it follows that their* vapour pressures must be
equal. Hence, according to H. L. Callendar : The osmotic pressure of a solu-
tion represents the external pressure which must be applied in order to make
its vapour pressure equal to that of the pure solvent. With this hypothesis,
H. L. Callendar has calculated the osmotic pressures of sugar solutions of different
concentration from published vapour pressure data, and the results are in close
agreement with observation :
Concentration (grams per litre)
Observed osmotic pressure (atmospheres)
Calculated osmotic pressure (atmospheres)
Hence it is inferred that osmotic equilibrium depends upon the equality of the
vapour pressure of the solution and of the pure solvent.
The semipermeable membrane has been styled a vapour sieve, and hkened to a
partition pierced by a large number of minute capillary tubes. Suppose that the
capillary tubes are not wetted by either the solvent or solution, then neither the
liquid solvent nor the solution can enter the capillaries — unless the pressure on
one of the Uquids exceeds 100 atmospheres — although vapour can diffuse through
the capillary tubes. But the vapour pressure of the solution on one side of one
of the capillary tubes is less than the vapour pressure of the solvent on the other
side ; consequently, vapour will pass through the capillary and distil from the
solvent to the solution. Hence the volume of the solution will increase, and if
the solution be confined in a closed vessel, the pressure must rise and continue rising
until the vapour pressure of the solvent and solute are the same. This increase
is the so-called osmotic pressure of the solution.
The agreement of this expression for dilute solutions with observation does not
necessarily mean that the molecules of the solute can move independently of the
solvent. It is difficult to beheve that the molecules of solute and solvent are in-
dependent of one another, and J. Larmor (1897) assumed that each molecule of the
solute forms for itself a nidus in the solvent ; that is, it sensibly influences the mole-
cules around it up to a certain minute distance so as to form a loosely connected
complex in the sense, not of chemical union, but of physical influence ; and, if the
solution be dilute, each such complex is very much the greater part of its time out
of the range of influence of other complexes, an application of the principles of
thermodynamics then necessitating the osmotic laws. J. H, Poynting (1896) ^ has
shown that the very same expression can actually be obtained by the assumption
that the molecules of the solute enter into some sort of chemical combination with
the solvent. Evaporation occurs when the molecules pass through the surface of
the solution with a sufficient velocity to overcome the attractions of the neighbour-
ing molecules, and if the molecules of the non-volatile solute are each loaded with
molecules of the solvent, the complexes will not pass out of the solution, and the
surface of the solution may be likened to a semipermeable membrane as regards
solute and solvent ; accordingly, the evaporation of a solution must proceed more
slowly than with the pure solvent. The complex hydrates, however, are supposed
to be always as effective as the solvent molecules in entangling the particles which
impinge on the surface, and accordingly condensation proceeds more rapidly than
560 INORGANIC AND THEORETICAL CHEMISTRY
evaporation until equilibrium is established. Hence, both evaporation and con-
densation proceed more slowly with the solution than with the solvent. A similar
state of things is supposed to prevail at the surface of a semipermeable membrane,
but owing to the rigidity of the latter, the excess of " condensation " over " evapora-
tion " gives rise to a hydrostatic pressure on the solution side which is a measure of
the osmotic pressure. T. M. Lowry (1897) has shown that it is not necessary to
assume, with J. H. Poynting, that the molecules of the solute are hydrated, and he
shows that in virtue of the mere presence of the molecules of the solute in the surface
layer of the solution, evaporation must be retarded ; a molecule of the solvent
rising from the interior of the solution may strike a molecule of the solute in the sur-
face layer and rebound back into the liquid without passing through the surface
into the region of the vapour ; condensation is more rapid because it will proceed
as if no solute were present. While T. M. Lowry's hypothesis gives a similar relation
between vapour pressure as that previously obtained, J. H. Poynting's hypothesis
requires that if the molecules of the solute are mono-hydrated the osmotic pressure
will be doubled ; etc.
Surface tension hypotheses. — M. THermite (1855), S. L. Bigelow (1907), I. Traube
(1904), G. Jager (1891), and B. Moore (1894) ^ tried to explain osmotic phenomena
as a result of the different surface tensions of the two liquids — solution and solvent.
Let T and Tg denote the respective surface tensions of the two liquids A and B
contained in vessels connected by a capillary tube of radius r so that the liquids
meet in the capillary. If T be greater than Tg the skin of liquid at the surface of
contact with the wall of the capillary tube will move in the direction of A with a
force equivalent to {T—Ts)27rr. If the radius of the capillary be small enough
to bring all the contained liquid within the range of the capillary force — as is probably
the case with ordinary osmotic membranes — the whole body of liquid in the tube
will be driven in the direction of A. Hence, osmosis should proceed from the region
of the less to that of the greater surface tension. Hence, (1) the surface tensions of
solutions obeying the other solution laws should be greater than those of the pure
solvents ; and (2) the surface tensions of solutions of a given substance in a given
solvent should be proportional to the concentrations. I. Traube (1904) compared
hundreds of measurements of surface and osmotic effects, and found one and all
in agreement with the hypothesis that the motive force of osmotic phenomena
is determined by a difference in the surface tensions of solution and solvent.
A. Battelli and A. Stefanini (1906) showed that aqueous solutions of salicine (or
of ethyl alcohol) have a less surface tension than water, and yet osmosis
takes place in the direction of the solution. Hence, they, like I. Traube, ascribe
osmotic phenomena to differences in the surface tensions of the liquids on the two
sides of the membrane ; but, unlike B. Moore and I. Traube, they consider that
osmosis proceeds in the direction which lends itself best to an equalization of surface
tensions. If the membrane be permeable to but one liquid, it alone will pass
through ; but if permeable to both, osmosis will proceed in both directions until
the surface tensions are equalized. If the passage of n molecules of water will
increase the surface tension of a given mass of alcohol less than the passage of n
molecules of alcohol would decrease the surface tension of water, water will flow
to the alcohol more rapidly than alcohol will flow to the water. Solutions with
the same surface tensions have the same osmotic pressure independent of their
concentration. For instance, a 1:78 per cent, solution of magnesium sulphate,
and a 1*11 per cent, solution of sodium sulphate have the same surface tension, and
they produce no difference in osmotic pressure when placed on each side of an osmotic
cell. In conclusion, no one has yet succeeded in giving an adequate account of
osmotic pressure, but that facts seem to indicate that osmotic pressure and surface
tension are related in some way so that the two phenomena exhibit in many cases an
interesting parallelism.
SOLUTIONS 561
References.
1 J. H. van't Hoff, Zeit. phys. Chem., 1. 481, 1887 ; Phil Mag., (5), 26. 81, 1888 ; Eep. Brit.
Assoc, 335, 1890; Lord Rayleigh, Nature, 55. 253, 1897; Lord Kelvin, ib., 55. 272, 1897;
J. Larmor, ib., 55. 545, 1897 ; Phil. Trans., 190. 266, 1897 ; A. Findlay, Trans. Faraday Sgc,
3. 30, 1907 ; W. C. D. Whetham, ib., 3. 34, 1907 ; Nature, 74. 54, 1906 ; P. Duhem, Journ.
Phys., (2), 6. 134, 397, 1887 ; (2), 7. 391, 1888 ; E. Bouty, ib., (3), 4. 154, 1895 ; A. Gouy and
G. Chaperon, Compt. Bend., 105. 117, 1887; Ann. Chim. Phys., (6), 13. 120, 1888; M. Dupin,
Der osmotische Druck und seine Berechungen zur freien Energie, Berlin, 1889 ; G. Bredig, Zeit.
phys. Chem., 4. 444, 1889 ; L. Meyer, ih., 5. 23, 1890 ; J. H. van't Hoff, ib., 5. 174, 1890 ;
9. 477, 1892 ; L. Boltzmann, ib., 6. 474, 1890 ; 7. 88, 1891 ; H. A. Lorentz, ib., 7. 36, 1891 ;
T. Ewan, ib., 14. 409, 1894 ; J. J. van Laar, ib., 15. 457, 1894 ; 31. 22, 1899 ; K. Schreber, ib., 28.
79, 1899 ; A. Noyes, ib., 28. 220, 1899 ; 35. 707, 1900 ; K. Ikeda, ib., 33. 280, 1900 ; M. Planck,
ib., 41. 212, 1902 ; 42. 584, 1903 ; C. Dieterici, ib., 29. 139, 1899 ; 37. 220, 1901 ; F. Barmwater,
ib., 28. 115, 1899 ; Uber die Natur des osmotischen Druckes, Kopenhagen, 1898 ; L. Meyer, Sitzber.
Akad. Berlin. 48, 993, 1892 ; 1. Traube, Zeit. anorg. Chem., 8. 232, 1895 ; J. H. Poynting, Nature,
55. 253, 272, 461, 545, 606, 1896 ; Phil. Mag., (5), 42. 289, 1896 ; S. R. Milner, ib., (5), 49. 417,
1900 ; A. H. Bucherer, Wied. Ann., 64. 549, 1898 ; H. Jahn, Ber., 30. 2982, 1898 ; J. Traube,
ib., 31. 154, 1898 ; Zeit. anorg. Chem., 8. 232, 1895 ; A. Smits, Rec. Trav. Pays-Bas., 22. 153,
1903 ; P. Fireman, Joiirn. Phys. Chem., 6. 636, 1902 ; G. Jager, Zeit. phys. Chem.y 93. 275, 1919.
2 F. Tinker, Nature, 97. 122, 1916 ; Phil. Mag., (6), 33. 428, 1917,
3 H. L. Callendar, Proc. Boy. Soc, 80. 466, 1908 ; Proc. Boy. Inst., 19. 485, 1911.
* J. H. Poynting, Phil. Mag., (5), 42. 289, 1896 ; T. M. Lowry, ib., (6), 13. 552, 1907 ; Trans.
Faraday Soc, 3. 12, 1907.
6 B. Moore, Phil. Mag., (5), 38. 299, 1894 ; I. Traube, ib., (6), 8. 704, 1904 ; Wied. Ann.,
62. 490, 1897 ; A. Battelli and A. Stefanini, Phys. Zeit., 7. 190, 1906 ; Journ. Phys., 6. 402, 1907 ;
M. I'Hermite, Ann. Chim. Phys., (3), 43. 420, 1855 ; S. L. Bigelow, Journ. Amer. Chem. Soc,
29. 1675, 1907 ; G. Jager, Sitzber. Akad. Wien, 100. 245, 493, 1891 ; A. SeUa, Atti Accad. Lincei,
(5), 16. ii, 384, 1907.
§ 12. The Relation between the Boihng Point of a Solution and the
Molecular Weight o£ the Solute
In Fig. 23 the curve PO represents the vapour pressure of the solid, and OQ
the vapour pressure of the pure liquid. The two curves intersect at the freezing
point 0. Let Q, Fig. 23, represent the boiling point
of the solvent at 760 mm. pressure, then since the
vapour pressure of a solution is less than the vapour
pressure of the pure solvent, let O^Q^ represent the
vapour pressure curve of a given solution. Then PM
will represent the freezing point of the solvent, and
PM' the freezing point of the solution. Since PM' is
less than PM, the freezing point of the solution will be
less than the freezing point of the solvent ; and since
PN' represents the boiling point of the solution and
PN the boihng point of the solvent, the boihng point Temperatures
of the solution must be greater than the boiling point Fig. 23. — Diagrammatic,
of the pure solvent. This agrees with experiment.
For instance, with solutions of potassium iodide in 100 grams of water, G. T.
Gerlach (1887) i found :
p
Solii'tioH Boils q'//
Solvent Boiis // Q'
V
//!
p
^
yy \
J^
V^"^ ' :
1
F
» M'M NN'
Boiling point
100°
101°
102°
103°
104°
105°
Potassium iodide
0
15
30
45
60
74 grams
If the solutions are very strong the relation is not quite the same, but the
raising of the boiling point o! a dilute solution is directly proportional to the
weight of the dissolved substance in a given weight of solvent. Double the con-
centration of the solution, and the elevation of the boiling point will be doubled.
An equal number of molecules of the dissolved substance in the same quantity
of a solvent give the same elevation of the boiling point — F. M. Eaoult (1886). ^
Hence the rise in the boiling point of a solvent is proportional to the number of
VOL. I. 2 o
562 INORGANIC AND THEORETICAL CHEMISTRY
molecules of the dissolved substance in solution, and inversely proportional to the
molecular weight of the solute.
In Fig. 23, QN and RN represent the respective vapour pressures of solvent
and solution at the boiling point PN of the pure solvent ; QS and RQ' represent
the vapour pressure curves as the temperature rises ; QQ' is parallel to the tempera-
ture axis, and represents the rise in the boihng point NN'dT. The solution is
supposed to be dilute, and therefore there will be no sensible error if the curves be
taken close together and parallel. "HhenQQ'^^QQ'.QS/Q'S ; 0TdT=QR{Q'SIQQ');
but Q'R=p—ps; and Q/SIQQ'=dp/dT. Consequently, dT={p—ps)dp I dT. From
Clapeyron's equation, neglecting V2 for small differences of temperature, and
putting the vapour density Z)=l/vj per unit mass of substance,, it follows
that dp/dT=XD/T ; and from p — ps=PD/s, where s denotes the specific gravity
of dilute solution or solvent, and P denotes the osmotic pressure of the solution.
Hence, dT=PTIXs, where A represents the latent heat of vaporization of the
solvent. This expression shows that the osmotic pressure is directly proportional
to the rise of the boiling point. Since P denotes the osmotic pressure of the solution
containing one gram-molecule of solute in a volume F, it follows that PV=^RT,
and if the same volume of solution has n gram-molecules of the solute PV=nRT,
or P=nRTIV ; if the solution, of specific gravity s, contains N gram-molecules
each of molecular weight M, it follows that V=NMIs. Remembering that R is
nearly equivalent to 2 cals., substituting in dT=PTIXs for the values of P and V
just determined ; and reducing to its lowest terms, dT=2nT^INMX ; or,
dT={n/N){2T^IMX). As previously indicated, if n=ivlm, and iV=100/M, it
follows that dT={wlm){0'02T^IX).
Let the symbol k stand for the so-called boiling constant ; then, since, for any
given solvent, T and A are physical constants, the
^ ... , ^ , 0-02^2
Boiling constant, Jq=
A
and m, the molecular weight of the vapour of the solute, becomes m=kiv/b, where b,
the raising of the boiling point, has been put in place of dT. The meaning of k is
obtained by assuming that one gram-molecule {w/7n) of the solute — say 342 grams
of cane sugar, C12H22O11 — is dissolved in 100 grams of the solvent, then ^=6,
and therefore k represents the elevation of the boiling point produced by
dissolving one gram-molecule of a substance in 100 grams of solvent ; k is called
the boiling constant, or the molecular elevation of the boiling point. E. Beck-
mann and 0. Liesche ^ have shown a number of other ways of computing the
boiling constant ; thus \mk=mp{dT jdp) ; k=0-0O^%MT ; k=0'00n0S9T^la^ ;
k=0'OOi34:3MT{l-TITc)^log {Pc/p), when dT/dj) represent the rise dT in the
boiling point for a change dp in the vapour pressure ; a^ denotes P. Walden's
specific cohesion of the solvent at the boiling point ; and Tc and pc respectively
denote the critical temperature and pressure of the solvent.
Example.- — What is the molecular elevation of the boiling point of water boiling
at 100°, when its latent heat of vaporization is 537 cals? Here T = 100 + 273; and
A;=0-02x{373)2-^-537 = 5-18, which is very near the observed value 5'2.
Each solvent has its own specific boihng constant :• e.g. acetone, 16*7 ; benzene,
26'7 ; ether, 21*6 ; carbon disulphide, 23*5 ; ethyl alcohol, 11*5 ; chloroform.
35'6 ; pyridine, 30' 1 ; etc. The boihng constant is determined by finding the
boihng point of, say, water and of aqueous solutions containing 0"02, 0*06, O'lO
gram-molecules of cane sugar, and calculating the results per 342 grams of cane
sugar. Suppose that w grams of a substance dissolved in 100 grams of water
raised the boihng point of the water b°. Then, if M be used to denote the mole-
cular weight of the substance, we have the proportion w : M=b : 5, or, for substance
dissolved in water :
w
Molecular weight==5"2 . . • (3)
SOLUTIONS
563
This enables the molecular weight of many substances to be determined from their
effect on the boihng point of water. The particular solvent to be used depends
on the solubility of the substance under investigation. If ether is used in place of
water, 5"2 must be altered to 21*6, etc. In no case must a solvent be employed
which reacts chemically with the substance under examination.
Examples.' — (1) E. Beckmann (1890) found that 2*0579 grams of iodine dissolved in
30*14 grams of ether raised the boiling point of the ether 0*566°. What is the molecular
weight of iodine ? Here, 2*0579 grams of iodine in 30*14 grams of ether correspond with
100 X 2*0579-^30*14 =t^; = 6*8278 grams of iodine in 100 grams of the solvent. Hence,
JIf = 21*6x6-8278-f-0*566 = 254*6. This corresponds with the formula lo when iodine has
a molecular weight of 253*84. The numbers seldom, if ever, coincide, but there can be no
mistake in the significance of the figures.
(2) E. Beckmann (1890) foxind that a solution of 1*4475 gram of phosphorus in 54*65
grams of carbon disulphide raised the boiling point 0*486°. What is the molecular weight
of the phosphorus ? Answer : Molecular weight, 129*16.
The atomic weight of phosphorus is 31, hence the mole-
cule of phosphorus is represented : P4.
(3) A. Helff (1893) found that 0*2096 gram of sulphur
in 17*79 grams of carbon disulphide raised the boiling
point 0*107°. Hence show that the molecular weight
of sulphur is probably Sg. Here w = l'll ; and the
molecular weight is 259. This is close to the theoretical
value 256 for Sg.
(4) A solution of 3*164 grams of cupric chloride in
100 grams of alcohol raised the boiling point 0*308°.
The boiling constant of alcohol is 11*5, what is the
molecular weight of the solute ? Answer : 134*5.
(5) L. Marchlewsky and J. Sachs (1892) analyzed
Roussln's salt and found it to contain 38*29 per cent,
of iron ; 16*54 of sulphur ; 16*70 of nitrogen ; and 6*54
per cent, of water. When 0*1826 gram of the salt was
dissolved in ether, the boiling point was raised 0*012°.
Show that this agrees with the recognized formula of
the salt: Fe4(NO):S3K.H20.
E. Beckmaim's process for the determination
of boiling points (1888-96) .--This method of
determining molecular weights has been much
employed by F. M. Eaoult, E. Beckmann,
H. C. Jones, and many others.'* The process
is applicable only to solutes which do not
give off an appreciable amount of vapour at
the boiUng point of the solution. The so-called
Beckmann's thermometer has contributed largely
to the successful application of the method.
The modern forms of this instrument are sensi- -p^^ 24.^Beckmann's Apparatus
tive to 0'001°. Great precautions must be taken for Boiling-point Determinations,
to measure the temperatures accurately since
a small error in the temperature readings has a large influence on the computed
result.
The apparatus consists of a glass boiling tube A, Fig. 24, with a piece of platinum wire
sealed in the bottom, and packed with beads to prevent irregular boiling. A side tube with
a condenser, C, liquefies the vapour given off during the boiling ; and the exposed end of
the condenser is closed with a calcium chloride tube, D. The boiling tube is surrounded
by a jacket of some non-conducting material, E, to prevent the radiation of heat. The
boiling tube is fitted with Beckmann's thermometer, T, which can be read to ^^^^^ of a degree,
and set so that the mercury is about halfway up the stem when the solvent is boiling.
Beckmann's thermometer has a reservoir of mercury at the top so that it can be set for use
at any desired temperature. As indicated in text-books of laboratory processes, this
avoids an inconveniently long, or an inconveniently large number of thermometers.
The thermometer is always tapped before a reading to make sure the mercurj' is not
lagging behind. The lens L facilitates the reading of the thermometer. The barometer
should be read to make sure no appreciable change occurs during a determination.
564
INORGANIC AND THEORETICAL CHEMISTRY
The boiling tube has a stoppered side tube, J, for introducing the solutions under investiga-
tion. The whole is clamped to a stand and rests on an asbestos tray, F. The boiling point
of the solvent is first determined. The boiling tube is weighed. The solvent is introduced
and its boiling point determined when the boiling is brisk and vigorous. A known weight
of the substance is then introduced, and the boiling point of the solution determined. A
correction is made by subtracting 0*2 to 0*4 gram from the weight of the solvent in order
to allow for the solvent condensed on to the walls of the apparatus and the condenser. The
actual correction depends upon the nature of the solvent and the particular form of the
apparatus used. The difficulty with this apparatus is to avoid fluctuations of temperature
in the boiling tube due to the radiation of heat ; dripping of the cold liquid from the con-
denser into the boUing solution, etc. Many other forms of apparatus for this determination
have been devised.
W. Landsberger's method for the determination of boiling points (1898).—
When the solution of a non-volatile solute boils, the vapour of the solvent
and solution are in equilibrium, and this condition can be established by leading
the vapour of the boiling solvent into the solution. When the solution is boiling,
the vapour will pass through the system without condensation ; if the solution
is below this temperature some vapour will condense, and the latent heat of con-
densation will continue heating the solution until the boiling point is reached. There
is virtually no danger of superheat-
^ /vj ing the solution. This method of
determining the boiling point of a
solution has been employed with
some success in molecular weight
determinations . 5
In W. Landsberger's apparatus, a
modification of which is shown in Fig.
25, the solvent is boiled in the flask A,
and the vapour passed into the solution
via the tube F. The temperature of
the solution is raised to its boiling
point by the latent heat of condensa-
tion of the vapour of the solvent. The
vapour of the solvent passes to the con-
-Landsberger's Apparatus for Boiling denser G through E around the boiling
Determinations. tube, and thus the inner tube is
jacketed with the vapour of the boil-
ing solvent. This reduces radiation losses. The boiling point of the solvent is first
determined, and a weighed amount of the solute is introduced into the inner tube B, which
is graduated so that the boiling can be interrupted for a moment before more solute
is added, and the volume of the solution read at a glance.
With this apparatus, the boiUng constant for water is nearly 5i; allowing for
this, and using the preceding notation, with water as a solvent,
Molecular weight=5*4:
w
{*)
where w denotes the weight of the substance per 100 c.c. of the solvent, and b repre-
sents the elevation of the boiling point. If other solvents be used 5*4 is altered thus :
for acetone (sp. gr. 1'827), 22'2 ; benzene (sp. gr. 0*879), 32-8 ; ether (sp. gr. 0-736),
30-3 ; carbon disulphide (sp. gr. 2-63), 26 ; ethyl alcohol (sp. gr. 0*80), 15*6 ;
chloroform (sp. gr. 1*526), 26*0 * aniline (sp. gr. 1'022), 38*2. If the boiUng tube
be weighed so that the amount of solvent is determined by weight, and not by
volume, the original formula is used.
Examples. — (1) 0-8686 grm. of boric acid with 7'73 c.c. of water raised the boiling
point 0-917°. What is the molecular weight of boric acid ? Ansr. : 63-7. Theory for
B(0H)3 is 62. Show that this result is concordant with an elevation of the boiling point
of 0-262° when 33-4 c.c. of water contains 1-015 grms. of the acid in question.
(2) W. Landsberger (1898) foimd that 0-4929 grm. of cadmium iodide with 7*30 grms.
of ethyl alcohol gave a rise of 0-218° in the boiling point of the solvent. Show that this is
in hai-mony with the theoretical molecular weight of 366 for Cdlj. The value computed
from the observed data is 356.
SOLUTIONS 565
Correction for the volatility of the solute.— The formula (3), for
calculating the molecular weight of a solute from its efiect on the boiling point of
the solvent, is valid only when the solute does not volatilize. If the solute
volatilizes along with the solvent, this formula must be replaced by
Molecular weight=5*4— 7 — ; or, Molecular weight =A;(1 — a)j- . (5)
where w represents the number of grams of the solute in 100 grms. of the solvent,
and Wi the number of grams of the solute per 100 grms. of the vapour, and its value
is obtained from the mean value of the concentration of the solute in 100 grms.
of the distillate when the solution is distilled ; and a represents the ratio of the
concentrations of the volatile solute in the vapour and in the solution ; k is the
boiling constant — for water, jfc=5"4.
Example. — Compare the molecular weights of iodine in carbon tetrachloride and in
benzene when it is found that 8'498 grms. of the distillate from iodine in carbon tetra-
chloride contained 0*168 grm. of iodine per 8'330 grms. of the solvent ; the initial and end
concentrations of the boiling solutions were 4-507 and 6'592 respectively ; 0*722 grm. of
iodine in 30*2 grms. of the same solvent raised the boiling point 0*315" ; and the boiling
constant for this solvent is 48*8. Similarly, for the solvent benzene, o =0*144 ; boiling
constant k=25'7 ; and 1*195 grms. of iodine in 22*3 grms. of the solvent raised the boiling
point 0*467°. For carbon tetrachloride, w is the mean value of 4*507, 6*592, and 5*550 ;
Wi is 100x0*168-^8-330=2*107 ; and a is 2*107^5*550=0-36. Again, the solution used
for measuring the effect of iodine on the rise of the boiling point of carbon tetrachloride,
w=0•722xl00-^30*2=2*391, and W& ><48-8=2*391 x49*8-^0•315 = 370. The imcorrected
molecular weight of the iodine in boiling carbon tetrachloride is 370, and when corrected
for the volatility of the iodine it is 370(1 —a) =370 X 0*66 = 244. Similarly, the uncorrected
molecular weight of iodine in boiling benzene is 294, and the corrected value 252.
In the laboratory, advantage is taken of the fact that the boiling point of a
solution is higher than the boiling point of the solvent to get liquids for baths, etc.,
boiling a few degrees higher than water, by using salt solutions in place of
water. For example, a saturated solution of sodium nitrate boils- at 120°, and
a saturated solution of sodium chloride at about 108°.
References.
1 G. T. Gerlach, Ueber Siedetemperahiren der Salzlosungen, Wiesbaden, 1887.
2 F. M. Raoult, Tonometries Paris, 1900; Compt. Rend., 103. 1125, 1886; 104. 976, 1430, 1887 ;
105. 857, 1887 ; 107. 442, 1888 ; Ann. Chim, Phys., (6), 15. 375, 1888 ; Zeit. phys. Chem., 2.
353, 1888.
3 E. Beckmann and 0. Liesche, Zeit. phys. Chem., 86. 337, 1914 ; P. Walden, ib., 65. 257,
1909.
* E. Beckmann, Zeit. phys. Chem., 6. 437, 1889 ; 7. 539, 1889 ; 8. 226, 1891 ; 15. 664, 1894 ;
21. 245, 1896 ; 40. 129, 1902 ; 58. 543, 1907 ; E. Beckmann and A. Stock, ib., 17. 107, 1895 ;
P. Fuchs, ib., 22. 72, 1897 ; H. C. Jones, The Freezing-point, Boiling-point, and Conductivity
Methods, Easton, Pa., 1897.
s W. Landsberger, Zeit. anorg. Chem., 17. 422, 1898 ; Ber., 31. 458, 1898 ; C. N. Riiber, ib.,
34. 1060, 1901 ; J. Walker and J. S. Lumsden, Journ. Chem. Soc, 73. 502, 1898 ; M. Reinganum,
Wied. Ann., 59. 764, 1897.
§ 13. The Relation between the Freezing Point of a Solution and the
Molecular Weight o£ the Solute
F. M. Raoult's method for determining the molecular weight of a substance is the most
significant contribution to the list of physical processes applicable to chemical investigations
since the discovery of the law of Dulong and Petit.' — V. Meyer (1888).
Similar remarks apply mutatis mutandis to the freezing point of solutions as were
made with reference to the boiling point. A study of Fig. 23 will show that if the
vapour pressure of a solution is less than that of the pure solvent, the vapour pressure
566 INORGANIC AND THEORETICAL CHEMISTRY
curve will cut the ice curve at a temperature below the freezing point of the pure
solvent. This means that the freezing point of a given solution will be lower than
the freezing point of the pure solvent, and experiment shows that the lowering
of the freezing point will be proportional to the weight of the substance dissolved
in a given weight of the solvent. This reminds us of C. Blagden's discovery (1788)
that the depression in the freezing point of a solvent produced by the addition of a
salt is directly proportional to the concentration of the solution ; in 1861 F. Riidorff
rediscovered the law and explained certain deviations he observed by assuming
that hydrates were formed ; and in 1871, L. C. de Coppet found that if the quantity
of salts dissolved is proportional to the molecular weight, salts of analogous con-
stitution produced the same depression. Experiments show that an equal number
of gram-molecules of difierent substances in the same solvent depress the freezing
point to the same extent — F. M. Raoult's law ( 1883-84). i The depression in the
freezing point is proportional to the weight of the dissolved substance in a given
weight of the solvent ; and inversely proportional to the molecular weight of
the dissolved substance. The same law was deduced from thermodynamics by
J. H. van't Hoff in 1886.2 Solutions of sugar (342 grms.), methyl alcohol
(32"03 grms.), etc., in 100 grms. of water depress the freezing point 18'5°. This
is the freezing constant for water ; Raoult called it the molecular depression
of the freezing point per 100 grms. of solvent, and its numerical value for each
solvent can be derived precisely like the boiling constant by substituting the
latent heat of fusion for A in place of latent heat of vaporization.
^ . /7x 0-02T2
Freezing constant (/J)= c —
Examples.' — (1) The latent heat of fusion of water is 79 cals; hence 2^ = 273, and
A = 79, and k is 18*9 ; the observed value is 18-5 (F. M. Raoult).
(2) The observed value for the freezing constant of acetic acid is 39; show that this
deviates from the theoretical value by about 2 per cent, when the latent heat of fusion
is 43*66 cals. The computed value is 3 8 "2.
(3) The freezing constant or molecular lowering of the freezing point of sulphur which
freezes at 119-25°, and has a latent heat of fusion 9'368, is 0-02 x(392-25)2-^9-368 = 328.
Each solvent has its own specific freezing constant ; e.g. acetic acid, 38'88° ; benzene,
49° ; mercury, 425° ; naphthalene, 69° ; phenol, 74° ; etc. If iv grms. of a substance,
molecular weight M, dissolved in 100 grms. of solvent, lowers the freezing point
/°, we have the proportion w : M=^f: 18'5 for water ; or, for substances dissolved
in water,
w
Molecular weight=18'5-^ . . . • (6)
This enables the molecular weight of a substance to be computed from its effect
on the freezing point of water. The particular solvent to be selected is of course
determined by the solubility of the substance under investigation, and the number
18*5 must be replaced by another if a different solvent be used. In no case must
a solvent be chosen which can enter into chemical union with the substance to be
examined.
Examples. — (1) W. Tammann (1889) found that a solution of 0-022 grm. of sodium
in 100 grms. of mercury lowered the freezing point of mercury 0-39°. What is the mole-
cular weight of sodium? Here, M = 425 x0-022-i-0-39 = 23-8. Hence the atomic and
molecular weights are the same.
(2) W. R. Omdorff and J. White (1893) found that a solution of 0-2735 grm. of hydrogen
peroxide in 19-86 grms. of water lowered the freezing point of water 0'746°. What is the
molecular weight of hydrogen peroxide ? Here w; = 100 x0-2735-M9-86 = l-3773 ;/ = 0-746 ;
hence, M = 34-2. This corresponds with the molecule H^Og. An earlier determination by
W. Tammann (1899) gave H4O4, but this was afterwards found to be due to the use of an
impure sample.
(3) J. Hertz (1890) found that 2*423 grms. of sulphur in 100 grms. of naphthalene
lowered the freezing jjoint of naphthalene 0641° ; hence show that the molecular weight
SOLUTIONS
567
of sulphur under these conditions corresponds with the formula : Sg- Ansr. : The mole-
cular weight by experiment is 262, and by calculation for Sg, 256.
(4) G. Buchbock (1897) found that a solution of 21 •740 grms. of ethyl ferrocyanide in
a litre of water lowered the freezing point 0*1270°; show that this corresponds with the
formula (C2H5)4FeCy6. The theoretical molecular weight is 328, observed 323.
(5) J. L. R. Morgan and H. K. Benson (1907) find that 0292 grm. of potassium chloride
dissolved in 100 grms. of molten calcium chloride, CaCl2.6H20, as solvent depressed the
freezing point 0-181°. The freezing constant of CaCl2.6H20 is 450. Hence show that
these observations agree with the formula KCl.
(6) G. Marchetti (1899) found that an aqueous solution of 1-703 grms. of a hydrated
molybdenum oxide in 100 grms. of water lowered the freezing point 0*072°. An analysis
of the compound gave 56-9 per cent, of molybdenum. Show that these results agree with
the formula M03O8.5H2O.
E. Beckmann's process for the determination of freezing points.—
Applications of the freezing process for the determination of molecular weights
present no particular difficulty. Several different forms of apparatus have been
devised by investigators who have followed Kaoult, but no
apparatus has proved so useful as the later forms of E.
Beckmann's. Precautions have to be taken against under-
cooling. Satisfactory results too are only obtained with those
solvents which do not separate out with the solute in form of
solid solutions.
Freezing-point determinations are usually made in Beckmann's
apparatus.* The tube A, Fig. 26, with a side neck, B, is weighed,
and about 15 c.c. of the solvent are added, and the tube is weighed
again. Beckmann's thermometer, reading to the 0*01° of a degree.
It is set so that the mercury is near the top of the scale at the freezing
point of the solvent, and it is provided with a reading lens. The
thermometer T and a stirrer S are placed in the solvent, and the whole
arrangement is placed in a glass tube A, which serves as an air jacket.
This is surrounded by a vessel D of water or some liquid at a tem-
perature about 5° below the freezing point of the solvent. This
vessel is fitted with a thermometer Tj and stirrer >S,. The tempera-
ture recorded by the thermometer slowly falls until the solvent begins
to freeze ; it usually falls from 0-2° to 0'3° below the freezing point of
the solvent, and then begins to rise to the freezing point proper. The
thermometer should always be tapped before a reading is taken to
make sure the mercury is not lagging behind. The highest point
reached by the mercury in the thermometer is taken to be the
freezing point of the solvent. Owing to undercooling, it is sometimes
difficult to start the freezing of the solution. In that case, a few
pieces of platinum foil, or a minute fragment of the frozen solvent,
will start the freezing. It is sometimes necessary to introduce a
correction for undercooling as indicated in text-books for the
laboratory. Each determination should be repeated two or three
times and the successive observations should agree within 0-002°
to 0-003°. When the freezing point of the solvent has been deter-
mined, add a sufficient amount of the substance under investigation
to give a depression of 0-3 to 0*5°. After the freezing point has
been determined again, find the freezing point after adding a second and then a third portion
of the substance under investigation.
The molecular weights of volatile substances relative to the weights of the
hydrogen molecule have been determined from the vapour density determinations
and Avogadro's hypothesis. The osmotic pressure and related properties of solutions
enable the molecular weights of liquids and solids in solution to be determined.
There is an extensive choice of solvents, and it is possible to utiHze such widely
different substances as stearic acid, mercury, ether, fused metals, etc. Molten
salts containing water of crystaUization may be used — e.g. sodium sulphate,
Na2S04.10H20 ; calcium chloride, CaCl2.6H20 ; lithium nitrate, LiN03.3H20 ; sodium
chromate, Na4CrO4.I0H2O ; etc. — the results agree with those obtained with other
solvents. The molecular weights of a great many substances in solution are in
agreement with those furnished by the vapour density method, yet there are some
Fig. 26. — Beck-
mann's Apparatus
for Freezing-point
Determinations.
568 INORGANIC AND THEORETICAL CHEMISTRY
irregularities. F. M. Raoult found that the molecular weights of substances in solu-
tions are sometimes greater and sometimes less than what we should expect. Organic
substances, like cane sugar and alcohol in aqueous solution, gave normal values for
their molecular weights, while inorganic salts like potassium and sodium chlorides
gave about half the values which correspond with their normal formula). The results
are then said to be abnormal. F. M. Raoult first assumed that organic substances
must form double molecules in solution, while inorganic salts are normal. He
then tried if extreme dilution would break down the supposed doubled molecules,
but the experiments returned a negative answer.
The relation between the osmotic pressure and the rise in the boiling point or depression o f
the freezing point.— From the Clapeyron-Clausius eciuation,log p — log Pg^{MX/R){dT/TT8),
where dT is put in place of Ts — T ; and T denotes the temperature at which the vapour
pressure is p, and T^ a slightly higher temperature where the vapour pressure is Pg.
Since it has been shown that (p—Pg)lp is an approximation for log p — log Pg, and
{MX/R){dT/TT8) can be substituted for {p—Pg)/p in equation (1), the result reduces to
the relation P=sXdT/T, which shows the relation between the osmotic pressure P and the
rise dT in the boiling point. A similar relation between the depression dT of the freezing
point and the osmotic pressure obtains if A denotes the heat of fusion of the solvent in
place of the heat of vaporization. The simpler form obtained by G. M. Lewis (1908),^
P = 12"06/— 0*21/2 atm., gives the osmotic pressure of a solution which lowers the freezing
point /°.
Example." — If an aqueous solution contains enough solute to raise the boiling point
1°, show that the osmotic pressure is nearly 47 atmospheres, dT = l ; T = 373 ; s = 0'96 ;
and A = 537 cals.
Keferences.
1 F. M. Raoult, Cryoscopie, Paris, 1901 ; Ann. Chim. Phys., (5), 20. 217, 1880 ; (5), 28. 133,
1883; (6), 2. 66, 93, 115, 1884; (6), 4. 401, 1885; (6), 8. 289, 317, 1886; (6), 9. 93, 1886;
Compt. Rend., 102. 1307, 1886; 95. 108, 1030, 1882; 101. 1056, 1885; 125. 751, 1897; Zeit.
phys. Chem., 27. 617, 1898 ; 2. 488, 1888 ; 9. 343, 1892 ; 20. 601, 1896 ; H. C. Jones, ib., 11.
110, 529, 1893; Phil. Mag., (5), 40. 383, 1896; The Freezing-point, Boiling-point, and Con-
ductivity Methods, Easton, Pa., 1897.
* J. H. van't Hoff, Zeit. phys. Chem., 1. 481, 1887.
8 E. Beckmann, Zeit. phys. Chem., 2. 638, 715, 1888 ; 7. 323, 1891 ; 21. 240, 1896 ; 22. 677,
1897.
* C. M. Lewis, Journ. Amer. Chem. See, 30. 668, 1908.
§ 14. The Relation between the Solvent Power of a Solvent and the
Molecular Weight of the Solute
The capacity of a pure solvent A to dissolve another liquid B is reduced when
another substance C is dissolved in B. For instance, an ethereal solution of naphtha-
lene (C) dissolves less water (B) than does pure ether (A). In following out the
analogy between vaporization and solution, W. Nernst (1890)i argued that what-
ever depresses the vapour pressure of A should also depress the solubility of A.
The relative lowering of the vapour pressure of A is proportional to the number of
gram-molecules of non- volatile solute C in A, so the relative lowering of the solu-
bility of B and A should also be proportional to the number of gram-molecules of
the solute C in A when the solute C is not soluble in B. Let S denote the solu-
bility of the liquid B in the pure solvent A ; and Sg the solubihty of the liquid B
in the solution of C in A ; the lowering of the solvent power of A is S—Sg, then,
by analogy with the vapour pressure formula :
p—Ps _ ^ S—Sg __ n
p ~N' S ~N
where n denotes the number of gram-molecules of the solute C in iV gram-molecules
of the solvent A. If w grms. of C are dissolved in A, n=whn, where m denotes the
SOLUTIONS 569
molecular weight of C. Then, (S—Ss)mlw=SIN ; and since S/N is a constant, say
k, whose numerical value can be determined experimentally for a particular tem-
perature and amount of solvent, then, if S^ denotes the lowering of the solubility
S — Ss,
Molecular weigh t= A; ^ . . . • (7)
Experimental details for the application of the principle were worked out by
S. Tollaczko (1895) for substances soluble in ether, and insoluble in water. The
accuracy of the process does not appear to be very great even though the temperature
be maintained constant for all the readings.
Example.— The constant k was found to be 536, and 0*1266 grm. of naphthalene de-
pressed the solubiUty of water in ether about 0*55; accordingly, the molecular weight of
naphthalene is 536 X 0*1266 -^0-55 = 123' — the theoretical value for CjoHg is 128.
If there is an appreciable change in the volume of the solvent on addition of the
solute, a correction factor v/vg is required — v represents the initial volume of the
solvent and Vg its volume after the addition of the solute.
The relation between osmotic pressure and the lowering of the solvent power
of a solvent. — By analogy with formula (3), above, the osmotic pressure P is related
with the relative lowering of the solvent power of the solvent by
P=
S-Ss sRT
S ' M
The connection between the osmotic pressure and the colligative properties of
a solution — the lowering of the vapour pressure, of the freezing point, and of the
solvent power of a solvent, and the raising of the boiling point — has thus been estab-
lished. Consequently, results obtained by these difierent methods cannot be re-
garded as independent evidence supporting any particular hypothesis, for the
different processes are simply different ways of measuring one quantity, and they
must necessarily lead to similar conclusions. The different methods are distin-
guished from one another by the degree of accuracy which can be obtained in the
application of a particular process ; by convenience in the theoretical (mathematical)
treatment ; and by the range of temperature over which they are applicable.
References.
1 W. Nemst, Zeit. phys. Chem., 6. 16, 573, 1890; 8. 110, 1891 ; S. Tollaczko, Ber., 28. 804,
1895.
§ 15. Anomalous or Abnormal Results for the Molecular Weights of
Substances in Solution
We never profit more than by those unexpected results of experiments which contradict
our theories and analogies.— Guyton de Morveau.
I thank God that I was not made a dexterous manipulator, for the most important of
my discoveries have been suggested to me by failures. — ^Humphry Davy.
The analogy between gases and solutes in dilute solution has been pursued
further. If the molecules of a dissolved substance are in a similar state to what
they would be if the substance were in the gaseous condition, the relation between
the pressure, temperature, and concentration will be represented by the expression,
PV=iRT ; or, since the concentration C is inversely as the volume, by PIC=iRT.
As before, if i be unity, the molecules of the substance in solution and in the gaseous
condition are presumably similar ; if i be greater than unity, the analogy with similar
phenomena with gases j)v=iRT, has led to the assumption that the molecules
dissociate when they pass into solution ; and if i be less than unity, the molecules
570 INOKGANIC AND THEORETICAL CHEMISTRY
polymerize. If we apply the uncorrected relation, PIC=RT, it is now easy to see
that if i be greater than unity (dissociation), the osmotic pressure will appear too
high ; and if i be less than unity (polymerization), the osmotic pressure will appear
too low. When we speak of the lowering of the osmotic pressure, we also imply
that the vapour pressure is increased, the boiUng point is lowered, and the freezing
point raised ; and conversely, the raising of the osmotic pressure implies that the
boihng point is raised, and the vapour pressure and freezing point are lowered.
Examples. — (1) F. Vogel (1903) found that the lowering of the freezing point of a 2N-
solution of barium nitrate, Ba(N02)2, was 3-931°, and of a 0-2N-solution, 0*479°. The
specific gravity of the former solution at 21° was 1-1604, and of the latter 1*0017. Calculate
the value of i=M (theoretical)/ikfi (observed) for each solution. Here 50 c.c. of the 2N-
solution contains 11*4740 grms. of barium nitrate and 46*5456 grms. of water; and the
2A/'-solution, 1*1474 grms. of barium nitrate and 489358 grms. of water. The computed
molecular weights of the salt in the two solutions are respectively 97-4016 and 99*7581, the
theoretical value for both solutions is 229*14, hence, i=2-54 for the 2N-solution, and 2*48
for the 0-2N-solution.
(2) Which is the more probable equation for the action of a potassium hydroxide solution
on aluminium hydroxide: KOH+Al(OH)3=Al(OH)20K+H20, or 2KOH+Al2(OH)e
=Al2(OH)4(OK)a + 2H20, when the freezing point of solutions of potassium hydroxide are
notchanged by the addition of aluminium hydroxide ? C. L. Speyers (1898). The former,
because one molecule of aluminate is produced per molecule of potassium hydroxide, etc.
Again, from the argument based upon Fig. 23, it follows that the osmotic
pressure P is to the lowering of the vapour pressure f—jfs of the solvent and
solution as the density s of the solution is to the density D of the vapour ; or
P : j)~pg=^s : D ; and since p—pg^pn/N, and p=DRTIM, where M represents
the molecular weight of the vapour, the osmotic pressure,
According to J. H. van't Hoff's relation, P==nRT, unit volume of all solutions,
with n molecules of the solute, have the same osmotic pressure. If unit volume
of a solution of specific gravity s has N molecules of the solvent of molecular
weight M', then NM'/s=l, or N=slM' ; and hence,
p j^j, Molecular weight of solvent p M „-,
Molecular weight of solute ' ' M
which shows that van't Hoff's relation — P=nRT — is valid only when the molecular
weight M' or the solvent is the same as that of the vapour, M.
Abnormally high osmotic pressures — dissociation oJ solute, or polymerization
of the solvent. — The molecules of the liquid may be more complex than the mole-
cules of the vapour, so that the molecular weight of the liquid molecules is M'=aM,
where a, the so-called association factor, represents the number of normal mole-
cules of molecular weight M which combine to form one molecule of the associated
liquid. The observed osmotic pressure P^ of a solution containing n gram-molecules
of the solute in unit volume of the solvent, whose molecular weight is a times that
of the vapour, will be a times the osmotic pressure P of a solution in which the solvent
is not so associated, or Pi=aP. A very considerable number of aqueous solutions
of acid, bases, and salts furnishes a much greater osmotic pressure than we should
naturally expect. The deviation of a gas from Avogadro's law is usually explained
by assuming that the molecules of the gas are dissociated into simpler forms. Iodine
molecules, I2, at high temperatures appear to behave as symbolized : I2=I+I.
S. Arrhenius (1887) sought to explain deviations of the molecular weights of salts,
acids, and bases in aqueous solutions by assuming that the molecules are dis-
sociated into simpler parts. The molecules of sodium chloride, for instance, are
supposed to be dissociated in aqueous solutions into two parts^ — Na and CI. The
idea came as a surprise, and much opposition has been raised against this interpre-
tation of the results, because there are no signs of chemical action which might be
SOLUTIONS
571
expected if the molecule of sodium chloride were dissociated into electrically charged
Na+ and CI— atoms on solution in water. Accordingly, other hypotheses have
been invented to make the first hypothesis fit the facts. In spite of this, Arrhenius'
hypothesis at once explains in a seductive and plausible manner the abnormally
high osmotic pressures obtained for these substances. There is a strange coinci-
dence. Arrhenius determined the value of i — the number of molecules in the
above equations — for ninety different substances. He noticed at once that these
substances could be roughly divided into two classes : those which gave values of
i nearly unity were either non-conductors or poor conductors of electricity ; whereas
those which gave values of i materially greater than unity were fair or good con-
ductors of electricity. In the following table is taken to represent, within the
limits of experimental error, the relative number of molecules formed when one
molecule of the substance is dissolved : —
Table VI.- — Normal and Abnormal Osmotic Pressures.
Non-conductors.
Conductors.
Substances in solution.
i
Substances in solution.
i
Methyl alcohol ....
Mannite .....
Cane .sugar ....
Ethyl acetate ....
Acetamide ....
0-94
0-97
1-00
0-96
0-96
Calcium nitrate
Magnesium sulphate
Strontium chloride
Potassium chloride
Lithium chloride .
2-48
1-25
2-69
1-81
1-92
We naturally inquire : What connection, if any, subsists between the alleged
dissociation of the molecules of a substance in a solution and the conduction of
electricity ? How can one molecule of sodium chloride, one molecule of lithium
chloride, each furnish what appears to be two molecules when dissolved in water ?
Is this dissociation hypothesis the only possible explanation which covers all the
facts ?
Chemical theory of osmotic pressure. — The circumstantial evidence adduced
to show that in liquid water at any given temperature, it is highly probable (i) that
definite relation subsisting between polymerized and simple molecules ; and that
(ii) the presence of salts in solution displaces the equilibrium in favour of the simpler
molecules. Hence, with some differences, J. J. van der Laar (1906) i and H. E. Arm-
strong (1902-6) argue that the osmotic pressure is not a result of a pressure produced
by the molecules of the dissolved substance, but rather a secondary effect of the
reduction in the concentration of the more complex water molecules by the dissolved
substance which causes the pure solvent, say, water, to travel towards the water
of the solution. Here, osmotic pressure is a consequence of the impulse of diffusion
which continues until the concentration of the complex water molecules is the same
on both sides of the separating membrane. A certain proportion of the complex
molecules are supposed to be continuously resolved into simpler molecules as the
pure solvent diffuses through the partition, and conversely, a certain proportion
of the simple water molecules which diffuse from the solution side of the partition
are polymerized on the pure solvent side. The addition of a non-electrolyte disturbs
the equilibrium (H20)n^wH20 in the direction (H2^)n'->'^^2^ , and the resulting
molecules of hy drone, H2O, exert an attraction on similar molecules in the region of
the pure solvent or more dilute solution, so that the solvent passes through the
membrane to the solution until equilibrium is re-established. Osmotic pressure is
then a measure of the disturbance of the equilibrium, (H20)n=^H20, produced by
the continuous depolymerization of the water molecules on the solution side of the
partition, and a continuous polymerization of water molecules on the pure solvent
side. The greater the difference in the concentration of the more complex water
572 INORGANIC AND THEORETICAL CHEMISTRY
molecules on the two sides of the partition, the greater is the impulse of diffusion,
and the greater the osmotic pressure. Electrolytes exert an attractive influence on
the solvent, which is superposed on the effect produced by dissociated water com-
plexes, so that solutions of electrol}iies exert a greater osmotic pressure than solutions
of non-electrolytes of equivalent molecular concentrations. H. E. Armstrong (1909)
explains the selective action of the membrane in the following manner :
The compounds which penetrate the membrane are all substances which attract water
presumably only to a minor extent and which exist to some extent in solution in an un-
hydrated condition ; those which cannot penetrate it, on the other hand, probably all form
hydrates of considerable stability in solution. I picture surfaces generally, colloidal surfaces
in particular, as not merely wetted by water, but as more or less hydronated and hydrolated,
that is to say, they are not merely wetted by water complexes, but associated with hydrone,
the simple fundamental molecule of which water is composed. The intra-molecular passages
in a colloidal membrane, if thus hydrolated, would be guarded by the attracted hydrone
molecules ; and the hydrolated molecules in a solution which attempted to effect an entiy
through such passages, would be seized upon and held back in virtue of the attraction which
two hydrolated surfaces- — ^that of the membrane and that of the solute — would exercise
upon one another. The hydrolated passages, however, would be indifferent to molecules
which were not hydrolated — consequently such a substance as acetic acid, of which probably
only a small proportion is present in solution in the hydrolated state, would gradually pass
through them.
H. E. Armstrong further assumes that the reason non-electrolytes in equivalent
concentration all exert the same osmotic pressure, while, in contradistinction, elec-
trolytes exert an excessive pressure, turns on the assumed fact that non-electrolytes
dissolve in water in the form of their fundamental molecules, which exerts no
appreciable attractive action (affinity) on the water molecules ; and that electro-
lytes have the power to attract water molecules in proportion to their efficiency as
electrolytes. It is inferred that the process of solution implies a strong affinity of
a chemical or quasi-chemical nature in order to break up the cohesion of the crystal.
The molecules of a compound like mercuric chloride which dissolves in water with
difficulty probably differ from the molecules of, say, calcium chloride in exerting
but a slight attractive force on molecules of water, while the so-called ionized mole-
cules are assumed to form complex reversible systems of solvent and solute, say :
K-Cl + 0-OH ^ K-Cl<^jj
The observed osmotic effects are thus assumed to be an effect of two distinct forces :
(1) the depolymerization of the complex water molecules ; and (2) the hydration of
the solute whereby the solvent is attracted towards the solute by the same forces
which cause it to dissolve in the water. The function of the electric current is to
break down such systems by drawing certain groups of atoms apart. As these
systems are broken down by the current new ones spontaneously form in the
solution. Molecules which exhibit no tendency to form such systems are virtually
non-electrolytes. Thus, in the words of M. Faraday, " the inability of a system
to suffer decomposition by electricity . . . may be dependent upon the absence of
that mutual relation of the particles which is the cause of aggregation." J. Larmor
has also shown that the osmotic Uws deduced for dilute solutions by thermodynamics
are the same, even supposing that a complex between solvent and solute is formed,
provided that the solution is so dilute that each such complex is for very much the
greater part of the time out of the sphere of influence of other complexes.
V. A. Vilde and A. J. Bogorodsky (1915) ^ explain the abnormal depression of
the freezing point of aqueous solutions by assuming with A. Hantzsch (1907) that
just as ammonia may attach itself to a hydrogen atom forming ammonium, NH4,
so can a molecule of water similarly form hydroxonium, called for brevity hydronium,
OH3. When a little water is dissolved in absolute sulphuric acid, H2SO4, hydronium
sulphate, OH3.HSO4, is formed : HS04H+H20=0H3.HS04 ; just as ammonia
under similar conditions reacts : HS04H-|-NH3=NH4.HS04. Assuming, how-
ever, that water is polymerized and contains, say, trihydrol molecules, (H20)3, then,
SOLUTIONS 573
when a little acid, HR, is dissolved therein, (H20)3+HR=H30.R+H30.0H;
with a base, M.OH, the action is represented : (H20)3+M.OH=H30.0H
+H2MO.OH ; and with a salt, MR, similarly : (H20)3+MR=H30.R+H2M0.0H.
In dilute solution, with the molecules (H20)3 in large excess, the effect of such
changes would make the solution behave as if each molecule of the solute formed
two molecules on passing into solution. The number of molecules is increased owing
to the reaction between solvent and solute. An analogous case occurs when various
salts or oxides are dissolved in fused alkali hydroxides : wKOH-j-PbO=Pb(OK)2
+H20+(w-2)KOH.
There are available at least three explanations of an abnormally high osmotic
pressure : (i) Dissociation of the solute ; (ii) Polymerization of the solvent ; and
(iii) Reaction between polymerized solvent and solute. Observations on osmotic
pressures and the related phenomena — raising of the vapour pressure, raising of
the boiling point, lowering of the freezing point, etc. — do not discriminate between
these three explanations unless it be to decide against the hypothesis that the mole-
cules of the electrolytes are dissociated, both on account of its revolutionary nature,
and a priori improbability. It is therefore necessary to seek evidence from other
sources. The solvent water is undoubtedly associated, and it has been shown that
the degree of association is diminished by the dissolution of salts in the liquid. It
has not yet been shown that the degree of association of dijSerent solutions varies
concomitantly with the changes in their osmotic pressure (or related phenomena) ;
it has, however, been shown that solutions of salts in non-associated solvents — e.g.
hydrocarbons and their halogen derivatives, ethers, esters, most aldehydes and ketones
— do not usually show abnormally high osmotic pressures ; while solutions of salts
in associated solvents — e.g. water, compounds containing hydroxyl groups like the
alcohols, organic acids, oximes, some nitro-compounds, and compounds containing
amidogen groups — do show phenomena corresponding with abnormally high osmotic
pressures. It will probably be found in later years that the abnormal behaviour of
water retarded the development of the generalizations concerning the effect of solutes
on the depression of the freezing points, the elevation of the boiling points, and the
lowering of the vapour pressures of solutions. F. M. Raoult's success followed as a
result of his experiments with non-associated solvents where the main phenomenon
was not obscured by secondary effects. If it turns out that electrolytes are salts
which are in the normal or unassociated condition when dissolved in associated
solvents, the so-called ionization hypothesis which has been elaborated on the
assumption that electrolytes have molecules which are dissociated into electrically
charged units when in solution, will be unnecessary. It is also remarkable how
tenaciously the ionization hypothesis is held when it is remembered how very
many properties of electrolytes, supposed to be uniquely and adequately explained
by the hypothetical dissociation of the solute can be satisfactorily explained by the
observed association of the solvent.
Abnormally low osmotic pressures — ^polymerization of solute. — If the liquid
solvent and its vapour have the same molecular weight, and the solute is so associated
that it has an association factor, i^ then, in place of n molecules of the solute, there
are in reality nji molecules, and the observed osmotic pressure P2 will be i times
smaller than the osmotic pressure P would be if the solute were not associated,
or, P=iP2. F. M. Raoult (1886) noticed that the depression of the freezing point
of a solution of alcohol in benzene, naphthalene, nitrobenzene, ethylene dibromide,
etc. — ^is just about half what we should expect if the molecules of alcohol were
represented by the regular formula : C2H5OH. This means that the molecules of
alcohol — C2H5OH — in benzene solutions are doubled, and may be represented
by C4Hio(OH)2 in benzene solution. This phenomenon is common with molecules
possessing hydroxyl, OH, groups — e.g. alcohols and acids. Formic — H.CO.OH—
and acetic — CH3.C0.0H^ — acids, and indeed water, behave in a similar manner,
and we know that if these substances be vaporized, they appear to have twice
the molecular weight they would have if their molecules could be really
574 INORGANIC AND THEORETICAL CHEMISTRY
represented by the ordinary formulae. Otherwise expressed, the molecules are
polymerized. Increasing dilution usually reduces the abnormality indicating that
the association becomes smaller and smaller ; W. Ramsay and J. Shields also noticed
that the dilution of associated molecules has a similar effect to an increase of tem-
perature. It is also necessary to remember that the dissolved substance may freeze
out along with the solvent so as to form a kind of solid solution — for example, the
ice which separates from a dilute solution of ether or alcohol in water contains
appreciable quantities of the solute. The observed reduction of the freezing point
of the solution may be less than that calculated from the regular molecular formula
of the dissolved substance. J. H. van't Hoff (1890) demonstrated the elevation of
the freezing point from this cause. In 1889, W. Ramsay 3 determined the
lowering of the vapour pressure of mercury by the solution of a number of metals
— lithium, sodium, barium, magnesium, zinc, aluminium, tin, lead, antimony,
manganese, silver, gold, etc. — and found the results agreed with the assumption
that the mefcals are in the atomic condition in mercurial solution. C. T. Heycock
and F. H. Neville examined solutions of several metals in molten tin and sodium.
There is abundant evidence of the existence of definite compounds of mercury with
the metals, the alkalies, or the alkaline earths, and there is no reason to assume
that these compounds break up when in mercurial solution. If there is no evidence
of the formation of chemical compounds, it might be assumed that the lowering
of the freezing point is an indication of the molecular state of the metals — e.g.
zinc, cadmium, bismuth, lead, and tin, dissolved in mercury.
If both liquid solvent and solute are associated, it follows from the above discussion
that the osmotic pressure P,, will be equal to Paj^ ; and if o = j8, the osmotic pressure will
appear to be normal, but both solvent and solute would have polymerized. It also follows
by analogous reasoning that J. H. van't Hoff's relation, k = {)'{y2T^lX applies only to the
unassociated solvent and solute ; if the solvent be alone associated, this relation becomes
0-02r2„y^ . and if the solute is alone associated, 0-02T^/fiX.
References.
^ J. J. van der Laar, Proc. Acad. Amsterdam, 9. 53, 1906 ; H. E. Armstrong, Proc. Roy. Soc,
78. A, 264, 1906 ; 79. A. 564, 576, 579, 586, 1907 ; 81. A. 80, 102, 1909 ; Trans. Faraday Soc,
3. 30, 1907 ; M. M. Garver, Journ. Phys. Chem., 15. 20, 1911 ; J. Larmor, Nature, 55. 545, 1897-
2 A. Hantzsch, Zeit. phys. Chem., 61. 257, 1907 ; 65. 41, 1908 ; 68. 204, 1909 ; V. A. Vilde
and A. J. Bogorodsky, Journ. Russian Phys. Chem. Soc. 47. 373, 1915.
3 W. Ramsay, Journ. Chem. Soc, 55. 521, 1889 ; C. T. Heycock and F. H. Neville, ih., 55.
666, 1889 ; 57. 376, 656, 1890 ; C. Tammann, Zeit. phys. Chem., 3. 441, 1889.
§ 16. The Cause of Solution
There is probably no sharp line of demarcation between chemical union and the process
of solution.— S. L. Bigelow (1907).
Clear answers to the questions : What determines the solubility of a substance ?
Why is caesium chloride, CsCl, so very soluble and silver chloride so very sparingly
soluble in water ? are not forthcoming. The alchemists considered solution to be
produced by the penetration of the particles of the solute between the particles of
the solvent, or to the result of some kind of union of the two. P. Gassend,i following
the old atomic theory, taught in 1647 that the atoms of one substance could enter
the interstices between the atoms of another substance. The cubic particles of com-
mon salt, for example, filled up the water molecules, and when all such pores were
filled, the water could dissolve no more salt ; similarly the octohedral alum. Hence,
it was supposed that water also contained octahedral pores which the alum but not
the salt could enter. The ideas of solution prevaiHng in the seventeenth century
may be gathered from R. Boyle's essay : Of the producibleness ofchymical 'principles
(Oxford, 1663). He says :
Dissolution depends not so much on the pretended cognation between the solvent and
SOLUTIONS 575
the body it is to work upon as upon the congruity, as to size and figure, between ihe porea
of the latter and the corpuscles of the former.
A similar view of solution was advocated about the same time by N. Lemery
in his Cours de chymie (Paris, 1675). Near the beginning of the eighteenth century
Isaac Newton naturally extended the idea of attraction to the particles or atoms
composing a substance ; and, reasoning from the fact that the particles of a salt
dissolving in water, in spite of their greater density, diffuse themselves uniformly
throughout the solvent so that the particles of the salt recede from one another
and endeavour to get as far asunder as the solvent will allow, Isaac Newton
asked : Does not this^ endeavour imply that the particles of the dissolved
substance exert a repulsive force on one another, or else attract the particles of
the solvent more strongly than they do one another ? Otherwise expressed :
Does not the solute dissolve because it exerts an attraction on the solvent ?
H. Boerhaave (1772) took a somewhat similar view, and added :
The particles of the solvent and those of the dissolved substance unite, after solution,
to form a new homogeneous substance. . . . The cause of this must be sought for in both
the solvent and the dissolved substance. It is common to them both. Particulce solventes
et solutce se affinitate suoe naturce coUigant in corpora homogenea.
About this time, J. K. Wallerius (1772),2 M. H. Klaproth (1806), and many others
also regarded solution as the result of the action of chemical affinity between solvent
and solute which must be stronger than the cohesion of the particles of the dissolved
substance each to each.
In 1789, A. L. Lavoisier distinguished between dissolution, a chemical 'process ,
typified by the dissolution of zinc in dilute acid ; and solution, a physical 'process,
typified by the solution of salt in water. In the latter process, said A. L. Lavoisier,
the salt molecules are simply torn apart from one another, but neither the solvent
nor solute suffers any other change, because both can be recovered in the same
quantity as before the operation. These views were adopted by A. F. de Fourcroy
(1801), who regarded a solution as a physical or mechanical mixture in which the
dissolved substance is in a state of fine subdivision in the solvent. With reference
to purely physical hypotheses of this kind, W. A. Tilden adds :
Such a theory, however, serves to account only for the initial stage in the process of
solution, and does not explain the selective power of solvent, nor the limitation of the
solvent power of a given liquid, etc.
The father of the chemical school, C. L. BerthoUet (1803), ^ took the contrary
view. A solution, said he, is a true chemical compound which is not very stable,
and which shows the characteristics of the dissolved body. Solution must be due
to a force which is great enough to overcome the cohesion of the dissolved substance.
The difference between a solution -and a chemical compound is to be found in the
firmness of the union of the parts. In solutions the parts are the less firmly united,
and the characteristic properties of the dissolved substance have not been lost.
J. P. Cooke 4 expressed similar views in 1881. C. L. BerthoUet, it will be remembered,
did not accept the laws of definite and multiple combination ; hence he could say :
Chemical union and solution must therefore follow the same laws.
After the laws of combination had been generally accepted, it became necessary
for those who regarded solution as a chemical process to explain the reason why
chemical combination takes place in certain definite proportions, while solution
occurs in any proportion up to a certain limit. This was done by J. L. Gay Lussac,^
who assumed that the force which leads to chemical combination is more powerful
than that which produces solution. He compared the process of solution with that
of vaporization — both are dependent on temperature and obey its variations. The
essential difference between the two processes consists in this : Gaseous molecules
do not need a solvent to maintain them in a given space, their own repulsive force
suffices for this purpose. On the other hand, in a sohition, the particles of the
576 INORGANIC AND THEORETICAL CHEMISTRY
dissolved substance cannot maintain themselves in the space if they are not united
by affinity to the molecules of the solvent.^
These ideas were fairly general during the greater part of the nineteenth century,
and they are typified by the views of L. Dossios,^ who referred solution to the
molecular attraction between unlike molecules exceeding the attraction between
molecules of the same kind. Consider the case of two fluids A and B in contact
with one another. Molecules of the one fluid A will enter among the molecules of
the other B if the molecules of the latter attract the former more strongly than the
attraction of either the molecules of A or B have for one another. When the number
of molecules of A passing into the fluid B is equal to those passing from B in a given
time is equal, the liquid B is saturated ; similar considerations apply to the fluid
A and the molecules of B. Finally, therefore, each hquid will be saturated with
the other. The same remarks apply, fnutatis mutandis, to the solution of a soHd in
a liquid. The solubihty will increase with a rise of temperature, if the temperature
diminishes the attraction of Hke molecules for one another more than it does for
unHke molecules. If the solution of a salt in water is a consequence of the attrac-
tion of the molecules of water for a molecule of the salt exceeding the attraction
of the molecules of a salt for one another, it follows, says W. W. J. Nicol ® :
As the number of dissolved salt molecules increases, the attraction of the dissimilar
molecules is more and more balanced by the attraction of the similar molecules ; when these
two forces are in equilibrium, saturation takes place.
D. I. Mendel^efi ^ modified the views of C. L. Berthollet and J. L. Gay Lussac
so as to bring the conception in accord with the law of definite proportions and the
phenomena of dissociation. He supposed that the solvent and solute formed chemical
compounds with so feeble an affinity that the compounds are more or less dissociated
at ordinary temperatures. According to the nature of the solute one or more such
compounds may be simultaneously in the solution. M. Berthelot (1879) ^^ also
advocated the view that a solution is a mixture of the free solvent and a compound
of solvent and solute — called, in the case of aqueous solutions, a hydrate — united
in definite proportions. Les phenomenes de la dissolution normale Lont en quelque
sorte intermediares entre le simple melange et la comhinaison veritable. Definite
hydrates are formed in solution by the union of the solvent water with the anhydrous
salt, and they are comparable with the crystalline hydrates of the salt, but with
this difference, the hydrates in solution are in une etat de dissociation partielle which
depends on the relative quantities of solute and solvent and the temperature in
accord with the law of mass action. In some cases, he added, the whole mass of
the solvent takes part in the action as in the case of strong acids ; and in other cases,
as with salts of the alkalies, only a part of the solute is combined with the solute ;
and lastly, there may be many hydrates, or equivalent bodies existing in the solution
in equilibrium at the same time.
D. I. MendeleefE could see no way of definitely determining the composition
of the supposed hydrates, and his attempt to determine the character of the
hydrates was not very successful. He plotted the specific gravities of
solutions of sulphuric acid of varying concentrations against the concentrations,
and assumed that the maxima in the curves represented definite hydrates.
S. U. Pickering ii also examined the physical properties of numerous aqueous solu-
tions, and when one of these properties is plotted against the concentration he
inferred that (i) abrupt changes in the curvature of the resulting curves, and (ii) the
supposed points of discontinuity of the first or second differential coefficient of a
function connecting the physical property of a solution with concentration, indicated
the presence of certain definite hydrates. ^2 gome of these discontinuities are possibly
due to experimental errors, since a small error may be much magnified in the
differential process. Definite breaks do exist in many curves connecting the com-
position with some physical property — specific heat, specific gravity, viscosity,
refractive index, electrical conductivity, compressibility, surface tension, thermal
SOLUTIONS 577
expansion, heat of solution, etc. — but striking irregularities are encountered, when
the attempt is made to connect these with specific hydrates. For example, Picker-
ing's hydrates did not always correspond with those deduced by other physical
methods — e.g. by the freezing points of the solutions.i^
According to T. M. Lowry, the breaks observed in the density-composition curves
of dilute solutions are due to distortions produced by alterations in the character of
the solvent water whereby complex water molecules are depolymerized.^* Apart
from the question whether the hydrates actually assumed are those really present,
the underlying hypothesis that the cause of solution is dependent on the chemical
affinity between solvent and solute is very generally accepted.
In the ionic theory, to be described later, it is assumed that in dilute solutions
the molecules of the salt, presumably after the salt has dissolved, are more or less
dissociated into parts which carry electric charges, and which are called ions ; and
in the so-called solvate theory, the ions as well as the non-ionized molecules are
supposed to be more or less hydrated, so that aqueous solutions of a salt may contain
hydrated molecules, and hydrated ions.
The formation of hydrates or of complexes between solute and solvent is evi-
denced by the so-called abnormal effects produced by the solute upon the freezing
point, boiling point, vapour pressure, and solvent power of a solvent ; by the
distribution of a solute between two solvents, or a solvent and a gas (Henry's law) ;
breaks and irregularities in the solubility curves; by deviations in the physical
properties — specific gravity, thermal expansion, heat of solution, specific heat,
surface tension, viscosity, conductivity, compressibility, index of refraction, mag-
netic rotation of the plane of polarization, diffusion, ionic velocities, hydration
and colour changes, etc. ; effect of salts on the velocities of reactions in aqueous
solutions, etc.
References.
1 P. Walden, Die Losungstheorien in ihrer geschichtlichen Aufeinanderfolge, Btuttg&it, 1910;
H. C. Jones, The Nature of Solution, London, ]917 ; S. Arrhenius, Theories of Solution, New
Haven, 1912; P. Gassend, Opera, Florentise, 1684; 1. Newton, Opticks, London, 1704; H. Boer-
haave, Elementa chemice, Ludguni Batavorum, 1732.
2 J. K. Wallerius, Physische Chemie, Schleusingen, 1772 ; F. A. C. Gren and M. H. Klaproth,
Systematisches Handbuch der gesamten Chemie, Halle, 1806 ; A. L. Lavoisier, Traite eUmentaire de
chimie, Paris, 1789 ; A. F. de Fourcroy, Systkme des connaissances chimiqueSf Paris, 1801 ; W. A.
Tilden, B. A. Rep., 444, 1886.
' C. L. Berthollet, Essai de statique chimique, Paris, 1803.
^ J. P. Cooke, Principles of Chemical Philosophy, Boston, 1881.
^ J. L. Gay Lussac, Ann. Chim. Phys., (2), 70. 424, 1839.
6 W, A. Tilden and W. A. Shenstone, Phil. Trans., 159. 30, 1884.
' L. Dossios, Zurich. Vierteljahrsschr., 13. 1, 1868.
8 W. W. J. Nicol, Phil. Mag., (5), 15. 91, 1883 ; (6), 16. 128, 1883; (6), 17. 537, 1884; (5), 21.
70, 1886.
* T>. 1. Mendeleeff, On the Chemical Combination of Alcohol and Water, St. Petersburg, 1865 ;
Pogg. Ann., 138. 103, 230, 1869 ; Journ. Russian Phys. Chem. Soc, 1. 9, 1869 ; 3. 248, 1871 ;
7. 147, 1875 ; 16. 93, 184, 455, 643, 1884 ; 18. 4, 64, 1886 ; 19. 242, 1887 ; Zeif. phys. Chem., 1.
273, 1887 ; Journ. Chem. Soc, 51. 782, 1887 ; The Principles of Chemistry, St. Petersburg, 1868 ;
London, 1892.
^" M. Berthelot, Essai de m^canique chimique fondce sur la thermochimie, Paris, 1879.
" S. U. Pickering, Proc. Chem. Soc, 1. 122, 1885 ; Journ. Chem. Soc, 51. 290, 593, 1887 ; 57.
64, 331, 1890 ; 63. 99, 141, 890, 1893 ; B. A. Rep., 311, 1890.
12 S. U. Pickering, Journ. Chem. Soc, 51. 290, 593, 1887 ; 57. 64, 331, 1890 ; Chem. News, 61.
305, 1891 ; 64. 1, 311, 1891 ; Ber., 25. 1104, 1892 ; Phil. Mag., (5), 32. 90, 1892 ; (5), 33. 132,
426, 1892 ; S. Lupton, ib., (5), 31. 418, 1891 ; J. F. Heyes, ib., (5), 31. 99, 1891 ; A. W. Riieker,
ib., (5), 32. 304, 1892 ; (5), 33. 204, 1892 ; T. M. Lowry, Science Progress, 3. 124, 1908.
1' B.. G. Jonea, Zeit. phys. Chem., 13. 419,1894; Amer. Chem. Journ.,lQ. 1,1894; S. Arrhenius,
Phil. Mag., (5), 28. 33, 1880 ; D. I. Mendeleeff, Ber., 16. 386, 1886.
1* T. M. Lowry, Phil. Trans., 205. 253, 1905.
VOL. I. 2 P
578 INORGANIC AND THEORETICAL CHEMISTRY
§ 17. The Physical Properties of Solutions
In a sense it is unfortunate that such an enormous number of observations
have been made on aqueous solutions, and so few with other solvents, because water
is so emphatically exceptional in its physical and chemical properties, and the
mechanism of solution will never be clearly demonstrated until these observations
have been supplemented by prolonged series of investigations with less complex
non-aqueous normal solvents. Nearly every physical property of water is pro-
foundly modified when it is used as a solvent for the various salts, and the evidence
as to the formation of hydrates is ambiguous, in that it may be interpreted to imply
that the complex molecules of this solvent are more or less depolymerized in the
presence of a solute.
According to the law of mixtures, the molecular physical properties Z of a
mixture are additive if Z=ZiiVi+Z2iV2+ . . ., where Z^, Z2, . . . represent the
magnitude of a molecular physical property of the components, and Ni, N2, . • .
denote the molecular fractions of the corresponding components of the mixture.
Deviations from this rule are attributed to changes in one or more of the components
— e.g. polymerization or depolymerization of one or more components of the mixture
— or to the chemical union of two or more of the components of the mixture.
The specific gravity of solutions ; the molecular volumes of salts in solution.
— The early observers, P. Gassend, A. Nollet, and M. EUer beheved that salts
dissolve in water without a change in volume, but R. Watson demonstrated that
the assumption has no foundation in fact.i Jn 1840, J. Dalton discovered that a
contraction occurs when some salts dissolve in water ; and in some cases, the con-
traction is as large as that of the volume of the anhydrous salt in solution, so that
the volume of the solution is not greater than that of the solvent alone. He
experimented with hydrated and anhydrous salts and concluded ;
I have tried the carbonates, the sulphates, the nitrates, the muriates or chlorides, the
phosphates, the arseniates, the oxalates, the citrates, the tartrates, the acetates, etc., etc.,
and have been uniformly successful ; only the water adds to the bulk, and the solid matter
adds to the weight.
J. Dalton said of this observation (184:0) : " This fact is new to me and I suppose
to others. It is the greatest discovery that I know of next to the atomic theory ; "
but, added L. Playfair, " Dalton was here inclined to generalize much further than
the observation would bear. There is, indeed, a class of salts which behaves in
this manner — magnesium, copper, zinc, and iron sulphates and a few other
salts Uke sodium borate and phosphate." Apparently unknown to J. Dalton,
E. Swedenborg (1721), a century earher, made a similar observation as is indi-
cated in a previous citation. J. Dalton's work was followed up by S. Holker
(1844), L. Playfair and J. P. Joule (1845), J. C. G. de Marignac (1846), and others.
The degree of contraction varies with different salts, and in some cases — e.g. Hme
water — ^the volume of the solution is even less than that of the contained solvent.
The contraction which occurs during the formation of a concentrated solution
continues as the solution is diluted and either approaches a constant value in a
dilute solution, or else becomes negative so that the volume of the solution is less
than the original volume of the water used as a solvent.^ In some cases — e.g.
silver nitrate — the solution occupies almost as great a volume as the sum of the
volumes of salt and water, whereas with solutions of sugar and water the solution
occupies the same volume as the joint volume of the two components in accord with
the law of mixtures. Again, solutions of organic compounds in hydrocarbons,
though seldom strictly conformable to the law of mixtures, show but shght devia-
tions.3 P. A. Favre and C. A. Valson, I. Traube, and J. Y. Buchanan ^ showed
that with lithium bromide and nitrate and the ammonium hahdes and nitrate, the
volume of the solutions is greater than the sum of the volumes of the water and dry
salt. According to G. P. Baxter and C. C. Wallace, lithium and ca)sium hahdes
SOLUTIONS
579
also produce expansion during solution ; lithium chloride is exceptional, but it too
produces an expansion if the concentration is high and the temperature over 25°.
Does the contraction which occurs when a solid is dissolved in water or when
water is progressively added to its solution indicate the formation of hydrates ?
The answer is in the negative. The curves shown in Fig. 27 represent the differ-
ences between observed molecular volumes and those calculated on the assumption
that the law of mixtures obtains for sul-
phuric anhydride with up to ten molecules
of water, SO3+IOH2O ; and for sodium
chloride with up to NaCl+100H20. Hence,
there is no point in the curve where we
should be justified in setting up a distinction
between the effect due to chemical com-
bination, and that due to other causes.^
Table VII, by G. P. Baxter and C. C.
Wallace, shows the change in volume a
which occurs during the formation of the
solid salt from the solid or liquid elements,
and the change in the molecular volume h
of the salt during the solution of the alkali
halides. The change in volume which occurs fig. 27.— Contraction during Solution
during the solution of a salt in water is a and Progressive Dilution,
highly complex phenomenon. According to
P. A. Favre and C. A. Valson, the observed change of volume during solution
is the joint result of two opposing influences : (i) the contraction of the solvent
under the influence of the solute, and (ii) an increase in the volume of the
salt par suite de la dissociation plus ou moins avancee de ses elements con-
stituents. The observed contraction which usually occurs is taken to prove
that the first effect is usually greater than the second. The contraction is
Table VII. — Changes in the MoLEcuiiAB Volumes of the Alkali Halides DURiNa
THEIR Formation from their Elements, and during Solution.
Salt.
Mol. wt.
Sp. gr.
fused salt
at 25°.
Sum of
at. vols.
Mol. vol.
c.c.
Contraction
in forma-
tion of
Change in
vol. during
solution
a+b c.c.
solid— a c.c.
at25°«=6c.c.
LiCl .
42-40
2-068
38-1
20-5
-17-6
-2-03
-19-6
LiBr .
86-86
4-364
38-7
25-1
-13-6
+0-16
-13-4
Lil .
133-86
4-061
38-8
33-0
- 5-8
+ 3-40
- 2-4
NaCl .
58-46
2-162
48-7
270
-21-7
-8-48
-30-2
NaBr
102-92
3-203
49-3
33-1
-17-2
-6-94
-241
NaT .
149-92
3-665
49-4
40-9
- 8-5
-4-60
-130
KCl .
74-56
1-988
70-4
37-5
-32-9
-8-71
-41-6
KBr .
11902
2-749
71-0
43-3
-27-7
-7-72
-35-4
KI
166-02
3-123
71-1
53-2
-17-9
-6-31
-24-2
RbCl .
120-91
2-798
80-8
43-2
-37-6
-9-19
-46-8
RbBr .
165-37
3-349
81-4
49-4
,-32 0
-8-70
-40-7
Rbl .
212-37
3-550
81-5
59-8
-21-3
-7-86
-29-2
CsCl .
168-27
3-974
96-0
42-4
-53-6
-1-09
-54-7
CsBr .
212-73
4-433
96-6
47-9
-48-7
0-00
-48-7
Csl .
259-73
4-509
96-7
67-6
-39-1
+ 1-77
-37-3
usually attributed to hydration. G. Tammann^ argues that since solutions
behave in approximately the same way when subjected to changes of tempera-
ture and pressure, as the same volume of water, at a higher pressure, there must
be a compression of the water by the solute owing to an increase of internal
pressure which he calls the Binnendruch. It is not clear whether the Binnendruck
is exerted throughout all the water or only to the portions in the vicinity of the
580 INOKGANIC AND THEORETICAL CHEMISTRY
solute molecules. This, howe'ver, is immaterial from the present point of view.
G. P. Baxter argues from T. W. Richards' hypothesis of compressible atoms that
under the influence of chemical affinity the atoms are" more or^^less' compressed- —
the greater the affinity, the greater the compression. This agrees with the observed
thermal values of the reactions. For related elements, the greater the heat of forma-
tion the greater the difierence between the sum of the atomic volumes and the
observed molecular volume of a salt. During solution, the compression due to
chemical affinity and molecular cohesion is more or less reUeved. In addition to
the contraction due to hydration, the expansion due to dissociation (or ionization),
and the expansion due to the partial release of the compression (T. W. Richards'
hypothesis), profound changes are produced in the degree of polymerization of water
during the dissolution of a salt.
C. A. Valson (1874) '^ noticed the curious fact that the differences between the
specific gravities of solutions containing one gram-equivalent of various salts per
litre of two specific metals with one acid are equal, and therefore independent of
the nature of the acid ; and conversely the differences between the specific gravities
of solutions of various salts of two specified acids with one metal are equal, and
therefore independent of the nature of the metal. Hence, the specific gravity of a
normal salt solution — ^hat is, a solution containing one gram-equivalent of the salt
per litre — is obtained by adding two numbers to the standard value — one is a character-
istic or modulus of the metal, and the other is a characteristic or modulus of the
acid. C. A. Valson used a solution of ammonium chloride of specific gravity 1*015
as his standard of reference, but water would have been the better standard.
Table VIII.— Valson's Moduli.
Ammonium .
. 0-000
Manganese .
. 0-037
Chlorine
. 0-000
Potassium
. 0-030
Iron .
. 0-037
Bromine
. 0-034
Sodium
. 0-025
Zinc .
. 0-041
Iodine .
. 0-064
Calciima
. 0-026
Copper
. 0-042
Sulphate
. 0-020
Magnesium .
. 0-020
Cadmium
. 0-061
Nitrate .
. 0-015
Strontium
. 0-055
Lead .
. 0-103
Carbonate
. 0014
Barium
. 0-073
Silver .
. 0-105
Bicarbonate .
. 0-016
Thus the specific gravity of a solution of silver nitrate is 1'015+0'105-f 0*015=1*135.
The rule is valid only for dilute solutions. The observed irregularities depend upon
the regular volume changes which accompany the formation of salts in solution.
It has been found that at a given concentration, within the limits of observational
errors, the physical properties of dilute aqueous solutions of strong electrolytes- —
e.g. volume changes on mixing, the thermal, optical, and other properties — are purely
additive functions of the constituent ions. Consequently, it is assumed that in these
solutions, the electrolytic solute is completely ionized. For example, if the partial
volumes of HCl, HBr, and KBr are known in Y^iV-solution, the partial volume of
KCl in the same concentration can be computed. C. Bender ^ tried to extend the
rule to concentrated solutions by showing that the difference between the specific
gravities of two substances containing an equal number of gram-equivalents per
litre was proportional to the number of equivalents. This extension of C. A.
Valson's moduli is made possible on account of the empirical fact that the dilution
of equivalent salt solutions is attended by almost the same contraction.
In 1878, W. Ostwald^ showed that the changes in volume observed by G. T.
Gerlach (1859), C. Tissier (1859), and J. Regnauld (1865), to accompany the
neutralization of solutions of ammonium and alkali hydroxides by various acids,
in solutions containing a gram-equivalent of base or acid per kilogram of solution,
are dependent on both the acid and the base, and that the difference in the changes
in volume which accompany the neutrahzation of different bases by one acid is the
same whatever acid is used ; and conversely, the difference in the volume changes
which attend the neutralization of the different acids by one base is independent
of the nature of the base. W. Ostwald showed that similar additive relations hold
good for many other properties of dilute salt solutions, and used the facts as an
SOLUTIONS 581
argument in support of the ionic hypothesis. These relations do not hold so well
for concentrated solutions. It is argued that a negative molecular volume for the
solute is impossible, and accordingly a portion of the solvent must be denser than
the pure solvent. Consequently, the increase in density which occurs when salts
are dissolved in water is attributed to the union of a portion of the solvent water
with the solute to form a salt which is hydrated in aqueous solution, and this is
supposed to be confirmed by the increased density of the water of crystallization
in soHd crystalline hydrates. The alternative hypothesis is that the increase in
the density of solutions is due to the depolymerization of the solvent water. The
attempts made to determine the formulae of the supposed hydrates from irregularities
or discontinuities in the density-composition curves have not been generally success-
ful, for the alleged breaks in the curve have been shown in many cases to be due to
experimental errors,^^ and that no reliable conclusions about the formation of
hydrates can be drawn from the density-composition curves.
The thermal expansion and compressibility o! salt solutions. — Some remarks
on these properties have been discussed in dealing with water. Aqueous solutions
of salts do not follow the mixture law.n In general, the thermal expansion of aqueous
solutions of salts is the more uniform the more concentrated the solution ; the
more dilute the solution the greater the curvature of the line showing the relation
between its volume and temperature ; and with the more concentrated solutions,
the more nearly does the volume-temperature curve approach a straight line. This
is attributed to the depolymerization of the so-called ice-molecules when salt is
dissolved in water. Similarly with the compressibility of salt solutions. W. C.
Rontgen and J. Schneider investigated the compressibilities of 1'5 and 0*7 normal
solutions of various salts of the alkalies and ammonium, and the corresponding acids,
and concluded :
The substitution of one constituent of the compound in solution by another, e.g. I by
NO3, Br, CI, OH, SO4, or COg, alters the compressibility of the solution to an extent which
is only slightly dependent on the nature of the other constituent of the compound (H, NH4,
Li, K, Na). It appears, then, as if each constituent of a salt exerted a specific effect on the
compressibility of the solution of that compound, which effect is only slightly modified by
replacing the other constituents by different substances ; or, in other words, it seems as
if the components of the dissolved body, and not the compound in which these components
are contained, had the greatest influence on the compressibility of the solution.
Water and ammonia are exceptions ; and each of the exceptional solutions is found
to be a relatively poor conductor of electricity.
The viscosity of solutions. — The viscosity curves of binary solutions are of
three kinds : those which follow the law of mixtures, and those which exhibit
maxima or minima. According to A. E. Dunstan, the minima are produced by
the depolymerization of one or both the associated components of the mixture ;
the maxima are produced by the formation of complexes between solute and solvent.
T. Graham attributed the maxima he obtained with aqueous solutions of many
common acids and alcohols to the formation of definite compounds, and in this he
is supported by A. E. Dunstan, D. E. Tsakalotos, 0. Faust, R. B. Denison, and
others. 12 As in the case of the specific gravities, the existence of numerous hydrates
has been deduced from points of discontinuity in the viscosity curves. Four
hydrates of acetone, six of methyl alcohol, and seven of ethyl alcohol have been
reported by E. Varenne and L. Godefroy to be formed in aqueous solutions of these
compounds. E. W. Washburn takes the view that the deviations of a physical
property from the mixture law cannot give conclusive evidence of the existence of
hydrates or other complexes ; and he states that in the case of aqueous solutions
of methyl alcohol, " the points of discontinuity in the viscosity curves are purely
imaginary, and due to experimental errors." Probably the larger part of the
abnormal effects of salts on the physical properties of water should be ascribed to
changes in the complexity of the solvent.
The specific heat of solutions. — The molecular specific heat of mixtures of
582 INORGANIC AND THEORETICAL CHExMISTRY
some organic compounds — e.g. carbon disulphide and chloroform — follow the
mixture law, but mixtures of alcohol and water, and indeed aqueous solutions
generally, do not follow this rule. This was noticed by A. A. B. Bussy and J. L. H.
Buignet i^ in 1865. The thermal capacity of mixtures of alcohol and water is
always greater than that calculated by the law of mixtures, but in the case of
aqueous solutions the thermal capacity is usually less than that calculated by the
law of mixtures. The total heat capacity of aqueous solution of salts is frequently
less than that of the contained water ; in illustration, J. Thomsen found that the
difference between the thermal capacity of 100 parts of water and the thermal
capacity of a 10 per cent, solution of sodium chloride is 201 cals. ; a 20 per cent,
solution, 0*36 cal. ; and a 30 per cent, solution, —2*66 cals. J. Thomsen has
also shown that the calculated molecular heat capacity of the solute in solutions
of electrolytes is nearly always less than for the solute alone, and it steadily decreases
on dilution, passing through zero, and finally assuming a negative value. A
negative heat capacity has no physical meaning ; consequently, the heat capacity
of the water as a whole, or of a certain portion of the water is lowered by the presence
of a solute. M. Berthollet believed that the formation of hydrates with a smaller
heat capacity than water explains the phenomenon satisfactorily. The depolymeri-
zation of the solvent water is the alternative hypothesis.
The heat of solution. — The heat developed or absorbed during the formation
of a solution may be expressed in different ways. The magnitude measured in
the calorimeter usually refers to the thermal change which occurs when a gram of
the substance is dissolved in so large a quantity of the solvent, that any further
dilution of the resulting solution is not attended by any thermal change ; or when
a gram of the substance is dissolved in w grams of water ; or when a gram of the
substance is dissolved in sufficient water to form a saturated solution ; or when a
small quantity of the substance is added to a saturated solution containing a gram
of the substance — this is, the reversible heat of saturated solution ; the heat evolved
or absorbed when a small quantity of water is added to a saturated solution con-
taining a gram of the solute ; etc. These magnitudes can all be represented in terms
of the gram-molecule instead of the gram.
In the solution of a solid, work must be performed in the separation of the
molecules against intermolecular attraction, this is equivalent to the latent heat of
sublimation or to the heat of fusion plus the heat of vaporization. This exerts a
cooling effect. In the solution of a liquid, work equivalent to the heat of vaporiza-
tion must be expended against intermolecular attraction. Thermal phenomena
of greater or less magnitude may also accompany a reaction between solvent and
solute, the formation of complexes, depolymerization of the solvent, etc. The
observed heat of solutions is a resultant of these several effects. The heat of solution
of gases includes the external work — fv or RT — performed in compressing the
gas, and this magnitude must be subtracted from the observed heat of solution. The
external work associated with the solution of solids and liquids is negligibly small.
All known gases have a positive heat of solution ; and this is usually the case with
liquids ; solids also may have a positive heat of solution, but more usually the heat
of solution of solids is negative, for they dissolve with an absorption of heat. If
the heat of vaporization of a liquid exceeds the heat of solution, it will dissolve
with an evolution of heat, and with an absorption of heat if the heat of vaporization
is less than the heat of solution. Similarly with solids, the nature of the thermal
change is conditioned by the difference in the heats of solution and sublimation.
Consequently, the heat of solution of a gas is usually greater than the heat of
vaporization ; and the heat of solution of a solid is usually less than the heat of
sublimation. As a rule, the heat of solution of a substance is smaller the less its
solubihty.14
Does the evolution of heat which occurs when a solid is dissolved in water, or
when a concentrated solution of the salt is diluted indicate the formation of hydrates ?
J. Thomsen i^ returns a negative answer. The progressive addition of water to
SOLUTIONS
583
sulphuric anhydride, SO3, gave him the curve indicated in Fig. 28, and it is asked :
At what point in such a curve should we be justified in setting up a distinction
between the effect due to chemical combination and that due to other causes ?
The volume of the solution obtained when anhydrous salts are dissolved in water
is always less than the sum of the volume of solvent and solute, and the subsequent
dilution of the solution is likewise followed by a contraction. P. A. Favre and
C. A. Valsoni^ calculated the amount of heat required
for the contraction which occurs when sulphuric acid,
H2SO4, is diluted with a gram-molecule of water
from the specific heat and coeflS.cient of thermal
expansion, and found it to be 179 calories less
than that actually observed. Hence, the hypo-
thesis that the observed change in volume is merely
due to such a change in the mean distances and
motions of the molecules as would be produced by
a change of temperature is not tenable. The effect
must be complicated either by the formation of
hydrates, or to a polymerization or depolymeriza-
tion of the solvent. Analogous results are ob-
tained with sodium chloride, only that the heats
of solution and dilution are negative.
There is a parallelism between the heat of
solution and the degree of hydration when the process of solution is attended by
the formation of one or more hydrates. This is shown when the heats of hydration
of salts with a common ion are compared with abnormal freezing points or vapour
pressures. The magnitude of the depression corresponds with the order of the
hydrates of these salts, which is that indicated in Table IX.
-^
—
• —
/
k '
i
.
/
/
'
1
Uo/ecu/es o
f Water
__
0 2 4 6 8 10 12 14 16 Ift 20
Fig. 28.— Heat of Solution and
Dilution of Aqueous Solutions
of Sulphuric Anhydride.
Table IX.—
Beats of
Solution (Calories).
Nitrate
Thiocyanate
Cyanide
Chloride
Bromide
Iodide
NO 3
SCy
Cy
CI
Br
I
Potassiiun
-8-64
-6 00
-312
- 312
- 504
-5-04
Ammonium .
-6-24
- 4-08
- 0-24
-3-60
Sodium .
-5 04
. ,■
- 1-20
- 0-24
+ 1-20
Lithium
+0-24
. —
—
+ 8-40
—
—
Barium .
.
. .
+ 2-16
+ 5-04
Strontium
—
—
—
-hll-04
+ 16-08
—
Calcium
—
—
—
+ 17-52
+ 24-48
—
Magnesium
—
—
— ■
+ 24-00
—
J. Thomsen found that of thirty-five salts he examined :
The chlorides of sodium^ ammonium, calcium, magnesium, zinc, jiickel, and copper ;
potassium bromide, potassium, cyanide ; the nitrates of sodium, ammonium,, strontium, lead.,
magnesium, manganese, zinc, and copper ; the acetates of potassium, sodium, ammonium,
and zinc ; the sulphates of ammonium, magnesium, manganese, zinc, and copper ; sodium
hydrogen sulphate ; ammonium tartrate, and bicarbonate,
eighteen salts which evolve heat when the anhydrous salt dissolves in water also evolve
more heat on dilution ; and eleven salts (italicized in the list) which absorb heat on
solution also absorb still more heat on dilution. All those salts which form definite
crystallizable hydrates evolve heat, the other salts do not. The six exceptions
included :
Ammonium and potassium bisulphates, sodium sulphate, sodium iodide, and the
carbonates of potassium and sodium.
584 INORGANIC AND THEORETICAL CHEMISTRY
J. Thomsen's opinion is that
There is no doubt that the salts which dissolve in water with the evolution of much
heat, and form crystallizable hydrates, are present also in solution as hydrated compounds ;
but a determination of the number of water molecules contained in such compounds must
be very difficult.
In 1858, G. R. Kirchliofi ^^ obtained thermodynamically an expression between
the vapour pressure and the heat of solution of a solid or the thermal change which
occurs when a gram of the substance is dissolved in sufficient water to form a satu-
rated solution. Gr. R. Kirchhofi also deduced an equation for the heat of dilution
of a saturated solution, and his formulae have been verified by F. Jiittner
(1901) and R. Scholz (1892), and improved by N. N. Schiller. Assuming that the
variation in the heat of dilution dl of a solution with change of temperature dd
is equal to the rate at which the thermal capacity dC of the solution changes with
concentration dm — that is, dl/dO—dC/dm — provided the thermal capacity of the
solution does not change with respect to temperature, H. Teudt has shown that the
change in the heat capacity of a salt solution with temperature is in general less
than that of water ; and F. R. Pratt, that the ratio dCldm or dljdd decreases consider-
ably with an increase of temperature.
It has been shown i^ that the solubility S — gram-molecules per litre — at the
absolute temperature T, is related with the heat of solution Q by the expression :
dlogS_ Q
dT ~ J?.T2' ^"^
when Si denotes the solubility of a compound in water at the absolute temperature
Ti, and S2 the solubility at a temperature T2 ; and Q denotes the heat of solution
on the assumption that the heat of solution Q does not vary with temperature.
If the heat of solution does vary with temperature — say Q=a-{-hT-\-cT^-\- . . . —
changes corresponding with those previously indicated must be made. The heat of
solution here refers to the solution of a gram-molecule in its own saturated solution.
This fictitious quantity may differ considerably from the observed heat of solution
in a large quantity of water, and it may even be of opposite sign, as L. T. Reicher
and C. M. van Deventer observed with copper chloride, CUCI22H2O, which evolves
heat when dissolved in a large quantity of water, but absorbs heat if " dissolved"
in its own saturated solution. With sparingly soluble substances, the difference
between the two heats of solution is negligibly small.
Example.- — Boric acid has a solubility 1-95 at 0°, and 2-92 at 12°. Accordingly, by
substituting 2^1 = 273, ^j = l'95; and 5 = 2-92 and T = 285 in the preceding expression,
and using natural logarithms, Q = 5'2, the observed value is 5'6.
For solutions of substances which dissociate so as to increase the number of
molecules in the solution to i per molecule of solute during solution, J. H. van't
Hoff introduces the factor i ; and J. J. van Laar the factor a, where a denotes
the degree of dissociation such that l-fa=^. The respective equations
are :
d log S_ Q d log S ^ 2— g
dT ~~ 2if^'' dT~~~2T^' 2
Equations (1) and (2) show that the change of the solubility of a compound
with temperature is of opposite sign to the heat of solution — if the solubility increases
with a rise of temperature, heat will be absorbed when the substance dissolves in
its own saturated solution — this phenomenon occurs with most substances :
cupric chloride, CUCI2.2H2O, etc. ; if the solubihty decreases with a rise of tempera-
ture, heat will be evolved — examples, gases, ether, carbon disulphide, bromine,
etc. ; and if the solubility is a maximum or a minimum, the solubility does not
change appreciably with a small variation of temperature, and the heat of solution
SOLUTIONS 585
will be zero — examples, the solubilities of isobutyl alcohol (W. Alexejeff),^^ calcium
sulphate (H. le Chatelier), and of sodium chloride (C. M. van Deventer and H.^ J.
van de Stadt) in water ; the heat of solution of two liquids at their critical solution
temperature is also zero.
The effect of chemical composition on solubility. — It has been empirically
observed 21 that while there is a marked tendency for unlike substances to react
chemically, there is a strong disposition for like substances to dissolve in like,
for ivhen there is a close co7inection in chemical constitution between a liquid and a
solid, the solid will usually dissolve readily in the liquid. In illustration, nearly
all salts which contain water of crystallization are soluble in water — calcium
sulphate is one of the least soluble, while magnesium phosphate and arsenates,
and some natural siHcates, are exceptional in being insoluble. Insoluble salts are
almost always anhydrous and rarely contain the elements of water. The solubility
and capacity for uniting with water of crystallization of a series of salts containing
nearly allied metals generally diminishes as the atomic weight increases- — e.g. the
sulphates of magnesium, calcium, strontium, and barium ; the chlorides or nitrates
of calcium, strontium, barium, and lead. These facts have been taken to imply
that the salts which readily crystallize with water of crystallization also dissolve
readily in water, because they are likewise readily hydrated in the solvent. Anhy-
drous copper sulphate is white, and its hydrated crystals as well as its aqueous
solutions are blue. It has also been demonstrated by P. Vaillant and by G. N.
Lewis 22 that the colour changes produced in aqueous solutions of copper and cobalt
salts are due to hydration. Hence, there is here direct evidence of the formation
of hydrates in solution. Chemical composition, however, is not a sufficient
criterion to determine whether a solid will be soluble or insoluble in a given
menstruum.
W. Herz 23 has tried to show that the solubility of a number of organic liquids
in water is greater the smaller the diameter of the molecule ; but the relationship
between molecular diameter and solubility is still largely conditioned by the specific
chemical properties of the substance concerned.
The relation between the solubility and the melting point of a soUd.—
A. L.Lavoisier (1793)24 expressed the opinion that the solubility of a solid must be
related with its fusibility. He stated that the solubility of a salt in cold and hot
water is greater the more readily it fuses. T. Carnelley, W. A. Tilden and W. A.
Shenstone have shown that there appears to be some connection between the two
constants in that the more fusible a substance, the more readily does it dissolve in
a given liquid. In illustration, the solubilities and fusibilities of the alkali chlorides
may be cited :
LiCl
NaCl
KCl
RbCl
CsCI
Melting point
602°
8or
790°
726°
646°
Solubility (15°)
80
36
33-4
80
179
Fusibility, however, is not sufficient in itself to determine whether a solid shall be
soluble or insoluble in a given menstruum. Silver chloride, AgCl, for instance, is
more fusible than any of the alkali chlorides — its melting point is 490° — but it is
very sparingly soluble — almost insoluble.
According to I. Schroder, if s denotes the molecular fraction of the dissolved
molecules to the total number in a given solution, s will be proportional to the
osmotic pressure, and J. H. van't Hoff's well-known equation, dp/p—Qdl/RT^,
becomes ds/s=QdTIRT^, as indicated above. Consequently, the integral log s
=QIRT plus the constant of integration. This result was obtained independently
by H. le Chatelier in 1894. To evaluate the integration constant, it will be observed
that when the absolute temperature T rises, the concentration of the solution
increases until at the melting point T^, 5'=unity, and Q is then identical with A
the latent heat of fusion of the solvent. Since log 1=0, the integration
constant =—QIRT; and therefore the relation between the melting point
586 INORGANIC AND THEORETICAL CHEMISTRY
Tm, and the solubility of the solid at a temperature T, can be represented by
the expression :
1 ^(Q ^\ 1 MX Tm-T
if T^n be the melting point of the solution and T that of the pure solvent when
MA is the molecular latent heat of fusion of the solvent (a negative quantity).
I. Schroder found the approximation X=Q to be applicable for a number of organic
compounds.
The above expression contains no term relating to the dissolved substance,
so that when the solubiUty is expressed by the number of molecules of the solvent
present in one gram-molecule of the solution, it is found to be independent of the
nature of the dissolved substance, and one and the same solubility curve records
the behaviour of a given solvent towards all the compounds which it may be capable
of dissolving. The term solvent here refers to the substance which crystallizes
first from the liquid mixture on cooling, or the substance whose melting point is
being lowered — e.g. if salt crystallizes out on cooling an aqueous solution, then the
salt is here regarded as the solvent. H. Crompton adds that since MX—V?tSTEv,
where 2Jv denotes the sum of the valency bonds, it follows by substitution for
MX, and solving for T^, that Tm=—(^'lTi:vl(\ogs—0-lEv), from which it
follows that, knowing nothing more than the chemical constitution and melting
point of a given compound, it should be possible to deduce its solubility curve,
and its general behaviour as a solvent.
If a mixture of two substances be in question, the preceding expression will
apply to the one, and T^'=— O'TT'iJv/llog {\—s)—0'lZv'] will apply to the other.
At the eutectic temperature Ttn=Tm, and therefore
TUv log s—0'7l!v
rZv' log (l-s)-0'lEv'
which shows that if T be greater than T', s will be less than 1—5, or, in a eutectic
mixture, the substance with the lower melting point will be present in the greater
proportion. A. Miolati was the first to show that the eutectic 'point always lies
nearest to the melting point of the lower melting constituent of the mixture ; and the
eutectic mixture always contains the larger proportion of the lower melting constituent.
With a mixture of potassium and sodium nitrates, for instance, Zv=Zv\ and
since T differs by about 3 per cent, from T\ s should be nearly equal to 1 — s ; or, more
exactly, the eutectic mixture contains 47 gram-molecules of KNO3, 53 of NaN03,
and melts at 217°. H. Crompton and M. A. Whiteley also showed that the above
relations hold for a number of pairs of organic compounds. The observed melting
points are higher than the calculated values in cases where the solvent does not
crystallize out alone but forms a solid solution with the solute.
References.
^ P. Gassend, Opera ^ FlorentiaB, 1. 130, 1727 ; Abbe Nollet, Legons de physique experimentale,
Amsterdam, 1754 ; Mem. Acad. Berlin, 6. 67, 1750 ; R. Watson, Phil. Trans., 59. 325, 354, 1770 ;
J. Dalton, Acids, Bases, and Salts, Manchester, 1840 ; E. Swedenborg, Prodromus principorum
rerum naturalium, Amsterdam, 1721 ; S. Holker, Phil. Mag., (3), 27. 207, 1845 ; L. Playfair and
J. P. Joule, ib., (3), 27. 453, 1845 ; J. C. G. de Marignac, ib., (3), 28. 527, 1846 ; J. A. Wanklyn,
W. Johnstone, and W. J. Cooper, ib., (5), 32. 473, 1891 ; M. L. Prankenheim, Pogg. Ann., 72. 422,
1847 ; W. W. J. Nicol, Phil. Mag., (5), 16. 121, 1883 ; (5), 18. 179, 1884 ; J. G. MacGregor, Trans.
Roy. Soc. Canada, 8. 19, 1890 ; C. M. Pasea, ib. 18. 27, 1900 ; G. A. Carse, Trans. Roy. Soc. Edin.,
25. 281, 1904 ; W. R. Bousfieldand T. M. Lowry, Phil. Trans., 204. A, 265, 1905 ; I. Traube, Zeit.
anorg. Chem., 3. 25, 1893 ; C. Forch, Ann. Physik, (4), 12. 591, 1903 ; R. J. Southworth, Amer.
Journ. Science, (3), 17. 399, 1879 ; P. A. Favre and C. A. Valson, Compt. Rend., 75. 334, 1872 ; C. A.
Valson, ih., 73. 441, 1871 ; E. Ruppin, Zeit. phys. Chem., 14. 482, 1894 ; G. J. W. Bremer, ib.,
3. 423, 1888 ; F. Kohlrausch and W. HaUwachs, Wied. Ann., 53. 14, 1893 ; L. de Boisbaudran,
SOLUTIONS 587
Compt. Rend., 120. 540, 1896 ; 121. 100, 1895 ; J. A. Wanklyn, W. Johnstone, and W. J. Cooper,
Phil. Mag., (5), 32. 473, 1891.
2 A. Michel and L. Krafft, Ann. Chim. Phys., (3), 41. 471, 1854 ; P. Kremers, Pogg. Ann., 92.
297, 1864 ; 94. 87, 255, 1855 ; 95. 110, 1855 ; 96. 39, 1855.
« J. S. Lumsden, Journ. Chem. Soc, 91. 24, 1907.
* P. A. Favre and C. A. Valson, Compt. Bend., 77. 802, 1873 ; I. Traube, Zeit. anorg. Chem.,
3. 1, 1892 ; J. Y. Buchanan, Amer. Journ. Science, (4), 21. 25, 1906 ; G. P.- Baxter and C. C.
Wallace, Journ. Amer. Chem. Soc, 38. 70, 1916 ; G. P. Baxter, ib., 33. 922, 1911.
» D. I. Mendeleeff, Ber., 19. 370, 400, 1886,
* G. Tammann, Ueher die Beziehung zwischen den inneren Krdften und Eigenschaften der
Losungen, Leipzig, 78, 1907.
7 C. A. Valson, Compt. Bend., 73. 441, 1874.
8 C. Bender, Wied. Ann., 20. 560, 1883 ; J. A. Groshans, ih. 20. 492, 1883.
^ G. T. Gerlach, Specifische Gewicht der gebrauchlichsten Salzlosungen bei verschiedenen Con-
centrationsgraden, Freiberg, 1859; J. Regnauld, Journ. Pharm. Chim., (4), 1. 401, 1865;
C. Tissier, Mem. Institut., 158, 1859 ; W. Ostwald, Journ. prakt. Chem., (2), 18. 353, 1878; E. A.
Schneider, Monatsh., 11. 166, 1890 ; I. Traube, Zeit. anorg. Chem., 3. 24, 1893 ; W. R. Bousfield
and T. M. Lowry, Phil. Trans., 204. A, 265, 1905 ; W. W. J. Nicol, PhU. Mag., (5), 16. 121, 1883 ;
(5), 18. 179, 1884.
" V. H. Veley and J. J. Manley, Proc. Boy. Soc, 69. 96, 1901 ; J. Domke and W. Bein, Zeit.
anorg. Chem., 43. 125, 1905 ; F. W. Kuster and R. Kremann, ib., 41. 33, 1904 ; R. Kremann and
R. Ehriich, Sitzber. Akad. Wien, 116. 733, 1907.
^^ G. T. Gerlach, Specifische Gewichte der gebrauchlichsten Salzlosungen bei verschiedenen Con-
centrationsgraden, Freiberg, 1859 ; P. Kremers, Pogg. Ann., 100. 394, 1857 ; 105, 306, 1858 ; 108.
115, 1859 ; 111. 60, 1860 ; 114. 41, 1861 ; 120. 493, 1863 ; J. C. G. de Marignac, Ann. Chim. Phys.,
(4), 22. 385, 1871 ; P. de Heen, Becherches touchant la physique comparee et la theorie des liquides,
Paris, 76, 1883 ; G. J. W. Bremers, Zeit. phys. Chem., 3. 423, 1888 ; W. C. Rontgen and J. Schneider,
Wied. Ann., 29. 105, 1886 ; M. Schumann, ib„ 31. 14, 1887.
12 T. Graham. Phil. Trans., 157. 373, 1861 ; A. E. Dunstan, Journ. Chem. Soc, 85. 817, 1904 ;
87. 11, 1905 ; A. E. Dunstan and R. W. Wilson, ib., 91. 83, 1907 ; A. E. Dunstan and J. A.
Stubbs, ib., 93. 1919, 1908 ; F. B. Thole, A. E. Dunstan, and A. G. Mussel, ib., 103. 1114, 1913 ;
A. E. Dunstan and F. B. Thole, ib., 95. 1556, 1909 ; The Viscosity of Liquids, London, 1914;
D. E. Tsakalotos, Bull. Soc Chim., (4), 3. 234, 1908 ; 0. Faust, Zeit. phys. Chem., 79. 97, 1912 ;
58. 436, 1907 ; R. B. Denison, Trans. Faraday Soc, 8. 20, 1912 ; F. H. Getman, Journ. Chim.
Phys., 4. 386, 1906 ; R. Kremann and R. Ehriich, Sitzber. Akad. Wien, 116. 733, 1907 ;
E. Varenne and L. Godefroy, Compt. Bend., 138. 990, 1904 ; 137. 993, 1903 ; E. W. Washburn,
Tech. Quart., 21. 399, 1908.
13 A. A. B. Bussy and J. L. H. Buignet, Ann. Chim. Phys., (4), 4. 5, 1865 ; J. C. G. de Marignac,
ib., (4), 22. 385, 1871 ; J. H. Schuller, Pogg. Ann. Erg., 5. 116, 192, 1871 ; A. Winkelmann, Pogg.
Ann., 150. 592, 1873 ; J. Thomsen, ib., 142. 337, 1871 ; Thermochemische Untersuchungen,
Leipzig, 1. 53, 1882 ; W. F. Magie, Phys. Rev., 25. 171, 1907 ; M. Berthelot, Essai de mecanique
chimique fondee sur la thermochimie, Paris, 1. 508, 1879 ; 2. 80, 176, 1879.
1* J. Thomsen, Journ. prakt. Chem., (2), 13. 241, 1876.
15 J. Thomsen, Thermochemische XJntersuchungen, Leipzig, 3. 20, 1884 ; W. A. Tilden, B. A.
Rep., 444, 1886.
i« P. A. Favre and C. A. Valson, Compt. Bend., 11. 577, 802, 907, 1873.
17 G. Tammann, Mem. Acad. St. Petersburg, 35. 9, 1887 ; W. E. Biltz, Ber., 37. 3036, 1904 ;
W. R. Bousfield and T. M. Lowry, Trans. Faraday Soc, 3. 123, 1907.
18 G. R. Kirchhoff, Pogg. Ann., 103. 177, 1858 ; 104. 612, 1858 ; Gesammelte Abhandlungen,
Leipzig, 454, 1882 ; F. Juttner, Zeit. phys. Chem., 38. 76, 1901 ; R. Scholz, Wied. Ann., 45. 193,
1892 ; N. N. Schiller, ib., 67. 292, 1899 ; W. F. Magie, Phys. Bev., (1), 35. 265, 272, 1912 ; (2), 10.
64, 1917 ; F. R. Pratt, Journ. Franklin Inst., 185. 663, 1918 ; H. Teudt, Ueber die Aenderung
der specifischen Wdrmen wdsseriger Salzlosungen mit der Temperatur, Berlin, 1900 ; L. Natanson,
Zeit. phtjs. Chem., 10. 748, 1892 ; P. Duhem, ib., 2. 568, 1888 ; C. Dieterici, Wied. Ann., 42. 613,
1891 ; 45. 207, 1892 ; R. von Hehnholtz, ib., 27. 542, 1886.
19 C. M. van Deventer and H. J. van de Stadt, Zeit. phys. Chem., 9. 45, 1892 ; L. T. Reicher
and C. M. van Deventer, ib., 5. 559, 1890 ; C. M. van Deventer, i6., 2. 92, 1888 ; V. Rothmund,
ib., 26. 433, 1898 ; J. J. van Laar, ib., 17. 545, 1895 ; A. A. Noyes and G. V. Sammet, ib., 43. 513,
1903 ; G. von Marseveen, ib., 25. 91, 1898 ; H. W. B. Roozeboom, Bee Trav. Chim. Pays-Bas.,
5, 343, 1886 ; P. Duhem, Traitd elementaire de mecanique chimique, Paris, 1. 181, 1897 ; H. le
ChateHer, Compt. Bend., 85. 440, 1877 ; 100. 50, 1885 ; Becherches experimentales et thdoriques
sur les equilibres chimiques, Paris, 138, 1888 ; J. H. van't Hofif, Studies in Chemical Dynamics,
London, 207, 1896.
20 W. Alexejeff, Compt. Bend., 100. 442, 1885 ; H. le ChateHer, ib., 85 440, 1877 ; 100. 50,
1885 ; C. M. Deventer and H. J. van der Stadt, Zeit. phys. Chem., 9. 43, 1892 ; V. Rothmund,
ib., 26. 433, 1898.
21 W. A. TUden, B. A. Bep., 444, 1886.
22 P. Vaillant, Ann. Chim. Phys., (7), 28. 257, 1903 ; G. N. Lewis, Zeit. phys. Chem., 52. 224,
1905 ; G. Rudorf, Jahrb. Bad. Elekt., 3. 422, 1907 ; 4. 380, 1908.
588 INOKGAlSnC AND THEORETICAL CHEMISTRY
2» W. Herz, Zeit. Elektrochem., 23. 23, 1917 ; 21. 373, 1915.
** A. L. Lavoisier, TraiU eUmentaire de chimie, Paris, 2. 104, 1793 ; T. Carnelley, Phil. Mag.y
(5), 13. 180, 1882 ; W. A. TUden and W. A. Shenstone, Phil. Trans., 175. 28, 1884 ; W. A. Tilden,
Journ. Chem. Soc, 45. 266, 1884 ; J. Walker, Zeit. phys. Chem., 5. 193, 1890 ; I. Schroder, ib.,
11. 449, 1893 ; A. Miolati, ib., 9. 649, 1892 ; H. le Chatelier, Compt. Bend., 118. 638, 1894 ;
H. Cromjpton, Journ. Chem. Soc, 67. 316, 1896; H. Crompton and M. A. Whiteley, ib., 67.
327, 1895 ; H. le Chatelier, Compt, Bend., 118. 638, 1894 ; E. W. Washburn, Journ. Amer. Chem.
Soc., 32. 653, 1910.
CHAPTER XI
CRYSTALS AND CRYSTALLIZATION
120
A
100
1. The Crystallization o! Salts from Solutions
The world is not a meaningless medley. We do not believe that blind chance reigm
supreme. On the contrary, we see order everywhere, and law is the regulating principle
in all things and processes. — P. Carus.
If a saturated solution of a salt be allowed to evaporate at a given temperaturej
crystals of the salt separate when the concentration of the solution becomes greatei
than that represented by a point on the solubility curve. The phenomenon becomes
a little more complex when the solution contains two or more salts which do not
act upon one another ; and more complex still if the salts react with one anothei
forming double salts or with the solvent forming hydrates.
The solubility of a mixture of sodium and potassium chlorides in water at 25^
is represented by the curves shown in Fig. 1. These salts form neither hydrates
nor double salts at this temperature. The ordinates represent quantities of sodium
chloride, NaCl ; the abscissae, quantities of potas-
sium chloride, KCl. The concentration of a
saturated solution of sodium chloride at 25° is
represented by a point A, and of a saturated solu-
tion of potassium chloride by a point B. The line
AG represents the composition of solutions of
sodium chloride saturated in presence of the propor-
tions of potassium chloride indicated by the abscissa)
of the curve AC ; and the line GB, the composition
of solutions of potassium chloride saturated in
presence of the proportions of sodium chloride
represented by the ordinates of GB. The point G
represents the composition of a solution saturated
with both salts. The composition of all possible
solutions of these two salts can be represented by , . ^ ,
a point inside the surface AQBO, points outside this ^ixed Solutions of Sodium^and
area can only represent super-saturated solutions not Potassium Chlorides,
in a state of equilibrium. Hence, when a solution con-
taining equal molecular proportions of both salts — say 50 gram-molecules per 1000
molecules of water — is evaporated, the relative proportions of the two salts will not
alter ; water alone is removed and the solution becomes more and more concentrated,
so that the abscissa and ordinate, representing the composition of the solution, change
from those of the point P to those of the point Q— when P is left of Q. At Q the
solution will be saturated with respect to the less soluble potassium chloride, and
this salt will accordingly crystallize from the solution ; as evaporation proceeds,
potassium chloride continues to separate ; the successive states of the solution are
represented by points passing from Q in the direction BQ.
The phenomenon is really wonderful. The molecules of both sodium and
potassium chlorides are uniformly diffused throughout the original solution ; but,
as soon as the evaporating liquid has attained a certain concentration, the mole-
cules of the potassium chloride alone commence crystal-building ; and ordered
689
80
60
40
20
^
_
Te/flperature
^
^
■sKClhda^eparale
\
■\
:«
\
5
K
20 40 60 80 B
Gram Mo/ecu/es of/fQ.
100
590 INORGANIC AND THEORETICAL CHEMISTRY
cosmos grows out of a chaotic mixture of molecules ; the molecules of the solute
appear to be dominated by some occult power, for they withdraw from the solution
in harmonious order, which is followed as rigorously as the bricklayer, when building
a mansion, places brick upon brick, according to the plan predetermined by the
architect. The operation continues until the solution has the composition repre-
sented by the point of intersection, C, of the lines of A and B. At C the solution
is saturated with respect to both salts. Any further concentration of the solution
will result in the deposition of sodium and potassium chlorides side by side, and
at rates which are proportional to the concentration of the solution. A solution
which has the composition represented by the point C continues to deposit a mixture
of crystals of a constant composition until it has been evaporated to dryness. J. H.
van't Hoff (1905) calls the point C the end-point of crystallization, and he draws
attention to the fact that when a solution of the two salts is depositing crystals of
one of them, the composition of the solution changes further and further away from
the composition of a saturated solution of that salt until the end-point of crystalliza-
tion is reached. Similar remarks would have been applicable for a solution with,
say, 90 gram-molecules of sodium chloride and 20 of potassium chloride. Sodium
chloride would separate along the curve AC until the end-point C was attained.
Here the molecules of both salts are simultaneously building crystals side by side.
§ 2. Fractional Crystallization
What chemist who has watched iinder the microscope the beautiful symmetrical manner
in which minute particles of a substance separating in solid form from solution, arrange
themselves in geometrical figures obeying well-established mathematical laws, can pretend
to explain the cause of the astounding behaviour of inert lifeless matter ? — H. C. Bolton.
The molecules of a substance in solution appear to be distinct individuals
before crystallization, while in a crystallizing solution each molecule appears
to exert some specific attraction on its fellow molecules to enable them to
separate from the solution in a definite orderly way so as to form crystals whose
architectural symmetry has been called " a miracle of beauty and delight." The
alignment of the molecules in a growing crystal can proceed so rapidly that a few
seconds of our time must appear a long era in the molecular world. Thus, if
concentrated solutions of aluminium and potassium sulphates be mixed, and
constantly stirred, a mass of transparent sparkling crystals of alum is immediately
precipitated. The molecule of alum is represented in its simplest form by
KA1(S04)2-12H20 ; and accordingly, in these few seconds, the atoms have had
ample time to arrange themselves in molecular groups each containing at least
48 atoms ; and the molecules, in turn, have had time to align themselves in a precise
methodical way to form an indefinitely large number of regular octahedral crystals.
Each tiny crystal contains more molecules than could be enumerated by continuous
counting for myriads of years.
G. la Valle noted in 1853 that under suitable conditions crystals grow in directions
in which growth is opposed by an external force ; this was denied by H. Kopp,
but the fact has been abundantly confirmed by 0. Lehmann and others. Still
further, the force — crystallizing force — exerted by a growing crystal as it builds up
its structure, molecule by molecule, must be comparatively great. Sodium sulphate
or thiosulphate crystallizing in the pores of earthenware will shatter the body into
small fragments.! G. F. Becker and A. L. Day (1905) 2 placed a plate of glass,
supporting a kilogram weight over a growing crystal of alum about one centimetre
in diameter, and found that it raised the weight several tenths of a millimetre.
The disintegration of rocks, etc., by the growth of ice crystals ; the bursting of a
test-tube when plaster of Paris is allowed to set therein ; and the disintegration of
porous bricks and tiles by crystallizing sodium sulphate, are illustrations of a definite
and powerful crystallizing force. F. E. Wright and J. C. Hostetter found that
CRYSTALS AND CRYSTALLIZATION
591
100'
80"
60'
40
20'
when crystals are grown under pressure, the results agree with an hypothesis made
by J. Thomson, namely, that during crystallization each particle — atom or
radicle — enters into the crystalline state in the condition of the crystal at the point
to which it becomes affixed ; and that if the crystal be under a state of strain, the
freshly deposited particle enters into the same state of strain.
The separation of a mixture of potassium chloride, chlorate, and perchlorate. —
When a solution of two (or more) salts is slowly evaporated, if the solubilities
of the salts differ appreciably, one salt may pass more or less completely out of
solution before the other commences to separate ; provided, at the temperature
of separation, the solubilities are independent of one another so that the salts
exhibit no tendency to unite chemically or physically.
Warm 50 grams of potassiimi chlorate, just above its melting point, in a new porcelain
dish, and keep the mass at that temperature until it becomes viscid and almost solid. This
will occupy from ten to fifteen minutes. Let the mass cool. It contains undecomposed
potassium chlorate, some potassiimi chloride, and potassium perchlorate. Add 50 c.c.
of hot water, say at 50°, and when all has disintegrated and the solution cooled, the crop
of crystals of potassium perchlorate can be filtered off. Evaporate the filtrate until a
drop crystallizes when rubbed on a cold surface. The first crop of crystals which separates
as the solution cools is mainly potassium perchlorate, because this salt is so very much
less soluble than the other two ; 100 c.c. of water, at
15°, holds in solution about 36 grams of potassium
chloride, 6*6 gram of the chlorate, and 1"5 gram of
the perchlorate. The solubility curves of these three
salts are shown in Fig. 2. If the evaporation be
carried too far, crystals of potassium chlorate wUl
separate. The first crop of crystals is redissolved
and again allowed to crystallize by cooling the hot
solution ; potassium perchlorate can thus be obtained
almost free from the other two salts ; and by repeated
recrystallization it is possible to isolate the salt in a
high degree of purity. Recrystallization is needed to
get a more pure product, because the crystallizing
salt often carries down with it some of the mother
liquid, or some of the other salts dissolved in the
mother liquid. Indeed, it is perhaps impossible to
prepare crystals quite free from the imprisoned
solvent. If the evaporation be continued after the
separation of the perchlorate, potassium chlorate will
eventually separate, and the product must be re-
crystallized in order to isolate a purer salt. The remaining mother liquid contains
potassium chloride contaminated with the chlorate.
This operation — fractional crystallization — is sometimes a useful method of
separating salts which differ appreciably in solubility. In some cases it is the
only method of separation available, even though the salts in solution do not differ
very much in solubility. The process of fractional crystallization is then very
laborious, involving, maybe, scores of crystallizations and recrystallization s. In
other cases it is impossible to separate the salts in this way, because double salts
separate.
As a rule, the slower the process of crystallization, the larger and more perfect
the crystals. The chemist must learn from the mineralogist many facts concerning
the slow growth of crystals because some natural phenomena cannot be imitated
in the time at man's disposal, for, said J. W. Judd in his work, The Rejuvenescence
of Crystals ^ :
Nature is unstinting in the expenditure of time upon her handiwork, and her slow
elaboration of crystals during millions of years accounts for the presentation of some
natural products of curious phenomena that are not reproducible in test-tubes and crucibles.
Crystals are usually more or less distorted because, owing to local differences in
concentration, the crystallizing solution, in the vicinity of some faces of the crystal,
may be more concentrated than the others. Perfect crystals are rarely found in
nature or in the laboratory.*
4
r
^
^
7
f
7^
k'
f
'J
r
/
^'
7
/
/i
lO 20
50 Grms,
30 40
So/ubi/ity.
Fia. 2.- — Solubility Curves of Potas-
sium Chloride, Chlorate, and
Perchlorate.
592 INORGANIC AND THEORETICAL CHEMISTRY
The purification of salts by recrystallization is a well-known process. The
Latin Geber frequently alludes to the purification of salts by recrystallization, and
the process was recommended by R. J. Haiiy ^ in 1801 for the purification of nitre.
The operation, however, was the subject of a controversy in 1811. Thus, F. Clement
and J. B. Desormes ^ in their De Vepuration des corps far la cristallisation, cited
many experiments which demonstrated the fact, and remarked that chemists who
did not believe in the process should cease to purify their salts in this manner.
The operation of fractional crystallization has been compared with fractional
distillation ; the former is determined by the solubility of a body at a given tem-
perature, the latter by the temperature of vaporization.
Inclusions in crystals. — Ideal crystals are homogeneous, but crystals are some-
times coloured with pigments — e.g. smoky quartz — and they may also have
other inclusions discernible under the microscope. There may be cavities in the
crystal containing gases of various kinds— e.^. air, carbon dioxide, hydrocarbons,
sulphur dioxide, etc. — or the cavities may be wholly or partially filled with liquid
— e.g. water, liquid carbon dioxide, salt solutions, etc. Fig. 3 shows a photograph
of quartz (from Cornish granite) with four cavities containing a liquid, each liquid
inclusion has a bubble of air or gas. If the liquid be condensed carbon dioxide,
the crystal was probably formed under great pressure. Similar cavities are found
in natural crystals of rock salt, calcite, fluorspar, topaz, beryl, barytes, etc. A
cavity formed at an elevated temperature may be
filled with liquid ; as the temperature falls the liquid
contracts faster than the solid, and a space contain-
ing a vapour bubble results. The cavities may also
contain crystals which have separated from the
solution. Then again, crystals of a totally different
substance may be embedded (included) in the larger
crystal — e.g. in the so-called sagenitic quartz {aayrjvr],
a net) — needle-like crystals of rutile cross one another,
giving a reticulated or net-like appearance to the
Fig. 3. — Quartz in Cornish quartz ; and in the variety poetically called veneris
wIX^L.t^d'^JCo'^rand -'«- -/--' hair ov filches f a,nour (love' . darts).
Gas Bubbles ( X 1000). ^""^ quartz encloses bunches of reddish-yellow rutile
needles ; this variety of quartz is probably the
chrysothrix (golden hair) of the Orphic poem. Aventurine quartz has imprisoned
golden or brassy-yellow spangles of, presumably, mica. H. C. Sorby (1858) 7 has
shown that crystals deposited from solutions usually contain cavities enclosing
small quantities of the mother liquid, and this the more the quicker the rate of
crystallization.
Experiments on this subject can be made conveniently by allowing a solution of
potassium chloride to evaporate slowly ; the crystals which form are more or less opaque
towards the centre, and clear and transparent elsewhere. Under a high magnification
(60 to 400 diameters) in a shallow glass cell containing a cold saturated solution of the salt
itself, the opacity appears to be produced by vast numbers of minute cavities arranged in
band's parallel to the sides of the crystals. The cavities are full of liquid. This is demon-
strated by allowing sodium chlofide to crystallize from a solution tinted with potassium
dichromate, the crystals of sodium chloride appear yellow to the naked eye ; and on
magnification, this coloration is evidently an effect due to a large number of cavities in
the colourless crystals filled with yellow liquid.
The decrepitation of common salt on heating is due to the vaporization of the
included water ; similarly, the decrepitation of some varieties of quartz just over
1000° is due to the release of imprisoned gases. So important is this imprisoned
solvent in the preparation of pure material for exact work that, according to T. W.
Richards (1903), many records of painstaking determinations of atomic weights
can be safely ignored because the contamination of the materials from this cause
has been entirely overlooked, or inadequate means have been taken to counteract
the efiects.
CRYSTALS AND CRYSTALLIZATION 593
References.
1 J. Brard, A, deThury, and L. J. yic&t, Ann. Chim. Phys.,(2), 38. 160, 1828; L. M. Luquer,
Trans. Amer. Soc. Civ. Eng., 33. 236, 1895; J. W. Cobb, Journ. Soc. Chem. Ind., 26. 390, 1907.
2 G. F. Becker and A. L. Day, Proc. Washington Acad. Sciences, 7. 283, 1905 ; Journ. GeoL,
24. 313, 1916 ; J. C. Hostetter, Journ. Washington Acad. Sciences, 9, 85, 1919 ; F. E. Wright
and J. C. Hostetter, ih., 7. 405, 1917 ; F. E. Wright, ib., 6. 325, 1916 ; J. Thomson, Phil. Mag.,
(4), 24. 395, 1862 ; Proc. Roy. Soc, 11. 473, 1862 ; W. Bruhns and W. Mecklenberg, Jahresber.
niedersdchs. geol. Ver. Hannover, 6. 92, 1913; S. Taber, Proc. Nat. Acad. Sciences, 3. 297, 1917;
Scient. Amer. SuppL, 83. 410, 1917 ; G. la Valle, Compt. Rend., 36. 493, 1853 ; H. Kopp, Liebig's
Ann., 94. 124, 1855; 0. Lehmann, Molekularphysik, Leipzig, 1. 342, 1888.
3 J. W. Judd, Proc. Roy. Inst, 13. 250, 1891 ; G. Rauber, Die Regeneration der Krystalle,
Leipzig, 1895-6.
* J. C. Hostetter, Journ. Washington Acad. Sciences, 9. 85, 1919 ; J. J. P. Valeton, Ber. sacks.
Ges. Wiss., 67. 1, 1915 ; J. Johnston, Journ. Amer. Chem. Soc, 36. 16, 1914 ; R. C. Moore, ib.,
41. 1060, 1919.
5 R. J. Haiiy, Traite de mineralogie, Paris, 1. 161, 1801 ; H. Kopp, Beitrdge zur Geschichte der
Chemie, Braunschweig, 3. 39, 1875.
6 F. Clement and J. B. Desormes, Ann. Chim. Phys., (1), 92. 248, 1814 ; A. Seguin, ib., (1),
92. 70, 1814 ; L. N. Vauquelin, ib., (1), 13. 86, 1792 ; J. T. Lowitz, ib., (1), 22. 26, 1797 ; A. Arzruni,
Die Beziehungen zwischen Krystallform und chemischer Zusammensetzung, Braunschweig, 1898.
' H. C. Sorby, Quart. Journ. Geol. Soc, 14. 453, 1858 ; J.'G. Konigsberger and W. J. MiiUer,
Centr. Min., 72, 1906.
§ 3o Crystals
In whatever manner, or under whatever circumstances, a crystal may have been formed,
whether in the laboratory of the chemist or in the workshop of nature, in the bodies of
animals or in the tissue of plants, up in the sky or in the depths of the earth, whether so
rapidly that we may literally see its growth, or by the slow aggregation of its molecules
during perhaps hundreds, perhaps thousands of years, we always find that the arrangement
of the faces of the crystal, and therefore its other physical properties, are subject to fixed
and definite laws. — H. P. Gurney.
When homogeneous substances solidify from a state of vapour, fusion, or
solution, their particles often cohere so as to form solid figures — crystals — with
regular symmetrical shapes bounded by plane faces. The solids are then said
to be crystallized. M. A. Capellar's Prodromus crystallographice, published at
Lucerne in 1723, was the first book devoted to crystallography ; and in an essay on
the crystal forms of calcspar, in 1773, T. 0. Bergmann ^ made what may be regarded
as a first approximation to a definition of crystals. He said : " Crystals are bodies
which, though destitute of organic structure, yet externally resemble geometrical
figures more or less regular." The term crystal originally referred to the ice-
like appearance of rock crystal or quartz, but the angular shape of this substance,
as well as that of garnet, beryl, and other minerals, seems to have been regarded
by the ancients as an accidental and not an essential characteristic. The alchemists
must have studied many salts and noticed that on evaporation of their solutions,
definite and regular crystals were obtained which were to some extent character-
istic of particular salts. Thus, at the end of the sixteenth century, A. Libavius,
in his Ars prohandi mineraUa (Francofurti, 1597), stated that the nature of the
saline components of mineral waters could be ascertained by an examination of the
crystalline deposit left on evaporating the water to dryness ; and in his Chemical
Lectures (London, 1712), J. Freind said :
Let these salts be never so divided, and reduced into minute particles, yet when they
are formed into crystals, they each of them reassume their proper shape; so that one
might as easily divest and deprive them of their saltness, as of their figure* This being
an immutable and perpetual Law, by knowing the figure of the crystals, we may under-
stand what the texture of the particles ought to be, which can form those crystals.
The constancy o£ interfacial angles. — In 1669, N. Steno,^ in an essay
De solido intra solidum (Florentise, 1669), showed that in spite of numberless
variations in the size and shape of crystals of different specimens of rock crystal,
he could detect no variation in the angles between the laces. Thus, by cutting
VOL. I. 2 Q
594 INOKGANIC AND THEOEETICAL CHEMISTRY
a series of specimens at right angles to the faces of the prism, he obtained six-
sided sections with sides of varying length, and apparently different figures, but the
angles were all equal, each to each — Fig. 4. Soon afterwards, D. Guglielmini, in
his Riflessioni filosofche dedotte delle figure de' salt (Bononese, 1688), and in his dis-
sertation De salihus (Venetise, 1688), generalized N. Steno's observation, and
asserted that the crystals of every salt have their own peculiar shape which never
changes, and that even in imperfect and broken crystals, the interfacial angles are
always constant. As a result of an examination of over four hundred crystal forms,
J. B. L. Rome de I'lsle confirmed the earlier generalization of D. Guglielmini,
and developed the idea further in his Essai de cristallographie (Paris, 1772) . Every
crystalline substance of definite chemical composition has a specific form
characteristic of that substance. This is sometimes called Haiiy's law,
because R. J. Haiiy (1801) may be said to have " erected the science of
mineralogy on a crystallographic basis which was in turn founded on this principle."
The faces of crystals of the same substance may vary in size and shape ; but if
the crystals possess the same chemical composition, and are at the same temperature,
the interfacial angles have the same numerical value. In other words, the
angles between similar faces of crystals of the same substance are precisely
the same, and are characteristic of that substance. As indicated above,
this generalization was first announced by D. Guglielmini (1688). This means that
the crystalline form of a substance is not determined by the absolute position nor
by the sizes of the faces of the crystal, but rather by the dimensions of the interfacial
angles. The primary dominant faces, so to speak, may persist, but the angles and
edges of some of the crystals may or
may not be truncated and bevelled
{cf. Fig. 4), giving rise to new facets,
or secondary faces. As a general rule,
^''';V^A^*?'''''^^n'^T*P*'^^';^*V^^«"'^'^^. when crystals are formed rapidly the
of the Angles of Quartz Crystals of Different . ^ . , i -i ^ i ^i- "^ r
gj^es. faces are simple, while the faces of
crystals which have grown slowly may
be more complex. The dominant form of the crystals of a given substance *' persists
in spite of these variations ; although the primitive fundamental form can some-
times be recognized, the crystal can be interpreted only after careful study." The
doctrine of the identity of the primitive form of the crystals of a substance was not
at first generally accepted ; and even so late as 1783, G. L. L. de Bufion, in his
Histoire naturelle des mineraux (Paris, 1783-88), combated the idea. He said :
No crystallization will ever afford a specific character, for the variety is infinite ; not
only are there forms of crystallization common to several substances of a different nature,
but, on the contrary, there are few substances of the same nature which do not present
different forms on crystallization. It would thus be more than precarious to establish
differences or resemblances, real and essential, by means of this variable and almost acci-
dental character. . . . Our crystallographers thus propose ... to substitute ideal com-
binations for the real facts of Nature.
This is a remarkable testimony to the difficulties which the contemporaries of
D. Guglielmini and R. J. Haiiy encountered in seeing order among the diverse
forms of the crystals of a substance.
In 1767, C. E. G. H. Westfeld,^ and in 1773, T. 0. Bergmann, stated their opinion
that the different forms of crystals could be regarded as variations of a very small
number of primitive forms, and J. B. L. Rome de I'lsle (1783) followed up the
idea, using the method employed by geometers — e.g. W. Janitzer (1568) and J. Kepler
(1619). J. B. L. Rome de I'lsle derived all the different forms of crystals from six
primitive forms by replacing similar edges and corners by one or more planes.
The idea was still further developed by R. J. Haiiy (1782) in a remarkable Essai
d'une theorie sur la structure des crystaux (Paris, 1784) ; it was extended by F. Mohs
(1820) and others ; * and finally culminated in the seven systems now in general use.
Some emendations must be made to the law of the constancy of crystalline form.
CRYSTALS AND CRYSTALLIZATION 595
Crystal mimicry.— At first sight there appears to be a kind of mimicry among
the crystals of some minerals, for a mineral sometimes has the external crystalline
form characteristic of a totally different mineral species. Such crystals are said to
be pseudomorphs, and they appear to have been formed by secondary chemical
processes whereby the original mineral has been decomposed, and its place taken by
another. As a result of some such process of infiltration, quartz crystals are some-
times found with the external form characteristic of calcite or fluorspar, or barytes ;
tin-stone in the form of felspar; galena (cubic) in the form of pyromorphite
(hexagonal) ; etc. The apparent mimicry is confined to the external form of the
crystals, the internal structure is that peculiar to the normal crystals. A crystalline
substance with its own characteristic outlines is said to be idiomorphic.
The twimiing of crystals.— It appears as if, during the building of some
crystals, the structural units instead of continuing to deposit in layers with units
oriented all in the same direction, suddenly commence to deposit in layers turned
through an angle — sometimes 180° — about an axis perpendicular to those previously
laid down. As a result, two individual crystals appear to be united in a common
plane, or to penetrate one another symmetrically. The phenomenon is called
tioinning ; and the double crystal is called a twin. Twinned crystals of selenite
often have the appearance of an arrow-head. Fig. 5 ; tinstone similarly forms
twinned crystals. Sometimes too, after a number of layers of the crystal units
have been laid in the new direction, there is an abrupt reversion to the original
form. The result is a kind of twinning band which has different texture from the
Mgk
Fig. 5.— Twinned Fig. 6.— Twinned Crystal Fig. 7. — ^Dimorphic Forms of AlkaU Di-
Selenite Crystal. of Pyrite. phosphate.
material on either side. 5 Sometimes, too, a series of parallel bands are formed
in this way. R. J. Haiiy was interested in twinning. He pointed out that the
phenomenon is subject to certain laws, so that " instead of precipitating themselves
tumultuously on one another, the crystals have in a prearranged manner pre-
arranged their disposition." There is a plane of juncture *'so that the two
structures follow their regular development, each in its sphere, towards their
common plane, which forms their respective limits." Fig. 6 is a photograph of a
twinned crystal of pyrite from Minden (Prussia). The twinned crystals may not
only be juxtaposed but they may also be interpenetrant. The interpenetra-
tion may be so complete as to form a single crystal, which may then appear
to have a higher degree of symmetry than it really possesses. Thus a rhombic
crystal may be so twinned that externally it cannot be distinguished from a
tetragonal crystal. Crystals so twinned are pseudo-symmetric, and this form of
mimicry is called mimetic twimiing — e.g. rhombic aragonite may be so twinned as
to form a pseudo-hexagonal prism.
Substances with crystals of more than one form. — Crystals of different
substances usually have different forms ; crystals of the same substance developed
under the same conditions have the same form ; but crystals of the same substance
developed under different conditions may or may not have the same form. For
instance, crystals of sulphur formed above and below 94*5° are different ; there
are two differently shaped crystals of sodium phosphate — Fig. 7. These are cases
of folymor'phism. The hahit of crystals of sodium chloride is octahedral, if growji
596 INORGANIC AND THEORETICAL CHEMISTRY
in alkaline solutions ; and cubical, in neutral solutions ; and conversely, crystals
of alum are usually octahedral, but cubical if grown in alkaline solutions. The two
phenomena, change of habit and polymorphism, are quite difierent. Change of
habit does not mean a change of phase, while polymorphic modifications represent
different phases. There is no change of properties in crystals with a different
habit, but the transformation from one polymorphic form to another usually involves
a discontinuous change of scalar and vectorial properties, and it takes place within
a definite range of temperature, or at a definite temperature, the so-called transition
point.
A. G. Werner mistook aragonite from Spain, described by J. B. L. Rome de
risle,^ for apatite, but M. H. Klaproth showed that the composition is the same
as that of calcspar, and he said that aragonite must be regarded as ordinary calcspar,
which while retaining the same constituents has been altered by a change in the
disposition of its constituent parts. If this be the correct interpretation, it was
recognized that R. J. Haiiy's law required revision. In consequence, a great
many new analyses were made by French and German chemists — A. F. de Fourcroy
and L. N. Vauquelin, L. J. Thenard and J. B. Biot, L. J. Proust, C. F. Bucholz,
J. B. Trommsdorif, etc. A. F. de Fourcroy and L. N. Vauquelin said that their
analyses rendered it necessary for mineralogists to inquire if it is not possible for
the same substance to assume difierent forms ; and L. J. Thenard and J. B. Biot
added that aragonite and calcspar are compounded from the same elements
{principes chimiques) united in the same proportions ; and that the same elements
uniting in the same proportions can form compounds with different physical pro-
perties, because the molecules of the constituent elements can combine in many
different ways. The analyses showed that impurities are present in the native
minerals. In 1801, L. N. Vauquelin and M. A. Klaproth showed that anatase and
rutile consisted of titanic oxide along with some impurities ; while A. Laurent and
P. Dejussieu found that marcasite and pyrite were modifications of iron disulphide —
comme V analogue de Varragonite. A. Stromeyer, in his De arragonite ejusque differentia
a spatho calcareo rhomhoidali chemica (Gottingen, 1813), recorded the presence of
strontium in aragonite, and that this had been overlooked in previous analyses.
This lent support to R. J. Haiiy's contention that small quantities of a foreign
agent with a great power of crystallization may cause a compound to change its
crystalline form. A. Laugier found no difference in strontian-free and strontian-
iferous aragonite. The question was not decided until E. Mitscherlich's work on
polymorphism of sulphur and sodium dihydrogen phosphate, NaH2P04.H20, had
been accepted.
Each kind of crystal is stable only within a limiting range of temperature, so that
the amended form of Haiiy's law is : Every crystalline substance of definite chemical
composition has one specific stable form within certain definite limits of tem-
perature. Substances which crystallize in two different forms are said to be
dimorphous. Fig. 7 ; and substances which crystallize in three different forms
are said to be trimorphous. Titanic oxide, Ti02, for example, is known in three
forms — rutile, anatase, and brookite. Two of these forms will probably be found
to be in a metastable condition at ordinary temperatures. Magnesium meta-
silicate, MgSiOs, exists in four different crystal forms, and it is tetramorphous.
Polymorphism is the general term applied to the phenomenon when a substance
crystallizes in more than one form. When the crystals are similar and yet so
fashioned that one is the mirror image of the other, the crystals are said to be
enantiomorphic — e.g. d- and /- quartz, d- and I- tartaric acid, etc. Enantiomorphic
crystals are not usually regarded as polymorphic forms.
One or more faces of a crystal may be abnormally developed or stunted in
growth. During the growth of a crystal, the concentration of the mother liquid
is rarely so evenly balanced on all sides as to allow the growth to proceed with the
same rapidity in all directions. The crystal will grow fastest where the solution
is most concentrated. If a crystal grows on the bottom of a liquid at rest, flat
CRYSTALS AND CRYSTALLIZATION 597
plates, almost parallel with the bottom of the vessel, may be formed ; while if the
solution be agitated during crystallization, a more uniform growth in all directions
may prevail. This is not all the story, for a crystal may habitually grow more
rapidly in one direction so as to form a prismatic crystal — prismatic habit —
A, Fig. 8. This may take the form of needle-like or acicular crystals; hair-
like or trichitic crystals ; arborescent, branching, or dendritic crystals (ScVSpov,
a tree) ; or fibrous masses ; or the crystals may grow in two directions so as to
form a tabular or plate-like crystal — tabular habit — B, Fig. 8, this may take the
form of groups of thin separate plates, fan-like more or less divergent plates,
feather or branching aggregates (e.g. the six-rayed snow crystals) ; or again the
crystals may grow uniformly in all three directions, and the ideal cube — C, Fig. 8
— will be produced. In the well-known en tremies or hopper-shaped crystals of
bismuth, quartz, or of sodium chloride — 2), Fig. 8 — growth has been fastest along
the edges and corners of the cube, and the crystals assume the form of hollow cubes —
hopper salt. These crystals can be obtained by the addition of a little alum to the
salt. In every case of distorted cubical crystals, the angles between the faces,
however, will remain unchanged — 90°. Similar remarks apply to the different
forms of crystal other than the cubes.
The most suitable condition for uniform growth in all directions occurs when
the growing crystal is suspended in
the middle of the given solution by
means of a thread. The crystal then
sometimes approximates more or
less closely to the ideal form. If ""^ " b ' c d
the crystallization of a solution be Fig. 8. — Ideal and Distorted Cubic Crystals,
rapid, the crystals are usually much
smaller than if the process be slow. 0. Lehmann laid down the empirical
rule :
When a substance crystallizes from a solution owing to the withdrawa of the solvent,
the crystals are the more irregular in shape (i) in proportion as the separation occurs with
greater rapidity ; (ii) in proportion as the solvent is the more viscid ; and (iii) in proportion
as the substance is less soluble.
The minute crystals precipitated by alcohol from solutions of barium chloride,
copper sulphate, lead nitrate, etc., were found by P. Gaubert to be almost perfect
in form.''
As a rule, needle-shaped [acicular), fibrous, hair-like (trichitic), or branching
fern-like, moss-like, or tree-like forms (dendritic) grow in labile solutions
where crystallization is rapid, while well-formed characteristic crystals grow
in metastable solutions where growth is slow and uniform. The dendritic
forms are also common with highly viscous solutions, and with crystals grown in
the presence of colloidal suspensions. Ammonium chloride, calcium sulphate
(Fig. 10), copper sulphate (Fig. 9), or potassium dichromate (Fig. 11) can be used
in illustration. When, say, the last-named salt is deposited slowly from metastable
solutions, well-developed crystals are formed, while if crystallized from labile solu-
tions feather-like aggregates are developed. This is illustrated by Fig. 11, which
shows that the first crystals deposited have arborescent branches, while the later
crystals have the regular crystalline form.
Again, the habit of a crystal may change when grown from liquids containing
other salts in solution. This fact was noted by Robert Boyle 8 in 1666, in the
words : " Notwithstanding the regular and exquisite figures of some salts, they
may by the addition of other bodies, be brought to constitute crystals of very differ-
ing, yet curious shapes." Again, J. B. L. Rome de lisle (1783) ^ showed that the
crystals of sodium chloride which separate from an aqueous solution containing
fresh urine are octahedral, and not cubical. Urea or carbamide can be used pro-
vided, as J. W. Retgers emphasizes, not too small a quantity of urea is present. The
presence of colloids in a solution usually inhibits or modifies crystallization ; thus,
598
INORGANIC AND THEORETICAL CHEMISTRY
sodium chloride crystallizing from a solution containing a trace of gum arabic
forms tree-like or dendritic-masses. Again, atnmonium chloride crystallizing from
cold aqueous solutions forms dendritic masses, but if the saturated solution is
tinted yellow with ferric chloride, and a few crystals of ammonium acetate are
dissolved in the solution, well-formed cubic crystals of ammonium chloride separate
out. Further, if a few grains of octahedral alum crystals be added to a super-
saturated solution of potash alum, octahedral crystals are formed ; whereas, if
potassium carbonate be added to a boiling solution, the liquid, on cooling, deposits
cubic crystals. Z. Weyberg has studied the effect of hydrochloric acid on the
habit of the crystals of alum ; and H. Gerhardt the effect of foreign salts on the
crystals of barytes. Potassium chlorate gives jprismatic crystals when grown in
aqueous solutions ; but if much calcium chloride be present, small needle-like
crystals are obtained, and if some potassium iodate be present long plate-like
crystals separate from the solution — the crystal angles, however, are the same in
each case ; or else the different sets of planes which bound the crystals are all
derived from the same internal structure by the suppression of certain planes, and
the abnormal development of others. Near the beginning of the nineteenth century
R. J. Haiiy i^ showed that the crystals of axinite which are coloured violet by
manganese show additional faces not found on the crystals of axinite coloured
Fig. 9.— Copper
phate grown in
Labile Solution.
Fig. 10.— Calcium
phate grown in
Labile Solution.
Sul- Fig. 11. — Potassium Di-
chromate grown in a
Labile Solution.
green by chlorite, and that the green-tinted crystals are more regularly shaped and
do not present the striated surface observed on the violet-tinted crystals. He said :
The molecules of a substance in a solution and disposed to unite in the formation of a
crystal, are at the same time attracted towards each other and even by the molecules of the
solvent itself; and it is because their mutual affinity exceeds that of the solvent that
crystallization operates. ... If there are foreign substances in the solution, they will modify
the action of the liquid upon the coalescing molecules.
The general shape, of a crystal, or, as it is usually called, the habit of a crystal,
thus depends upon the character of the distortion it has suffered during growth,
and the distortion is often a characteristic feature of the crystals of a given substance.
Calcite occurs as rhombohedra, scalenohedra, or hexagonal prisms, but all can be
referred to one fundamental form. However much the crystals may be distorted,
the angles between like faces have the same value, and in consequence, similar
faces can be imagined at the same distance from the centre of the crystal. In this
way, the ideal form of a crystal can be derived from that of a distorted crystal.
The term distortion, as distinct from habit, is also applied to certain crystals which
appear as if they had been mechanically deformed — e.g. twisted crystals of topaz.
If a layer of gelatine containing a solution of salt be allowed to dry slowly at
the temperature of the room, the water is lost quickest at the edges, and as the
solution approaches a state of supersaturation, crystals of the salt begin to separate.
CRYSTALS AND CRYSTALLIZATION 599
There are three ways in which the crystals may appear : (i) single crystals irregularly
distributed are formed — e.g. sodium chloride ; (ii) dendritic lines or branches of
crystals may extend from the edges towards the middle of the gelatin — e.g. potas-
sium chromate ; or (iii) a number of strips, concentric lines, or bands of crystals
separated by zones free or almost free from crystals may be formed — e.g. trisodium
phosphate or ferrous sulphate. E. Liesegang's explanation of the phenomenon of
rhythmic or periodic crystallization is as follows :
Before the first crystals are formed at the outermost edge there is present at this place
a zone of supersaturated salt solution. As the degree of concentration increases a spon-
taneous deposit of the salt necessarily occurs. These rows of crystals act as a nucleus
upon the supersaturated solution in their vicinity, i.e. the solute diffuses towards the crystals
and increases the size of the latter. In this manner an area poor in salt content is produced.
As the loss of water proceeds this area becomes supersaturated in its turn. Accordingly
the impoverishment in salt content becomes still more marked through the solute travelling
towards the nuclei. Finally, however, the migration in this area ceases entirely because
of the fact that the solution is entirely dried up. In the mean time, at a certain distance
from the first row of crystals, the salt solution reaches such a degree of supersaturation,
that the separating out of the salt begins spontaneously to occur once more. This row
of crystals grows just as the first one did, and the process is repeated indefinitely. The
more abrupt the precipitation, the closer together the bands of crystals. Since the first
spontaneous deposit of crystals occurs very suddenly, while their increase in size through
diffusion takes place more slowly, each band may exhibit a very different aspect as regards
its inner side respectively, i.e. dense aggregates of small crystals on the outer side and larger
crystals on the inner side.
The phenomenon of rhythmic crystallization is analogous with that of rhythmic
precipitation.il
Cleavage* — Crystals can usually be readily split along certain definite planes.
Thus, a fragment of calcite when struck sharply and lightly will split into a number
of fragments of variable size but similar form — rhombohedrons ; a crystal of rock
salt similarly treated splits into little cubes ; while mica splits into thin plates. Fluor-
spar, selenite, galena, and hornblende also illustrate the property very well.
Orthoclase— normal potash felspar — splits in two directions with an angle of 90°
between them, whereas the two cleavages of albite — ordinary soda felspar — are
inclined at an angle of about 86°. Galena, like rock salt, splits in three directions,
fluorspar in four. This property is termed cleavage. The cleavage of calcite into
six-faced rhombohedra was noticed by T. 0. Bergmann in 1773 ; and this suggested
to R. J. Haiiy that the ultimate components of crystals are irreducible formative
nuclei, so that J. Herschel could say :
From the moment that the genius of Haiiy discovered the general fact that the crystals
could be cloven or split in such directions as to lay bare their peculiar primitive or formative
forms, from that moment, mineralogy ceased to be an xinmeaning list of names, a mere
laborious cataloguing of stones and rubbish.
Fracture is different from cleavage, for it is irregular and has no definite relation
with the crystalline form. Cleavage is obviously a structural weakness related
to the mode of aggregation of the particles which build up the crystal, and it will
be obvious that once a crack is started in a cr}''stal, it will follow the line of least
resistance, that is, where the structural units exert least cohesion. If a number of
spheres or ellipsoids be packed together in a regular way, the cleavage will naturally
follow the direction where the number of contacts per unit area is least. If a layer
of spheres be formed so that each sphere touches six others, and a second layer be
laid directly over the interstices of the first one ; and a third layer over the inter-
stices of the second ; a pyramid is eventually obtained representing the half of a
cubical crystal. In such a configuration, the number of sphere contacts per unit
area parallel to the base is 2Vi/d^, where d denotes the diameter of each sphere ;
and 4/(^2 sphere contacts per unit area parallel to the front face. Hence, such a
crystal will break more easily along a direction parallel to one of the cubic faces of
600 INORGANIC AND THEORETICAL CHEMISTRY
the crystal since 2\/3 is less than 4. This corresponds with the fact that the cubic
crystals usually have an octahedral cleavage.12
Cleavage is an important factor in the identification of some of the conmioner minerals
in rocks. The possible cleavage forms given by A. J. Moses, The Characters oj Crystals
(New York, 1899), in the cubic system the chief cleavages are cube- — -e.g. galena ; octahedron' —
e.g. fluorspar, diamond ; rhombic dodecahedron — e.g. sphalerite. In the hexagonal system,
hasal pinacoid — e.g. beryl, pyrosmalite ; hexagonal prism — e.g. nephelite, apatite ; rhombo-
hedron — e.g. calcite, siderite ; hexagonal pyramid (rare) — e.g. pyromorphite. In the tetragonal
system, hasal pinacoid — e.g. apophyUite ; tetragonal prism — e.g. rutile, wernerite ; tetragonal
pyramid (rare)- — e.g. scheelite. In the rhombic system, pinacoid- — e.g. topaz, anhydrite ;
prisms or domes- — e.g. barytes ; pyramid- — e.g. sulphur. In the monoclinic system, clino-
pinacoid — e.g. orthocla^e, gypsum ; basal pinacoid — e.g. muscovite, orthoclase ; ortho-
pinacoid — e.g. epidote ; orthodome- — e.g. epidote ; prism' — e.g. pyroxene, amphibole ;
pyratnid (rare)' — e.g. gypsum. Triclinic crystals usually admit of equally easy cleavages
parallel to any principal plane.
It has been known for a long time that many substances in the act of crystalliza-
tion in the dark exhibit a bright sparkling light. G. Pickel (1787), ^^ for example.,
noted the phenomenon with potassium sulphate crystallizing from an aqueous solu-
tion ; C. G. Schonwald, with a mixture of sodium sulphate and potassium sulphate
crystallizing from aqueous solution, and he noted that the crystals also became
luminous when rubbed. E. Wiedemann applied the term tribo-luminescence to
the property exhibited by many crystalline substances of emitting a characteristic
phosphorescent light when rubbed or crushed, while the emission of a phosphorescent
light during crystallization is called crystallo-luminescence. The phenomenon
was also observed by J. A. Giobert, and he said that if the potassium sulphate
contains magnesium sulphate as impurity, the phenomenon does not occur,
and he adds that previous exposure to sunlight strengthens the effect, although
this observation has not been confirmed. K. S. L. Hermann observed the light
during the crystallization of cobalt sulphate at —12°. The phenomenon with
potassium sulphate has also been the subject of comment by J. J. Berzelius,
F. Wohler, and M. Sager ; with strontium nitrate by C. H. Pfaff and E. Stieren ; with
potassium acetate by J. A. Biichner ; with potassium chromate, selenate, sulphate,
and chloride, and arsenious oxide by H. Rose. Numerous other salts have been
added to the list by E. Brandrowsky, M. Trautz, J. Guinchant, D. Gernez, W. J.
Pope, E. F. Farnan, etc. H. Rose thought the phenomenon with arsenious oxide
was due to the change from an amorphous to a cry.stalline state ; 0. Lehmann said
from one crystalline form to another — rhombic to cubic (octahedral). E. Brandrow-
sky disproved both these suggestions, and he tried to show that it was an electrical
effect of the union of charged ions in the act of crystallizing from aqueous solution.
He thought that the light was white, but H. B. Weiser showed that with sodium
chloride the light is blue. The general cause of crystallo-luminescence is now
attributed to chemical action due to the union of ions — e.^. Na'-fCr=NaCl — but
it is not at all clear what is really the nature of the reaction, if one such does occur.
The light also has been spectroscopically examined, and so has the varying effects
of the nature and concentration of the precipitant ; of agitating the precipitating
solution ; of adding a colloid ; of increasing the viscosity of the solution ; and of
variations of temperature. J. Plotnikoff has described a tribolumini scope for
exhibiting the phenomenon.
The Florentine Academicians (1660) were the first to note the tribo-luminescence
of many substances — flint, sugar, salt, agate, jasper, and quartz — when rubbed in
darkness. 1* R. Boyle (1663) noted that diamonds exhibited the phenomenon ;
C. Mentzel (1675), W. Homberg (1730), J. Bernoulli and J.Cassini (1707), F. Hauksbee
(1709), C. F. du Fay (1723), and many others, made analogous observations in
the eighteenth century. A great many observations were afterwards recorded
showing that the phenomenon is fairly general. L. Tschugaeff, for instance,
tried over five hundred substances and found that about 25 per cent, showed
tribo-luminescence, and he found some cyclic carbon compounds exhibited the
CRYSTALS AND CRYSTALLIZATION 601
phenomenon to a marked degree. The early observers thought tribo-luminescence
is due to small particles of the solid becoming heated to incandescence by friction,
but T. de Saussure demonstrated that this explanation is very improbable when Jie
was able to obtain the effect by stroking a mass of calcium phosphate with a quill,
J. P. Dessaignes is tried to demonstrate that the luminescence is an electrical effect,
but J. Schneider showed that neither electrification nor heating are adequate ex-
planations of the phenomenon. After an examination of the main hypotheses
propounded to explain the phenomenon, H. B. Weiser concluded that tribo-lumi-
nescence is caused by chemical action, and photographic observations are in
agreement with the same hypothesis. The colour of the light is specific in that it
depends on the nature of the chemical reaction by which it is produced. All crystallo-
luminescent substances are tribo-luminescent, but converse of this does not obtain.
The emission of light in lum inescence has been explained, in part, in terms of the elec-
tronic hypothesis. In electro- luminescence, where the effects are produced by cathode
rays, X-rays, Becquerel rays, or canal rays, the result is connected with the displacement or
separation of electrons from th e constituent atoms of a substance ; in tribo-luminescence,
where the effects are produced by friction or by crushing, frictional electricity is likewise
involved in the displacement o r separation of electrons ; in chemi-limainescence, where
the effects are produced by chemical action, the result is attributed to the rupture of a
chemical bond which in turn is e ffected by the rupture of a valency electron ; and finally,
in photo-luminescence whether it be transient (fluorescence) or persistent (phosphorescence),
where the effect is due to the actio n of light, it is also supposed that the partial or complete
separation of electrons from the co nstituent atoms is directly concerned in the emission of
light.
References.
* T. 0. Bergmann, Opuscula physica et chemica, Lipsije, 1779.
2 G. Struever, Atti Accad. Torino, 4. 285, 1869.
^ C. F. G. H. Westfeld, Mineralogische Abhandlungen, Gottingen, 1767 ; T. 0. Bergmann,
Opuscula, Upsala, 1780 ; J. B. L. Rome de I'lsle, Essai de cristallographie, Paris, 1772 ; Cristallo-
graphie, Paris, 1783.
* R. J. Haiiy, Traite de mineralogie, Paris, 1801 ; C. R. Weiss, De indogando formarum
crystallinarum, Lepsiae, 1809 ; Abhand. Akad. Berlin, 289, 1815 ; V. von Lang, Lehrbuch der
Krystallographie, Wien, 1866 ; F. Mohs, Charakteristik des naturhistorische Mineralsy stems,
Dresden, 1820 ; J. F. C. Hessel, Gehler's Worterbuch, 5. 1023, 1830 ; Krystallometrie oder Krystallo-
Tbomie und Krystallographie, Leipzig, 1831 ; Ostwald's Klassiker, 88, 89, 1897 ; A. Gadolin, ib.,
75, 1896 ; Acta Soc. Scient. Fennicce, 9. 1, 1867.
5 A. Liversidge, Journ. Chem. Soc, 17. 1125, 1897.
* J. B. Rome de I'lsle, Cristallographie, Paris, 1. 517, 1783 ; A. G. Werner, Berg. Journ.,
1. 76, 1788 ; M. H. Klaproth, ib., 1. 299, 1788 ; A. F. de Fourcroy and L. N. VauqueUn, Ann.
Mus. Hist. Nat., 4. 405, 1804 ; L. J. Thenard and J. B. Biot, Bull. Soc, Philom., 1. 32, 1807;
Mem. Soc. Arcueil, 2. 206, 1809; L. J. Proust, Journ. Phys., 62. 226, 1806; A. Stromeyer,
Ann. Chim. Phys., (1), 92. 255, 1814 ; (1), 58. 205, 1806 ; C. F. Bucholz, Neues allg. Journ. Chem.,
3. 72, 1804 ; J. B. TrommsdorfP, Gehlen's Journ., 8. 152, 1809 ; P. Berthier, Ann. Chim. Phys.,
(1), 56. 300, 1806 ; L. N. Vauquelin, Journ. Mines, 11. 425, 1801 ; A. Laurent and P. Dejassieu,
ib., 30. 240, 1811 ; A. Laugier, ib., 36. 313, 1814 ; Ann. Mines., 3. 113, 1818 ; E. Mitscherlich,
Ann. Chim. Phys., (2), 19. 407, 1821 ; (2), 24. 264, 1823.
' P. Gaubert, Bull. Soc. Min., 25. 223, 1902 ; 0. Lehmann, Zeit. Kryst, 1. 463, 1876.
8 R. Boyle, Origin of Forms and Qualities, Oxford, 1666.
» J. B. L. Rome de I'lsle, Cristallographie, Paris, 1. 379, 1783 ; J. W. Retgers, Zeit.
Chem., 9. 297, 1892 ; Z. Weyberg, Kosmos, 35. 487, 1910 ; H. Gerhart, Tschermak's Mitt.,
185, 1910.
1" R. J. Haiiy, Traite de mineralogie, Paris, 1801.
^^ E. Liesegang, Bemmelen's Gedenksboek, 33, 1910; Zeit. Kryst., 50. 40, 1911 ; Scient. Amer.
Suppl, 87. 299, 1919 ; Roll. Zeit., 10. 225, 1912 ; E. Kiister, ib., 14. 307, 1914 ; Naturwiss., 2.
4, 1914.
12 G. D. Liveing, Proc. Roy. Inst., 13. 375, 1891.
13 G. Pickel, Taschenbuch der Scheidkunstler, 8. 55, 1787 ; J. M. Schiller, *., 12. 45, 1791 ; C G.
Schonwald, CreWs Ann., 2. 401, 1786 ; J. A. Giobert, Journ. Phys., 36. 256, 1789 ; Gren's Journ.,
2. 437, 1790 ; C. H. Pfaff, Schweigger's Journ., 15. 275, 1805 ; K. S. L. Hermann, ib., 40. 75,
1824 ; J. S. C. Schweigger, ib., 39. 247, 1823 ; 40. 271, 1824 ; J. A. Buchner, ib., 41. 221, 228,
1824 ; Repert. Pharm., 15. 441, 1823 ; A. M. Pleischl, Zeit. Phys. Math., 3. 220, 1835 ; J. J. Berzelius
and F. Wohler, Jahresb., 4. 44, 1824; 5. 41, 1825; H. Rose, Pogg. Ann., 35. 481, 1835;
52. 443, 585, 1841; J. Schneider, ib., 96. 282, 1855; E. Stieren, Pharm. Centrb., 400, 1836;
M. Sager, Archiv. Pharm., 36. 274, 1831 ; F. Penny, Phil. Mag., (4), 10. 401, 1855 ; W D. Bancroft,
602 INORGANIC AND THEORETICAL CHEMISTRY
Joum. Franklin Inst., 175. 129, 1913 ; W. D. Bancroft and H. B. Weiser, Journ, Phys. Chem.,
19. 319, 1904 ; H. B. Weiser, ih., 22. 439, 480, 576, 1918 ; E. Brandrowsky, Zeit. phys. Chem., 15.
323. 1894 ; 17. 234, 1895 ; M. Trautz, ib., 53. 1, 1905 ; Zeit. wiss. Phot., 2. 217, 1904 ; M. Trautz
and P. Schorigin, t6., 3. 80, 1905 ; J, Guinchant, Compt. Rend., 140. 1101, 1905 ; D. Gernez, ib., 140.
1134, 1234, 1337, 1905 ; W. J. Pope, Journ. Chem. Soc., 67. 985, 1895 ; P. Horing, Ber., 37. 1542,
1904 ; J. Burke, Chem. News, 78. 156, 1898 ; B. A. Rep., 810, 1898 ; E. Weideraann, Wied. Ann.,
34. 446, 1888; 0. Lehmann, Molekularphysik,'L&\^zig, 1. 217, 1888; E. F. Farnan, Intern. Cojig.
App. Chem., 20. 133, 1912 ; H. Kayser, Handbuch der Spektroscopie, Leipzig, 4. 678, 1904.
^* P. van Muschenbrock, Tentamina experimentorum naturalium captorium in Academia del
Cimento, Lugduni Batavorum, 2. 185, 1731 ; R. Boyle, Observations upon Diamonds, London,
1663 ; C. Mentzel, Lapis Bononensis in dbscuro lucens, Bielefeldiae, 1675 ; W. Homberg, Mem.
Acad., 445, 1730 ; J. Bernoulli and J. Cassini, Hist. Acad. Roy. Paris, 1, 1707 ; C. F. du Fay, ib.,
347, 1735 ; F. Hauksbee, Physico-mechanical Experiments, London, 194, 1719 ; E. Becquerel, La
lumikre, Paris, 1. 22, 1867 ; Ann. Chim. Phys., (3), 55. 5, 1859 ; H. F. Delius, CrelVs Ann., 3.
265, 1785; J. F. Henkel, Kleine mineralogische und chemische Schriften, Dresden, 99, 1744;
P. Henrich, Die Phosphorescenz der Korpen, Niirnberg, 425, 1820 ; J. H. Pott, Chymische Unter-
suchungen, Berlin, 1757 ; F. Hofmann, Hamburgen Mag., 5. 288, 441, 1750 ; B. Wilson, A Series
of Experiments relating to Phosphori, London, 92, 1775 ; T. Wedgwood, Phil. Trans., 1. 28, 270,
1792 ; C. de Bournon, ib., 92, 233, 248, 1802 ; D. de Dolomieu, Rozier's Obs. Phys., 39. 3, 1791 ;
T. de Saussure, ib., 40. 160, 1792; J. Plotnikoff, Prometheus, 30- 235, 1919.
" L. Tschugaeff, Ber., 34. 1820, 1901 ; Journ. Russian Phys. Chem. Soc., 36. 1245, 1904 ;
I. I. Ostromisslensky, ib., 42. 591, 1910 ; J. P. Dessaignes, Journ. Phys., 68. 444, 1809 ; 68. 5,
1809 ; 69. 5, 1809 ; 73. 41, 1811 ; 74. 101, 173, 1812 ; J. Schneider, Pogg. Ann., 96. 282, 1855 ;
F. Noggerath, ib., 150. 325, 1873 ; T. L. Phipson, Compt. Rend., 50. 316, 1860 ; J. Guinchant, ib.,
140. 1170, 1905 ; A. Karl, tb., 146. 1104, 1908 ; H. Becquerel, ib., 133. 199, 1901 ; D. Gernez, ib.,
140. 1337, 1905; 147. 11, 1908; Ann. Chim. Phys., (8), 15. 516, 1908; H. C. Lewis, Science,
(1), 3. 267, 1884; W. G. Levison, ib., (2), 19. 826, 1904; F. Krafft, Ber., 21, 2265, 1888;
J. Reuland, ib., 22. 3011, 1889 ; H. Decker, ib., 33. 2277, 1900 ; P. Horing, ib., 37. 1556, 1904 ;
P. Gucci and G. Grassi-Cristaldi, Oazz. Chim. Ital, 22. 1, 1892 ; A. Andreocci, ib., 25. 462, 1895 ;
29. 516, 1899 ; L. Brugnatelli, Zeit. Kryst., 27. 78, 1897 ; W. Arnold, ib., 27. 92, 1897 ; E. Weide-
mann and G. C. Schmidt, Wied. Ann., 54. 614, 1895 ; W. J. Pope, Nature, 59. 618, 1899 ;
T. Steel, ib., 59. 295, 1899 ; J. Burke, B. A. Rep., 810, 1898 ; Chem. News, 78. 256, 1898 ; H. Church,
ib., 85. 276, 1902 ; J. Precht, Phys. Zeit., 3. 457, 1902 ; J. Dewar, Proc. Roy. Soc., 68. 360, 1901 ;
F. Rinne, Centr. Min., 2Q2, 1902 ; H. Kayser, Handbuch der Spektroscopie, Leipzig, 4. 672, 1904 ;
M. Trautz, Ion., 2. 77, 1910 ; Zeit. phys. Chem., 53. 1, 1905 ; W. J. Pope, Journ. Chem. Soc, 67.
985, 1895 ; M. Trautz and P. Schorigen, Zeit. wiss. Phot., 3. 80, 1904 ; P. Imhoff, Phys. Zeit., 18.
78, 374, 1908 ; B. A. Lindener, Bull. Acad. St. Petersburg, 6. 999. 1910 ; W. Vernadsky, ib., 6.
1037, 1910 ; W. D. Bancroft, Journ. Franklin Inst., 175. 129, 1913 ; H. B. Weiser, Journ. Phys.
Chem., 22. 482, 576, 1918.
§ 4. The Crystallization of Solids en masse
In crystals, we see one of the many ways of judging the internal world of molecules and
atoms, and one of the weapons for conquering the invisible world of molecular mechanics
which forms the main object of physical chemistry.- — D. I. MENDELfeEPF.
When a supercooled liquid commences to solidify, rays of the solid grow into the
liquid with a definite velocity,^ and H. A. Wilson has shown that the relation between
the velocity of solidification and the supercooling of some liquids fits the hypothesis
that solidification is due to the difference between the internal pressure in liquid
and solid, and that the molecules at the surface of a liquid are urged from liquid
to solid by this difference in internal pressure. Modifying J. H. van't Hofi's method
for calculating the osmotic pressure of a salt in solution from the lowering of the
vapour pressure, he obtained the formula : Velocity of solidification=yt(Tw— T)/F,
where Tj^ represents the melting point of the solid on the absolute scale ; T, the
temperature at the surface separating liquid and solid, so that T^—T measures
the supercooling of the liquid ; F, denotes the viscosity of the liquid ; and A; is a
constant. Hence, the velocity of solidification of a pure substance is directly
proportional to the actual supercooling of the solidifying liquid, and inversely as
the viscosity. The results obtained with a number of substances agree with this
hypothesis, although there is a disturbance due to the production of heat which
accompanies solidification, and which raises the temperature at the surface of
solidification.
CKYSTALS AND CRYSTALLIZATION
603
Fig. 12. — Idealized Dendrite of Copper.
Metals, during solidification, have a characteristic tendency to assume dendritic
forms rather than simpler crystals. Even in the natural state, crystals of native
gold, silver, and copper — all in the cubic system — frequently appear as dendritic
growths. E. S. Dana, in a paper on the Crystallization of native copper,^ has indicated
a variety of these forms. The idealized representation, Fig. 12, illustrates a common
method of growth. .- , ...
It seems as if the nuclei of a crystallizing metal first stretch out long branching
lines, each nucleus secures for its own
crystal a large share of territory, 3 and
then proceeds " in a leisurely manner to fill
up the gaps." Under high magnification,
it appears as if the branches radiating from
adjacent crystals stop growing before
they actually meet, and so form a kind of
neutral territory between the tips of the
branches. When the growth of all the
radiating branches has been arrested in
this way, the effect is almost the same as
if each nucleus was enclosed by a cell-like
boundary. As shown by D. K. Tschernoff
and others, the crystal grains of ordinary
cast metals thus appear to be produced
by the mutual interference of adjacent
dendrites, as illustrated diagram matically
in Fig. 13.
0. Lehmann,^ in his Molekularphysik,
gives a satisfactory explanation of the
phenomenon of dendritic growth in the case of supersaturated solutions : In the
immediate vicinity of a growing crystal, a zone of liquid no longer supersaturated
is formed owing to the removal of the solute, by the crystal. Further growth can
occur only when the concentration of the solution in the immediate neighbourhood
of the growing crystal is increased by diffusion or convection. The growth of the
crystal thus depends upon the rate of supply of dissolved material from the surround-
ing supersaturated solution. The concentration currents- — Diffusionsstrdme — thus
set up are stronger where the difference of concentration is
greatest, and this must be at the sharp angles of the grow-
ing crystal, because the crystal there presents the greatest
surface to the liquid. Growth is thus accelerated with
increasing velocity in the vicinity of the crystal augles.
As a result, the crystal extends most rapidly in the direc-
tion of the concentration currents or supply columns
bringing the food, so to speak, necessary for further
growth. When the supersaturation is so reduced that
the crystal can grow but slowly, the concentration currents -p^^ ^3 —Diagrammatic
will virtually cease, and the spaces between the branchlets illustration of the
will be gradually filled up. Crystallization of Pure
0. Lehmann's explanation can be extended to the case of Metals,
a cooling metal, where a certain amount of heat (latent
heat of fusion) must be locally developed about each crystallizing centre ; and the
crystal must cool down to the crystallizing point before further growth can occur.
Local currents of undercooled liquid will flow quickest from the surrounding liquid
in the direction of the sharpest angles of each growing crystal, because there the
temperature gradient is steepest. Hence, growth will be fastest in the direction of
the currents of undercooled liquid.
G. Quincke's hypothesis of the cellular structure of metals. — When two
partially miscible liquids — e.g. benzene with a little concentrated potash lye ; or
604 INORGANIC AND THEORETICAL CHEMISTRY
paraffin with one per cent, of a one per cent, soap solution — are shaken together,
an emulsion is formed ; and when one of the liquids is in large excess, the
other may be distributed throughout the mass in such a way as to form what
G. Quincke ^ called foam cells — Schlaumkammern — consisting of thin cell- walls of
the liquid present in small quantity, enclosing drops of the other liquid. In
G. Quincke's study of the formation of the emulsoidal structure, he showed how a
similar structure might be obtained with solutions in which the cell-wall difiered
from the contents only in the concentration of the solute. He also suggested the
hypothesis that the first stage in crystallization is the separation of the liquid into
two immiscible phases, one of which is in relatively small quantity, and that the two
liquids assume the foam-cell structure. He said that the purest of liquids still
contains enough impurity to produce an eutectic which forms the cell-walls
separating the primary metal. Crystallization then takes place within the foam
cells, and the cell-walls are represented in the solid mass by the boundaries of the
crystal grains. The arrangement of the crystalline particles, when actual solidifica-
tion begins, is determined by the size and shape of the foam cells. This attempt
to explain the cellular structure of metals does not appear to be satisfactory, since,
although a homogeneous liquid at a certain temperature may separate into two
immiscible liquids on cooling, there is nothing at all to show the phenomenon is
so general as is required by G. Quincke's hypothesis. In those cases where crystalli-
zation has been observed under the microscope, the nuclei first formed control the ulti-
mate structure, and there is no sign of a pre-existing foam-cell form. C. H. Desch
has sought evidence of the assumed phenomenon and concluded that the grains
of a solidifying metal have a tendency to assume the shape of foam cells, but
he was not able to decide if foam cells are actually formed, or if the solidification
of the metal proceeds from nuclei.
A cellular or tesselated structure is sometimes observed on the surface of cooling
liquids. E. H. Weber (1855) * has recorded the formation of polygonal areas during the
slow evaporation of a mixture of alcohol, water, and gamboge on a microscopic slide ; and
L. Frankenheim (1860) noticed a similar phenomenon during the evaporation of a solution of
sulphur in turpentine. J. Thomson (1882) obtained a similar structure on the surface of
soapy water; and C. Dauzere on the surface of mixtures of beeswax and paraffin or salol.
According to H. Benard (1901), the best results can be obtained by exposing a layer of sper-
maceti, 0*4 to 2"0 mm. thick, in a metal trough 15 cm. diameter. The lower surface must be
uniformly heated, and great care taken to eliminate disturbing influences. The partitioning
approximates to an arrangement of regular hexagons. According to J. Thomson, the
upper surface of the liquid is cooler than the portion below. Convection currents are set
up, and the warmer liquid below ascends in vertical columns, spread out, and descends
vertically downwards. The polygonal areas are the boundaries where the descending
currents meet. The currents can be rendered visible, and photographed, by using fine
particles of gamboge, lycopodium, graphite, etc., in suspension in the liquid. The surface
polygonal areas are thus the upper surfaces of a series of prismatic columns produced by
convection currents- — tourhillions celliiJaires- — their axes are always vertical, never hori-
zontal, and therefore the phenomenon is not the same in kind as that which gives rise to
the cellular structure of metals. The phenomenon was studied mathematically by Lord
Rayleigh in 1916.
Pure metals are aggregates of crystals which have been prevented from assuming
a regular geometrical form by the crowding which occurs during the growth of
neighbouring crystals. The crystals have grown more or less simultaneously
and independently from a number of independent centres of crystallization. The
facts must have been known to Robert Hooke 7 in 1665, and to R. A. F. de Reaumer
in 1722 ; but H. C. Sorby was the first to show, in 1864, that a polished and etched
surface of a metal is cut up into a number of polyhedral parts suggesting that the
metal has a kind of cellular structure when examined under a suitable magnification
and illumination. Under the microscope, also, the crystal boundaries appear as
dark lines which are developed by the etching liquid attacking the surfaces of the
various crystals at different rates. Although the internal structure of opaque
crystals cannot be established so readily as would be the case if the crystals were
transparent and could be examined under polarized light, yet, by etching the
CRYSTALS AND CRYSTALLIZATION 605
surface with suitable liquids, or by casting metals like cadmium or tin against a
smooth glass surface, the evidence of an internal oriented structure is unmistakable.
Thus, J. A. Ewing ® has shown that —
The surface of each grain consists of a multitude of geometrically similar pieces, parallel
to one another so that their corresponding facets are all oriented one way. They are
oriented in different ways as we pass from grain to grain, but in any one grain they face
one way, and in consequence of that the light which falls on the grain is reflected in a
perfectly uniform manner over the whole expanse of that grain, although it is reflected in
a very different manner from the surface of any other grain. Over each grain the brightness
is uniform, because the little surfaces are acting equally as regards the reflection of light.
The idea is well shown by the photograph by J. E. Stead, Fig. 14, from a specimen
of iron with 4*5 per cent, of silicon in solid solution, and deeply etched. Cubic
crystals of iron and octahedral crystals of copper have been obtained in an analogous
form.
Although the cohesion of the molecules of a homogeneous liquid or amorphous
solid can be explained by intermolecular attractions, auxiliary hypotheses must
be invented to explain the stability of crystals, for the orientation of the structural
units makes it appear as if the attractive forces are to some extent polar, because
they act most favourably in certain directions. With aggregates of crystals, the
distance between the surface molecules in
two different systems of adjacent crj^stals
must be much greater than between the
molecules within each of the cr}'^stals ;
and consequently, it would appear as if
the attraction of crystals for crystals should
be less than the molecular attraction within
the crystals themselves ; or as if intra-
crystalline cohesion should be greater than
intercrystalline cohesion. The inter-
crystalline boundaries or joints of a piece
of metal of normal purity might therefore
be expected to be a surface of weakness.
W. Rosenhain ^ has emphasized the fact
that the converse is usually the case. J. A.
Ewing and W. Rosenhain have shown ^^^^ i4.~Surface of Silicon Steel. (J. E.
that the fracture of a piece of Swedish Stead.)
iron under a tensile load does not usually
follow the intercrystalline boundaries, but rather cuts across the crystals themselves ;
otherwise expressed, the intercrystalline boundaries of normally pure metals are
not surfaces of special weakness, but are rather surfaces of special mechanical
strength.
The attempt has been made to explain the greater strength of the intercrystalline
boundaries by assuming that the crystals grow, not by accretion layer by layer,
but rather by the shooting out of dendritic branches which continue to grow until
they meet one another, and finally interpenetrate one another so that the crystals
are bound together by the interlacing of dendritic branches. W. Rosenhain and
D. Ewen, however, reported that the " study of a large number of crystal boundaries
does not seem to show sufficient evidence that the process is universal enough to
account for the whole of the phenomena," and they favour the hypothesis
that there is a cementing medium between adjacent crystals in the inter-
crystalline boundaries which makes the crystals adhere together with special
firmness.
Although J. 0. Arnold lo has shown that a deleterious impurity may considerably
weaken intercrystalline boundaries, for the presence of 0"1 per cent, of bismuth in
gold was sufficient to surround the ductile crystals of the latter metal with a brittle
envelope which readily fractured under a blow, yet many metals are known to be
606 INORGANIC AND THEORETICAL CHEMISTRY
of sufficient purity to render highly improbable the assumption that the inter-
ctystalline cement consists of impurities, more fusible metals, or eutectics. Con-
sequently, Rosenhain concludes that the cement, if it exists at all, must be chemically
the same material as the crystals themselves, and it is accordingly inferred that
the crystals of a metal of normal purity are surrounded by a thin layer of metal
in an amorphous condition which acts as a cement, and which determines the
mechanical strength of the metal itself.
It is assumed that the forces at work during crystallization are such as to
prevent the last vestiges of mother liquid from crystallizing, and that this liquid
retains the amorphous condition while the metal cools down to the ordinary tem-
perature. If the formation of a crystal requires not only the orderly arrangement
of the structural units in the crystal, but also a grouping together of the molecules
into aggregates to form the structural units, the liquid residue in the interstices
when the growing crystals are nearly in contact will be unable to crystallize because
there is not sufficient space for the aggregation of the molecules into the structural
units.
The mechanical strength of a metal is determined by the strength of the inter-
crystalline cement ; and if this cement be weakened or destroyed, the metal will
show intercrystalline weakness. An intercrystalline cement was postulated by
M. Brillouin, J. E. Sears, C. D. Bengough, and others n to explain the behaviour of
metals when subjected to deformation under various conditions. For example, the
effect of temperature on the tensile and elastic properties of metals is explained by
assuming that at low temperatures the cement will accommodate itself to stresses in
virtue of its elasticity while the crystals accommodate themselves to stresses by
plastic deformation ; as the temperature rises, the cement weakens, and its elasticity
diminishes ; when the cement has weakened sufficiently, the fracture under stress
will be intercrystalline, and permanent elongation will occur under very small
stresses because the cement does not prevent the crystals from sliding over one
another.
References.
1 D. Gemez, Journ. Phys., (2), 2. 159, 1904 ; G. Tammann, ZeiL phys. Chem., 23. 326, 1897 ;
J. Friedlander and G. Tammann, ib., 24. 152, 1897 ; H. A. Wilson, Phil Mag., (6), 50. 238, 1900.
2 E. S. Dana, Amer. Journ. Science, (3), 32. 413, 1886.
» J. A. Ewing, Journ. Inst. Metals, 8. 4, 1912.
* G. Chautaud, Ann. Mines, (9), 17. 110, 1900; D. K. Tschernoff, Eev. Univ. Mines, (2),
7. 129, 1880 ; 0. Lehmann, Molekularphysik, mit besonderer Beruchsichtigung mikroskopischer
UrUersuchungen, Leipzig, 1. 337, 1888.
6 G. Quincke, Ann. Physik, (4), 7. 631, 1902 ; (4), 9. 1, 1902 ; (4), 18. 1, 1905 ; Per. deut.
phys. Ges., 5. 102, 1903 ; Proc. Boy. Soc, 76, 431, 1905 ; 88. 60, 1907 ; Internat. Zeit. Metalog.,
3. 23, 1912; J. Plateau, Statique experimentale et theorique des liquides, Paris, 1873; S. U.
Pickering, Journ. Chem. Soc, 91. 2001, 1907; C. H. Desch, Journ. Inst. Metals, 11. 57, 1914; ^
22. 241, 1919.
« E. H. Weber, Pogg. Ann., 94. 452, 1855 ; L. Frankenheim, ib., 111. 1, 1860; J. Thomson,
Proc. Phil. Soc. Glasgow, 13. 464, 1882 ; H. Benard, Les tourbillons cellulaires dans une nappe
liquide propagie de la chaleur par convection en regime permanent, Paris, 1901 ; C. Dauzere, Journ.
Phys., (4), 4. 892, 1907 ; (4), 7. 930, 1908 ; G. Cartaud, Bev. Mit., 4. 819, 1907 ; C. H. Desch,
Journ. Inst. Metals, 11. 57, 1914 ; 0. Lehmann, Molekularphysik, Leipzig, 1. 271, 1888 ; Lord
Rayleigh, Phil. Mag., (6), 32, 529, 1916.
' R. Hooke, Micrographia, London, 1665 ; R. A. F. de Reaumer, L'art de convertir le fer
forge en acier, Paris, 1722 ; H. C. Sorby, B. A. Bep., 189, 1864.
8 J. A. Ewing, Journ. Inst. Metals, 8. 4, 1912 ; J. E. Stead, Journ. Iron Steel Inst., 53. i, 145,
1898 ; J A. Ewing and W. Rosenhain, Phil Trans., 193. 353, 1900 ; 195. 279, 1901 ; G. Cartaud,
Ann. Mines, (9), 17. 110, 1900.
» J. A. Ewing and W. Rosenhain, Proc. Boy. Soc, 65. 85, 1899 ; W. Rosenhain, Journ. Iron
Steel Inst., 70. ii, 212, 1906 ; W. Rosenhain and D. Ewen, Journ. Inst. Metals, 8. 149, 1912 ;
10. 119, 1913.
10 J. 0. Arnold and J. Jefferson, Engineering, 61. 176, 1896 ; F. Osmond and J. Werth, Ann.
Mines, (8), 8. 1, 1885 ; H. Behrens, Das mikroskopische Gefuge der Metalle und Legierung, Leipzig,
1894.
" M. Brillouin, Ann. Chim. Phys., (7), 13. 377, 1898 ; J. E. Sears, Trans. Cambridge Phil
CRYSTALS AND CRYSTALLIZATION 607
Soc, 21. 105, 1909 ; G. D. Bengough, Journ. Inst. Metals, 7. 176, 1912 ; W. Rosenhain and
D. Ewen, ib., 8. 149, 1912 ; 10. 119, 1913 ; W. Rosenhain, B. A. Sep., 427, 1913 ; F. Osmond,
Journ. Iron 8teel Inst., 80. ii, 61, 1911.
§ 5. The Internal Structure of Crystals
Imagine two hundred brilliant violin players playing the same piece with perfectly
tuned instruments, but commencing at different places selected at random. The effect
would not be pleasing, and even the finest ear could not recognize what was being played.
Such music is made for us by the molecules of gases, liquids, and ordinary solids. They
may be highly gifted molecules with a marvellous internal structure, but in their activity,
each disturbs the others. A crystal, on the other hand, corresponds with the orchestra
led by a vigorous conductor when all eyes intently follow his nod, and all hands follow the
exact beat. This picture enables us to understand how crystals can exhibit whole ranges
of phenomena quite wanting in other bodies. To me, the music of physical law soimds
forth in no other department in such full and rich accord as in crystal physics.- — ^W. Voigt.
Crystals are not only peculiar in the regularity o! their external shape,
but they also possess a definite internal structure. — ^In illustration, E. Bartho-
linas (1669) noticed that if a rhombohedral crystal of Iceland spar be placed
over a black spot on a strip of white paper, Fig. 15, the spot appears to
be doubled and one of the two spots appears clearer than the other. If the
rhombohedron be rotated, the spot which seems to be nearest to the eye appears
to rotate about the other. There is, however, one direction in which the dots
viewed through the crystal seem to coincide. It
is readily demonstrated that the beam of light in
passing through the prism (in all but one direc-
tion) is split into two rays, for if a single beam of
light be passed through the crystal, two beams of
light will emerge. This property of splitting the
ray of light into two different rays is called
double re&action, and the crystal of Iceland spar ^ ,^ -r^^ . .• r., -r^ ,,
• J , 1 J 1,7 r s' i,-i i-i, -T Fig. 15.- — Illustration of the Double
is said to be doubly refracting— while the cry- Refraction of Iceland Spar,
stal is singly refracting in the one direction
in which the image of the two spots is not doubled. One of the two rays
obtained by double refraction is called the extraordinary ray- — R. H. Haiiy's
rayon d' aberration — and the other the ordinary ray — B.. J. Haiiy's rayon
ordinaire. In 1690, C. Huyghens discovered that each of these rays has certain
peculiar properties different from those of an ordinary ray of light. Each ray is
said to be polarized. The splitting of a ray of light into two rays by double refraction
is called polarization of Ught. The effect of polarization on a beam of light
can be illustrated as follows : The ray of light is supposed to be vibrating in all
possible directions perpendicular to the path of the ray ; 3*c ; on passing through the
prism these vibrations are resolved into vibrations in two planes at right angles
to one another : + ; and are at the same time separated into two rays, — and |,
vibrating in different planes at right angles to one another. The resolution of the
heterogeneous mixture of vibrations into linear vibrations in one direction is called
the plane polarization of light. If a crystal of tourmaline, also doubly
refracting, be placed between the eye and the crystal of Iceland spar, Fig. 15,
and then rotated, the two spots will alternately appear and disappear.
That direction in the crystal of Iceland spar in which light is singly refracted
is called the optic axis of the crystal, and generally the directions parallel to which
there is no double refraction are called the optic axes of a crystal. Crystals
which have one optic axis — or one axis of no double refraction — are said to be
uniaxial crystals in contrast with crystals with two optic axes — or two axes
of no double refraction — which are said to be biaxial crystals. Examples of
uniaxial crystals are calcite, quartz, borax, sugar, sodium nitrate, potassium
608 INORGANIC AND THEORETICAL CHEMISTRY
thiosulphate, potassium or ammonium phosphate, etc. Examples of biaxial
crystals are potassium nitrate, sodium sulphate (glauberite) , cerussite, aragonite,
gypsum, etc.
Polarizing microscope.— In the polarizing microscope, a prism of Iceland spar — called
a Nicol's prism or simply a nicol — is so arranged that the ordinary ray from a parallel
ray of light is reflected to one side, and the extraordinary ray alone passes through the prism.
The extraordinary ray of polarized light can then be passed through a second Nicol's prism
which can be turned about its axis. When the nicols are crossed, the field will appear
dark in spite of the fact that there is nothing but two transparent prisms between the eye
and the source of light. The first Nicol's prism is called the polarizer and the second the
analyzer. When the two prisms are at right angles to the position where the field has
its maximum darkness, the nicols are said to be parallel. When the nicols are parallel,
the field appears of maximum brightness, and the extraordinary ray passes through both
prisms.
If a doubly refracting crystal be placed between crossed nicols, instead of com-
plete darkness, the crystal appears to be more or less illuminated on a dark back-
ground (double refraction). D. Brewster (1821) showed that the behaviour of
crystals in polarized light (between crossed nicols) is an important means
of distinguishing doubly refracting crystals from cubic crystals. "Very few
crystals fail to show light between crossed nicols when examined in a suitable
instrument, the polarizing microscope ; and crystals can be classified into families
according to their peculiar action on polarized light, because each system interferes
with polarized light in a characteristic way. Let a crystal be laid flat on a glass
plate on the stage of a microscope and the eyepiece, with cross wires, so fixed
that an edge of the crystal is parallel to one of the cross wires ; let the polarizer
and analyzer be placed perpendicular to each other. The angle through which the
crystal must be rotated in order to produce darkness is called the angle of optical
extinction. If the crystal extinguishes or disappears when the edge of reference
is parallel with the cross hairs of the eyepiece, and is brightest midway between,
the crystal is said to have parallel or straight extinction with respect to that
edge ; and if the position of extinction is inclined or oblique to the cross-hairs,
the crystal has oblique extinction. Tetragonal, hexagonal, trigonal, and rhombic
crystals show parallel extinction, while monoclinic and triclinic crystals give oblique
extinction.
A pencil of parallel polarized light. — If the direction of vibration of the plane
polarized light is continuously rotated as it passes through the crystal, the
phenomenon is called rotatory polarization, and the crystal is said to be optically
active. Optically active crystals rotate the plane of polarization of light even in
sections perpendicular to the optic axes. In 1811, D. F. Arago discovered the
optical activity of quartz. Sodium chlorate also exhibits the phenomenon.
When a pencil of parallel light rays is passing through the polarizer, the inter-
position of a plate of glass or of an isotropic crystal between the crossed nicols
will not affect the field of vision which will remain dark ; on the contrary, with
plates of certain other crystals — quartz, sodium chlorate, cinnabar, etc. — cut
perpendicular to the optic axis, the field of vision becomes more or less clear
according to the thickness of the plate interposed, and this in spite of the
fact that the nicols are crossed. The plates should be not less than 0-2 mm.
thick. In order to re-establish the original darkness, the analyzer must be
rotated through a certain angle, showing that the interposed plate of, say, quartz
rotated the plane of vibration of the polarized rays from the first nicol. If the
analyzer be turned through a certain angle clockwise (viewed from the front) in
order to restore darkness, the crystal is said to be dextro-rotatory, because the
crystal rotates the plane of polarization from the left to the right of an observer
receiving the light and not towards the right in the direction the wave of light
progresses. The converse of this applies for laevo-rotatory crystals. Since the
angle of rotation is proportional to the thickness of the crystalline plate under
examination, it is conventionally referred to a plate 1 mm. thick. The angle of
CRYSTALS AND CRYSTALLIZATION 609
rotation is also dependent on the wave-length of the light used, and therefore the
character of the light should be specified — yellow sodium light, designated after
Fraunhofer's D-line in the solar spectrum, is commonly employed, and the angle
of rotation is symbolized an. Examples : The angle of rotation aj, for sodium
bromate, NaBrOg, is 217°; for sodium thioantimoniate, Na3SbS3.9H20, 2-37°;
for sodium chlorate, NaClOg, 3-14° ; potassium lithium sulphate, KLiS04, 3*44° ;
potassium dithionate, K2S2O6, 8-39° ; quartz, 21-723° ; and for sodium periodate,
NaI04.3H20, 23*3°. When these substances are dissolved, or fused, the rotatory
power is lost ; and it is accordingly inferred that the cause of the rotatory power
must reside in the way the molecules are structurally grouped in the crystals. On
the other hand, a great many compounds of carbon possess this rotatory power
when in the liquid state, in solution, and sometimes even in the state of vapour,
e.g. camphor, etc. Consequently, it is inferred that with these substances the
rotatory power is not due to the structure of the crystal, but is produced by
arrangement of the atoms in the molecule itself, and is concealed by the orienta-
tion of the molecules in the crystal. Strychnine sulphate is optically active in
solution and still more so in the solid state, so that the optical activity is
determined by the orientation of the molecules in the crystal, as well as by the
arrangement of the atoms in the molecule.
It is not possible to calculate the rotatoiy power of a substance with certainty from the
rotatory power of its aqueous solution. In 1838, J. B. Biot ^ noticed that the rotatory
power of tartaric acid in aqueous solution increased with increasing dilution, but R. Pribram
failed to obtain constant values for the rotatory power even with veiy great dilutions.
The decrease in the specific rotatory power of dilute solutions of the alkali tartrates with
increasing dilution was attributed by R. von Sonnenthal to ionization. G. H. Schneider
found that concentrated solutions of malic acid were dextro-rotatory ; dilute solutions,
Isevo-rotatory ; and that 34 per cent, solutions appeared inactive. Hence, it is highly
probable that the so-called inactive solvent does exert some kind of action on the solute- —
maybe by breaking down molecular aggregates ; forming unstable hydrates ; changing
the configuration of the atoms in the molecules, or by ionizing the salt in solution.
In 1817, J. B. Biot showed that the magnitude of the angle of rotation of
solutions of optically active substances depends upon (i) the thickness of the layer
of solution observed ; (ii) the concentration — i.e. the mass per unit volume ;' (iii) the
temperature ; and (iv) the wave-length of the light employed. The specific
rotatory power or the specific rotation of a substance at a particular temperature
for a particular wave-length of light is the rotation produced by a decimetre of
solution which contains one gram of the active substance per cubic centimetre.
If w grams of an optically active substance are dissolved in v c.c. of a solvent, and
the observed angle of rotation a with a column of liquid I decimetres long, the
specific rotatory power, [a], of the solution will be [a]=avllwy on the assumption that
the solvent is without influence on the result. This assumption is usually justifiable
in the case of aqueous solutions, but not with other solvents. The change in the
rotatory power of certain substances when dissolved in solvents upon which they
exert no apparent chemical action, is explained by assuming that the solvent either
unites with the solute, or else exerts a specific attraction so that the motions of the
molecules of the solute are modified. A temperature of 20° is often taken as a
standard of reference.
Example.- — If 20-2 grms. of a substance is dissolved in 79 '8 grms. of water and the
solution has a specific gravity of 1*0842 ; and when a layer 2 d.m. thick has an angle of
rotation aD = 29°, what is the specific rotation ? Here, 1 = 2 ; w = 20-2 ; t'= weight -i- specific
gravity = (79-84-20-2)-r-l-0842; a=29 ; hence, [a]u = 66-2r.
The product of the specific rotatory power and the molecular weight of a substance
is called the molecular rotatory power. The numbers so obtained are often divided
by 100 in order that smaller numbers may be used. Special instruments called
'polarimeters are used to measure the rotatory power of optically active solutions,
for they do this more accurately than is possible in the ordinary polarizing microscope.
VOL. I. 2 E
610
INORGANIC AND THEORETICAL CHEMISTRY
A convergitig cone of light. — A thin plate of a doubly refracting uniaxial crystal,
cut at right angles to the optic axis, and placed between crossed nicols with a cone
of convergent light incident on the polarizer, furnishes a series of circular, coloured,
and concentric rings traversed by a dark cross. Fig. 16 ; and a biaxial crystal
treated in a similar manner exhibits a double series of elliptical (lemniscate) rings
traversed by a narrow bar, and separated by a broad bar. Fig. 17. These figures
are called interference figures. When the analyzer is rotated, the distribution of the
colours varies, and the interference figures change in a characteristic way in crystals
of different mineral species. The black cross of uniaxial crystals changes into a
white one, Fig. 16, and the colour of the rings changes to the complementary
tints when the nicols are parallel. Similarly, with biaxial crystals. Fig. 17, rotating
the analyzer breaks up the cross and develops two dark brushes. Fig. 17, each
of which traverses one of the systems df rings. The crystals of some substances
require the analyzer to be rotated to the right, and other crystals to the left in order
to develop a given succession of colours^ — e.g. different specimens of quartz, for
example, may show what J. Herschel called right- and left-handedness, and some
samples again, principally the purple crystals of quartz (amethysts), may show
both right- and left-handedness in one specimen.
The properties of crystals axe not always the same in different direc-
tions.— The hardness, elasticity, crushing strength, rate of solution in acids,
optical, thermal, and electrical properties, are generally different in different direc-
tions. This means that different results are usually obtained when the elasticity,
Nicols crossed. Nicols parallel. Nicols crossed. Analyzer rotated 45°.
Fig. 16. — ^luterference Figures — Fig. 17. — Interference Figures — Biaxial
Uniaxial Calcite. Potassium Nitrate.
refraction of light, thermal expansion, etc., of a crystal are measured in different
directions. Thus, the coefficients of thermal expansion of quartz along the axes
designated a and c are respectively 1515 X 10" » and 807 X 10-^ — so that one is nearly
twice as large as the other ; with adularia (felspar) the coefficients along the three
axes are respectively 1569x10-8, 65*9x10-8, and 291x10-8— corresponding
nearly with the ratio 24 : 1 : 4. A substance, apparently homogeneous, may there-
fore exhibit privileged directions for the propagation of any particular form of
energy — thermal, electrical, optical, magnetic, or elastic. Any medium in which
any natural phenomenon is not produced with the same intensity in every
direction is said to be anisotropic (avto-os, unequal ; TpiireLv, to turn) or
seolotropic (aioAo?, changeful) for that phenomenon, for the body is dissym-
metrical with regard to that phenomenon. A body whose optical, magnetic, thermal,
electrical, elastic, or other property depends upon direction, is anisotropic. All
these forms of anisotropy can exist simultaneously in a body. When the pro-
perties of a substance are the same in all directions, it is said to be an isotropic
substance (to-os, equal) — e.g. gases, most liquids, unstrained glass, and, so far as
the optical, thermal, and electrical properties are concerned, unstrained crystals
belonging to the cubic system.
In H. de Senarmont'a experiment (1847),^ a slice of quartz is cut perpendicular to the
long axis and another slice is cut perpendicular to this ; each slice is covered with wax,
and pierced at the centre so that a hot wire can be inserted. The wax naturally melts
about the hot wire. In the former case, the molten wax will form a circle-—^, Fig. 18 ;
CRYSTALS AND CRYSTALLIZATION
611
and in the latter case, an ellipse — By Fig. 18. This shows that the thermal conductivity of
the crystals is different in different directions. In H. L. Bowman's experiment one end of
a heated wire in contact with a face of a crystal of gypsum gave a white, opaque, elliptical
area owing to dehydration. The ratio and direction of the axes of the ellipsoid varied for
different faces- — on the cleavage face, the ratio was ri29.
If a crystal of calcite be hung in a beaker of dilute hydrochloric acid by means
of a platinum wire, solution does not occur at a uniform rate over the whole surface,
but the rate of solution of the crystal is faster in one direction than in another.
Plates have been cut parallel to the different faces of different crystals, and the
edges protected with wax, and measurements made of the amount dissolved after
the plates had been immersed in the selected solvent for a given time. The velocity
of the attack by hydrochloric acid is about, 1'15 times greater when the surface
exposed to the action is perpendicular to the principal axis than when the surface
is parallel therewith. 3 A. Wolff found all the faces of crystals of Mohr's salt
dissolve at the same rate ; and A. Korbs noticed very little difference in the rates
of solution of the different faces of crystals of sodium chloride, alum, and potassium
nitrate, but with copper sulphate and potassium ferrocyanide wide differences
were observed — in the former case, 37 per cent. ; and in the latter, 86 per cent.
Still further, the velocity of propagation of light through crystals of the cubic system
is the same in all directions, but not with members of the other systems.
The attack of a crystal face by a reagent does not necessarily commence at the
same time at all points, but proceeds more rapidly in some parts than in others.
If the attack be stopped at the right time, the attacked face will be pitted mth
Vb
♦♦
♦.^
Fig. 18.- — -H. de S6narmont's
Experiment.
Fig. 19. — Etch Figures of
Galena (diagrammatic).
Fig. 20.— Etch Figures of
Sylvite (diagrammatic).
little [angular cavities, many of microscopic size, called etch figures, or corrosion
figures. Passing a moist cloth rapidly over the surface of an octahedral crystal
of alum will suffice to develop triangular cavities with sloping sides, and similar
to those sometimes found naturally on the octahedral faces of the diamond. With
the right conditions — solvent, time, and temperature — the etch figures will be in
parallel positions plane for plane. The etch figures on similar faces will be all
alike, but unlike on dissimilar faces. The etch figures conform to the symmetry
of the class to which the crystal belongs, and they therefore serve as an important
clue in determining the symmetry of a crystal. The etch figures on cubes of galena
are symmetrical to the nine planes of symmetry, while those on cubes of sylvite
are symmetrical only to the axes. The etched pits are bounded by a number of
minute faces with complicated indices which are generally known as vicinal faces.
Experiments on the rates of solution of crystal faces must always be affected
by errors owing to modifications in the surface area by (i) the pitting of the
faces ; and (ii) the rounding of the edges. V. Goldschmidt * has studied the etch-
figures of crystal faces and of the successive forms assumed by a sphere of calcite
when subjected to the action of phosphoric, hydrochloric, nitric, formic, or acetic acid.
The sphere is first etched in tracts which correspond with the natural faces most
prone to attack ; the etch-figures then disappear ; the entire surface is then affected so
that the poles of the faces most prone to attack become most sharply pronounced
corners of curved faces. Consequently, growth and dissolution are inversely related.
This relation is called V. Goldschmidt and F. E. Wright's law o! polarity :
the corners of a dissolving crystal hecome the poles of the faces of a growing crystal.
612 INORGANIC AND THEORETICAL CHEMISTRY
The inversion corners shift slightly as the crystal dissolves, until at last a final form
is reached which dissolves without change of shape. This final form depends on
the original form of a natural crystal and on the nature and concentration of the
acid. P. Hochschild obtained similar results with zinc blende ; and A. E. Fersmann
and V. Goldschmidt have shown the relation between these results and the corrosion
forms shown by natural crystals of the diamond. W. Schnorr also studied the
alteration of form brought about by a reversal of the growth process of crystals of
rock salt by the solvent action of slightly unsaturated solutions of sodium chloride
with a little carbamide. The first action is to bevel the cube edges— the final form
is the icositetrahedron. The final form obeys V. Goldschmidt and F. E. Wright's
law of polarity.
The essential difference between crystalline and amorphous substances is
one of internal structure, not necessarily external shape. — The external form
of crystals is their most obtrusive characteristic, and it was naturally the
first to arrest attention ; but the geometrical shape is by no means the most
characteristic property of crystals, because the external geometrical form may
be destroyed, and yet the fragments do not cease to be crystals, for they
behave in polarized light like perfect crystals. On the contrary, the most
perfect glass model of a crystal is not a crystal, because it lacks the characteristic
internal properties of crystals. According to L. Vegard,^ crystals of thorite show
nearly perfect tetragonal forms and internally they appear to be amorphous. The
shapes of gems cut and polished to accentuate the ornamental value of the gem
must not be confounded with crystal structure ; similarly, the term *' crystal "
applied to cut glass has a different meaning from the special use of the word crystal
in the text. Transparent glass is not crystalline ; some varieties of opaque glass
are microcrystalline. In the case of granite, the crystals of felspar, quartz, and
mica have been so crowded during their growth that they have had no chance to
develop their characteristic external shape. The internal structure of each mineral,
however, is characteristic. A crystal has therefore been defined as " a solid body
bounded by plane surfaces arranged according to definite laws, and possessed of
definite physical properties. Both the external form and the physical properties
result from a definite and unique internal structure." Amorphous substances
show no signs of the definite structure characteristic of crystals. The term
" amorphous " is applied, somewhat vaguely — often wrongly — to the pulverulent
substances, i.e. to substances occurring as fine-grained powders particularly when
the powders have not the definite external shape characteristic of crystals, or when
the grains are opaque and do not permit the application of the usual optical tests
to find if they have the internal structure characteristic of crystals. P. P. von
Weimarn has raised the question whether the finest precipitates are ever amorphous ;
he is right in saying that many precipitates usually classed as amorphous
are really crystalline, but there is no doubt that many precipitates are really
analogous with supercooled liquids, like glass. P. P. von Weimarn, however,
says that a *' super-cooled glass" is crystalline, and he even says that liquids and
gases are crystalline ; this makes it obvious that his definition of a crystalline
substance is different from that usually employed, and need not here be considered.
An amorphous substance is one which, during solidification, has not taken the
definite external shape characteristic of crystals, the properties when measured
in any one direction are the same as when measured in any other direction,
and there are no signs of a definite orientation of the molecules. In this case
it is assumed that the constituent molecules are arranged haphazardly or in a
chaotic manner. In crystals, on the contrary, where the properties along parallel
directions are the same, but different in directions that are not parallel, it is assumed
that the ultimate molecules, or their motions, are oriented or arranged in a definite
regular manner. W. Voigt (1906) aptly illustrates this idea by the metaphor
cited above.
The words *' haphazard," " chance," and " chaotic," applied to the arrangement
CRYSTALS AND CRYSTALLIZATION 613
of atoms or molecules in amorphous substances, are not intended to imply that there
is such a thing in nature as a " fortuitous concourse of atoms." The man of science
believes, by faith, that the irregular path described by a mote dancing in a beam of
sunlight is determined as certainly as the orbit of the planet about its sun. Words
like these are conventional modes of expressing our ignorance of the great design.
If this be ever discovered, we believe, by faith, that what is now regarded as a
chance coincidence will be part of an everlasting harmony.
Refeeences.
1 J. B. Biot, Ann. Chim. Phys., (3), 36. 257, 1852 ; R. Pribram, Sitzber, Akad. Berlin, 605, 1887 ;
R. von Sonnenthal, Zeit. phys. Chem., 9. 656, 1892 ; G. H. Schneider, Liebig'a Ann., 207. 1881.
2 H. de Senarmont, Ann. Chim. Phys., (3), 21. 466, 1847 ; H. L. Bowman, Min. Mag., 12.
355, 1899 ; C. Pape, Pogg. Ann., 135. 4, 1868.
8 W. Spring, Zeit. phys. Chem., 2. 13, 1888; J. Schiirr, Journ. Chim. Phys., 2. 246, 1904;
C. E. Carbonelli, AM Soc. Lug., 3, 1892.
* A. Korbs, Zeit. Kryst., 43. 433, 1907 ; G. Wulflf, ih., 34. 386, 1901 ; V. Goldschmidt and
F. E. Wright, Jahrh. Min. B. B., 17. 355, 1903 ; 18. 335, 1904 ; P. Hochschild, ih., 26. 178,
1908; A. E. Fersmann and V. Goldschmidt, Der Diamant, Heidelberg, 1911 ; W. Schnorr, Zeit.
Kryst., 54. 289, 1916.
« L. Vegard, Phil. Mag., (6), 32. 93, 1916 ; P. P. von Weimam, Zeit. Koll, 3. 166, 1908 ;
Zur Lehre von den Zustdnde der Materie, Dresden, 1914.
§ 6. The Seven Styles of Crystal Architectuie
In crystallography there is a beautiful instance of successful classification connected
with a nearly perfect physical hypothesis. — W. S. Jevons.
A symmetrical shape is one which consists of parts exactly similar, repeated a certain
number of times, and placed so as to correspond with each other. The symmetrical parts
of a crystal are, under like circumstances, alike affected.- — W. Whewell.
Crystal faces usually occur in sets so arranged as to preserve the symmetry of
the crystal with respect to certain imaginary points, axes, or planes which are
characteristic of certain groups or families of crystals. J. B. L. Kome de I'lsle
expressed the idea in 1783 by stating : " Every crystal face has a similar face
parallel to it," and the symmetry of the faces and angles of crystals has been
emphasized by calling it the law of crystal symmetry : In normally formed
crystals, every face has a similar face in all positions consonant with the
symmetry of the particular class to which the crystal belongs. When
a crystal shows the highest grade of symmetry pertaining to its system — that is,
when a crystal possesses all the faces required by the law of symmetry — *the crystal
is said to be holosymmetrical or holohedral — from oXos, whole ; eSpa, base
or face. An un symmetrical crystal may be derived from a holohedral crystal
by the suppression of half its faces, when it is termed hemihedral, from ^/>tt,
half — the tetrahedron, for instance, is the hemihedral form of the octahedron ;
or by the suppression of three-quarters of its faces, when it is termed tetartohedral
— from T€TapTo?, a quarter.
G. D. Liveing ^ assumes that if a solid be bounded by plane faces, the surface
tensions at the edges will have a resultant which tends to compress the mass ; and,
for equilibrium, there must be an opposing pressure on the opposite side of the
crystal, or else there will be internal stresses. Hence, a reason for the law of crystal
symmetry can be seen, for if one face of the crystal be developed, the opposite face
will also be developed ; and if one edge or angle be truncated, all the corresponding
edges or angles will be truncated ; if otherwise, there would be a stress in the
interior tending to deform the crystals. The surface tension, which produces this
stress, depends on the nature of the surfaces in contact, on their temperature,
electrical condition, etc. If therefore the surface tension on one face be balanced
by inequalities of temperature, etc., unsymmetrical faces may be developed, and
614
INORGANIC AND THEORETICAL CHEMISTRY
when the stress produced by, say, an inequaUty of temperature is relieved, an in-
ternal stress due to unequal surface tensions would persist. Such crystals would
exhibit signs of internal stresses. Crystals with unsymmetrical faces generally
do exhibit signs of internal stresses by developing electrifications of opposite signs
at the two ends when heated or cooled — pyro-electrification—oi they may affect
polarized light differently. In illustration, symmetrical crystals of tourmaline
do not usually exhibit pyro-electrification, while the unsymmetrical crystals do.
Likewise, substances which show rotatory power in solution develop unsymmetrical
crystals — e.g. the tartrates.
A crystal is said to possess a centre of symmetry when to every face of the
crystal there is a corresponding parallel face at the opposite side of the crystal.
A plane of symmetry is an imaginary plane
Ol I /\ which divides the crystal into two parts
•0 / '(X I KJ I such that one part is the exact but inverse
I I /_ \ counterpart of the other. In other words,
A B C D the two parts bear to one another the same
Fig. 21. — Axes of Symmetry. relation that the image in a mirror bears to
its object. The mirror is the equivalent of a
plane of symmetry. A crystal of potassium iodide, for example, has nine planes of
symmetry indicated in Fig. 24. The crystal of gypsum. Fig. 44, has only one plane
of synametry ; and a crystal of barium sulphate has three planes of symmetry.
Fig. 41. Crystals can be classified into groups, according to the disposition and
number of their planes of symmetry.
Then again, a crystal may be rotated about a definite axis through an angle,
which is a simple fraction — J, J, J, or J of the angle of complete rotation, 360° —
such that the faces, edges, and corners are brought into similar or symmetrical
positions, and the aspect of the crystal is the same as before rotation. The axes
of rotation are then called axes of symmetry. Thus we speak of dyad, triad,
tetrad, hexad axes of symmetry according as there are 2, 3, 4, or 6 positions of
symmetry during a complete rotation. These positions
correspond with rotations of 180°, 120°, 90°, and 60°.
Thus, Fig. 21, A, represents a horizontal cross-section
of a crystal with one hexad axis of symmetry, because
during the rotation of the crystal about the axis 0,
there are six positions— 60°, 120°, 180°, 240°, 300°, and
360° — where the original aspect of the crystal is the
same. In Fig. 21, B, C, D, respectively, denote tetrad,
triad, and dyad axes of symmetry. P. H. R. von
Groth (1876) 2 has shown that other grades of sym-
metry— pentad, heptad, octad, etc. — are not possible in
crystals, and J. W. Evans proved that the only possible
axes]^of symmetry are those with cyclic numbers 2,
3,^4, or^6,^ provided that the crystals have a homogeneous
cellular structure. In the subjoined outline discussion
of the crystal systems, the maximum symmetry is alone indicated.
Miller's system of crystal notation. — Under ordinary conditions the earth
is the standard of reference for both position and direction. This is merely
for convenience ; other standards are used in astronomy. In analytical geometry,
position and direction are referred to a set of arbitrary lines called axes. Similarly,
in the geometrical description of a crystal, it is convenient to refer the position of
the faces or bounding planes of the crystal to a set of imaginary coordinate axes
within the crystal, and which are called the crystallographic axes. There
are generally three, sometimes four, of these axes. The comparative length and
mutual inclination of the axes depend upon the symmetry of the ideal crystal.
The axes are chosen so as to furnish the simplest expression to describe the faces
of the crystals, and to allow similar faces to be described by similar terms. In
Fio. 22.
Crystallographic
Constanta.
CRYSTALS AND CRYSTALLIZATION 615
Fig. 22, let AOA'y BOB\ COC represent the three axes of a crystal, and call them
respectively the a-, 6-, and the c-axis. The virtual axis is called the c-axis ; that
passing forwards and backwards is the a-axis, and that passing right and left the
6-axis. The form and nature of the crystal is supposed to be determined by the
length and mutual inclination of these axes. The angles between the axes are
called the axial angles. The axial angle BOG is symbolized a, the axial
angle CO A, by j3; and AOB, by y. The axial angles with the axial ratios
a'.h:c are a characteristic for the crystals of each individual substance, and are
called the crystallographic constants of the crystals. Each of the crystallo-
graphic axes can be drawn from the intersection or origin 0 in two directions ; one
direction is arbitrarily called the positive and the other the negative direction. A
bar over the letters Ay B, C represents conventionally the negative direction, the
opposite directions are called positive. Mark off the equal lengths a and a on the
AOA axis, b and h on the BOB axis, and c and c on the COG axis. Suppose a set
of planes to lie in such a position that the extremities meet the AOA^ axis at a
distance a, the OG axis at c, and OB axis at h, and so on. The lengths a, 6, c and
a, b, c are called parameters. In the diagram, Fig. 22, the plane ABC with the
parameters abc is called the a6c-plane ; the plane ABC' is the a6c-plane ; AB'C
is the a6c-plane, etc.
The faces can be fixed when the directions of the a-, 6-, and c-axis and the
parameters a, 6, c, are known. Suppose a plane to cut the axis a at half the length
of a, 6 at J the length of b, and c at J the length of c, then the position of this plane
about the given axes would be fixed by Ja, J6, \c. Such a notation is considered
clumsy, and the reciprocals of the fractional values alone are used in describing the
plane, which would then be called the 234-plane. Similarly, a 123-plane is one
which passes through points corresponding with a, J6, \c. Similar remarks apply
to the other planes, allowing for the negative values as just indicated ; the plane
abc, for instance, becomes the 111 -plane. Each number is called an index, and
conventionally, the first index always refers to the a-axis, the second to the 6-axis,
and the third to the c-axis. Suppose that a plane corresponded with j^q^ a, yq^ott b, c,
it would be described as the 1000, 1000, 1-plane ; this is true, however small the
fractions be taken. The smaller the parameter, the steeper the plane, until finally
the plane oo «, co 6, c is written 001, is parallel to the axes in question. The cyphers
thus represent planes parallel to the a- and 6-axes respectively. The three indices
of a plane may be multiplied or divided by any desired number without altering
their relations one to another. Thus the plane 222 must be parallel to the Ill-plane,
and also to the 333-plane, etc. Hence, it is usual to reduce the ratios to their
simplest form. This system of notation was adopted by W. H. Miller in a classical
work entitled Treatise on Crystallography (Cambridge, 1829). There are several
other systems in use, but Miller's promises to oust them in the course of time.
R. J. Haiiy's law of rational indices. — It might be supposed that the angles
between the planes of crystals could have any indices, and the planes any inclina-
tion. Observations show that this is not the case. R. J. Haiiy noticed that the
indices can be generally represented by simple whole numbers, but never by what
mathematicians call irrational numbers. 3 Thus the indices of a plane might be 123,
457, etc., but never l\/23, 45\/7, etc. The fact that the indices of all crystal planes
can be expressed by rational whole numbers is called R. J. Haiiy's law of rational
indices. The simplicity of the indices of course is largely determined on the happy
though arbitrary choice of the axial directions AA\ BB\ and CC\ Like most other
so-called laws, the rationality is seldom fulfilled with strict accuracy, but it comes so
near the truth that it is regarded as an outward and visible symbol of the internal
structural simplicity of crystals
Sometimes the symbols hkl, hkl, etc., are used in a general way to express any set of
rational numbers. If the symmetry of a crystal and one face of a crystal form are known,
the other similar faces can be derived from the known face. A group of similar faces is
616 INORGANIC AND THEORETICAL CHEMISTRY
called the form of a crystal, £uad the form of a crystal can be represented by the same symbol
as that used for one of the faces. In that case the symbol for the face is enclosed in brackets,
thus, (hJd) represents all the faces included in the groups of faces similar in every respect to
the face Iikl.
The seven systems of crystal architecture. — The study of the physical pro-
perties and forms of crystals qud crystals is a special branch of chemical physics —
crystallography — and the study of the forms of the different varieties of crystals
is called the morphology of crystals. It is assumed that crystals are built of similar
molecules which are either similarly related to all the adjoining molecules, or else
similarly related to the adjoining molecules which are in the same plane, but
differently related to those in different planes. In the one case the arrangement
of the molecules is rectangular, and in the other case, oblique. All the physical
properties of crystals are closely correlated with the form of the crystal ; and the
form of a crystal is determined by the relative length, and the mutual inclinations
of the crystallographic axes.
It is supposed that the imaginary axes of all except hexagonal crystals can be
varied with respect to length in three ways, for they may be (1) all equal ; (2) all
unequal ; and (3) one may be unequal and two equal. The axes can conceivably
be varied in direction or slope in four ways : (a) All may be at right angles to one
another ; (6) two axes may be at right angles, and the third perpendicular to one
of them and oblique to the other ; or (c) the third axis may be oblique to both ; and
(d) all three axes may be oblique to one another. There is also an additional type
which has three axes lying in one plane and a fourth axis perpendicular to these
three. Every known crystal can be referred, on the basis of its symmetry, to one
of the following seven systems :
System.
( All equal .... Cubic
All rectangular | Two equal, one unequal . Tetragonal (uniaxial)
( All unequal . . . Rhombic (biaxial)
'Three J One perpendicular to two mutually oblique,
all imequal . . . Monoclinic (biaxial)
I Unequal .... Triclinic (biaxial)
Equal (with angles equal but
not 90°) .... Trigonal (uaiaxial)
Four One rectangular, to three oblique and equal . Hexagonal (uniaxial)
These seven systems are further subdivided into classes, each of which has its own
characteristic symmetry — described in standard works — e.g. A. E. H. Tutton's
Crystallogra'phy and Practical Crystal Measurement (London, 1911).
With the exception of the members of the cubic system, the crystals of no two
compounds are exactly alike ; but crystals of the same compound have their faces
inclined at the same angles. Consequently, it is possible to identify crystals quickly
from measurements of the angles between similar faces of one or two crystals by
reference to tables containing measurements of all those crystalline substances
whose angles have been measured. This mode of identifying crystalline substances
is called Fedoroff's crystallochemical analysis^
I. Cubic system. — The first class of crystals possesses three axes — a, h, c — of
equal length, and they make equal angles — a, j3, y — with one another, as illustrated
in Fig. 23. The axes are interchangeable, so also are the angles, so that what is
true of one axis or angle is true also of the other two. This is expressed in symbols :
a=h=c, and a=j8=y. The crystals have nine planes of sj^mmetry as illustrated
in Fig. 24, three of the planes of symmetry are principal planes, and six are secondary.
There are six dyad, three tetrad ; and four triad axes of symmetry. The primary
or simplest representative form is the cube — hence the name cubic system. By
cutting off the corners of the primary cube by planes variously inclined to the axes,
the octahedron, dodecahedron, and various secondary forms are derived. Since
each of the three crystallographic axes is exactly like the other two, every facet
formed on one corner of a crystal must be repeated symmetrically with regard to
Axes
CRYSTALS AND CRYSTALLIZATION
617
the other axes ; hence the forms produced are symmetrical or regular, and in place
of the cubic system, the term regular system is synonymously employed. In addition
to the terms cubic or regular system, this class of crystals has also been called the
isometric, monometric, tesseral, tessural, and octahedral system.
Typical crystals for examination are cuprite and garnet. These cubic crystals
are defined by the numerical values of the indices of the component forms. For
example, cuprite, Fig. 25, has a(lOO), d{110). This particular crystal, described
by A. E. H. Tutton, is apparently holohedral ; but H. E. Miers has shown that
in general cuprite is hemihedral. Garnet, Fig. 26, has d(llO), i{2ll). Potassium
b^
^b
Fig. 23. — Crystallographic Axes Fig. 24. — Planes of Symmetry in the Fig. 25.— Cuprite,
of the Cubic System. Cubic System — 3 Principal and 6
Secondary Planes.
iodide, barium nitrate, chrome alum, potash alum, sodium chlorate, and arsenic
trioxide also furnish good crystals for examination.
Examples.— Diamond ; potassimn chloride ; sodium chloride ; alum ; fluorspar ;
iron pyrites ; lead nitrate ; magnetic oxide of iron ; barium nitrate ; arsenic trioxide ;
galena ; garnet ; ammonium chloroplatinate ; silver chloride ; boracite ; indium ; alu-
minium ; iron ; platinum ; lead ; phosphorus ; gold ; copper ; silver ; nickel ; arsenic ;
metacinnabarite ; cerargyrite ; ammonium chloride ; amalgam- — HgAg ; nitrous oxide ;
carbon dioxide; ammonia; potassium thio-stannate- — K2SnS3.3H20 ; beryllium sulphate
• — BeS04.6H20 ; uranyl sodium acetate ; stannic iodide ; bismuth ; fahlerz ; spinel ;
argentite ; leucite ; franklinite ; nosean ; tetrahedrite ; ten-
nan tite ; allmannite- — 'NiSbS ; barium nitrate ; analcite ;
cobaltite : cuprite ; sodalite ; sodium chlorate ; sodium bro-
mate ; zinc blende ; mercury.
The equality and symmetry of the three axes in the
members of the cubic system is a mathematical expression
of the fact that the vectorial properties of the crystals
belonging to this system are alike in all direction — e.g. the
optical properties. When light or heat rays enter one of
these crystals, the rays spread with equal rapidity in all
directions just as they do in homogeneous gases, liquids,
and unstrained amorphous solids — e.g. glass. The crj^stals
are not doubly refracting unless the elasticity is modi-
fied by compression. Otherwise expressed, the crystals are optically isotropic,
with an index of refraction which is the same in all directions ; there is no change
between crossed nicols ; and there are no interference figures. When the crystals
are heated, they expand equally in all directions. Consequently, the mere state-
ment that a crystal belongs to the cubic system, is a sufficient indication that it
possesses these qualities in common with other members of its class.
n. Hexagonal system, — The system is so named because a horizontal section
is usually hexagonal — from c^aycovta, having six angles or corners. Here the
crystals have four axes — ai, a2, a^, c — of which the three — ai, a2> *3» — Iji^g i^ o^®
plane are of equal length, ai=a2—a^, and meet one another at angles of 60°, Fig. 27 ;
the fourth or c-axis is perpendicular to the a-axis, and is called the principal axis.
Fig. 26.— Garnet.
618
INORGANIC AND THEORETICAL CHEMISTRY
These facts are symbolized ai=a2=«3^c; ai=a2=a3=90°, y=60°. The crystals
have seven planes of symmetry, Fig. 28 ; and one hexad, and maybe six dyad
axes of symmetry. The crystals are uniaxial ; the interference figure is a symmetri-
cal black cross with concentric spectrum coloured rings ; there are two principal
indices of refraction.
Typical crystals for examination are beryl and apatite. The crystals are defined
by the form development and the values of the interfacial angles from which the
axial ratios can be calculated. For instance, apatite has a : c=l : 0*7346 and m(lOlO) ,
\ ' ,>a2
<^
Fio. 27. — Crystallographic Axes
of the Hexagonal System.
Fig. 28. — Seven Planes of Symmetry
of the Hexagonal System.
Fig. 29. -Apatite.
a;(1011); beryl has the angle a:c=l-0499, and a(lOlO), r(1121), c(OOOl). Lead
iodide, and cadmium iodide also, furnish good examples for examination.
Examples." — Beryl ; apatite ; copper sulphide ; lead iodide ; magnesium ; beryllium ;
zinc ; cadmium ; calcium ; pyrrhotite ; proustite ; pyrargyrite ; silver iodide ; strontium
and lead antimonyl tartrates ; pyromorphite ; mimetite ; vanadinite ; iodyrite ; nephelite ; etc.
Owing to the disposition of the axes in the tetragonal, trigonal, and hexagonal
systems, the physical properties are alike in all directions perpendicular to the
principal axes, but are different in other directions. The crystals are not doubly
refracting in the direction of the principal axis, but they are doubly refracting in
\t;
-,P!\
a V-
T T r
Fig. 30.— Beryl.
<b-^
Fig. 31. — Crystallographic Axes
of the Trigonal System.
Fig. 32. — Trigonal Pianos of
Symmetry.
other directions. Thus, a crystal of beryl does not exhibit double refraction in the
direction of the principal axis, but light is doubly refracted in every other direction.
Accordingly, these crystals are uniaxial. Heat is cond'ucted at the same speed in
directions parallel to the principal axis, but with different speeds in directions
perpendicular to this axis.
in. Trigonal system. — The name comes from rpiyoivov^ having three angles or
corners. The crystals of this system have three axes a=h=c, all equal and equally
included at an angle which is not a right angle, so that a=/3=y, Fig. 31 . The crystals
have three planes of symmetry, Fig. 32, one triad, and maybe three dyad axes of
symmetry. The crystallographic axes are not axes of symmetry, but are lines
CKYSTALS AND CKYSTALLIZATION
619
parallel with the edges of the fundamental rhombohedron. This system is some-
times called the rhombohedral system, and it is sometimes regarded as a special
development of the hexagonal system. The crystals are uniaxial ; the interference
figure is a symmetrical black cross with concentric spectrum coloured rings ; there
are two principal indices of refraction.
Typical crystals for examination are calcite and quartz. The crystals are defined
by the form development, and the values of the interfacial angles from which the
axial ratios are calculated. For example, calcite (Figs. 36 and 37) has a=101° 54' ;
and m{2U), rlOO^ ?;(201)^ f?(110) ; quartz (Fig. 3) has a=93° 57'; and m(211),
r(100),r'(122), §(421), a;(4:21). Sodium orthophosphate, Na3P04.12H20, strontium
chloride, SrCl3.6H20, sodium nitrate, NaNOs, furnish examples for examination.
b^
^b
^^
Pig. 33. — Rhombohedral Fig. 34. — Calcite (Dog's Tooth Fio. 35. — Crystallcgraphic Axes
Calcite. Spar). of the Tetragonal System.
ExAMPLES.^ — Sodium periodate— NaI04.3H20 ; quartz ; tourmaline ; antimony ;
bismuth ; calcite ; ice ; graphite ; sodium nitrate ; arsenic ; tellurium ; nickel sulphide
— millerite ; cinnabar; calcium chloride- — ^CaCla-BHaO ; corundum; cadmium carbonate;
bismuth iodide ; ferrous carbonate ; zinc carbonate ; manganese carbonate ; chabazite ;
brucite ; corundimi ; lead, barium, strontium, and calcium dithionates ; calamine ; dolo-
mite ; dioptase — CuSiOg.HoO ; benitoite — BaO.Ti02.3Si02 ; ilmenite ; phenacite; etc.
IV. Tetragonal system. — ^The name comes from rcrpaywvta, having four angles
or corners. The members of this system have three axes, two of which, a and b,
are equal to one another a=h'^c ; the axes intersect at right angles. Fig. 35, so that
a^=p=y=90°. The crystals may have five planes of symmetry, Fig. 36, one tetrad,
and maybe four dyad axes of symmetry. The crystals are uniaxial ;
the interference figure is a symmetrical cross with concentric rings ;
the crystals are isotropic in one position, and the optical extinc-
tion is parallel in the other two ; there are two principal indices
of refraction. This system is sometimes called the pyramidal,
quadratic, or the quaternary dimetric system.
Typical crystals for examination are nickel sulphate, NiS04.6H20,
and cassiterite. The crystals are defined by the form development,
and the values of the interfacial angles from which the axial ratio "^^'g^^^
is calculated. For example, nickel sulphate (Fig. 39) crystallized
from a warm solution has a:c=l : 1'912 ; and o(lll), ic(112), c(OOl),
a.(lOO), r(lOl), 5(203). Cassiterite (Fig. 38) has a : c=l : 0-673, and
wi(llO), a(lOO), r(lOl), s(431). Potassium cupric chloride, mercuric cyanide, and
potassium arsenite also furnish good crystals for examination.
Examples.- — Rutile ; cassiterite ; zircon ; mercurous chloride ; nickel sulphate ; potas-
sium hydrogen phosphate — KH2PO4 ; native lead molybdate or wulfenite — PbMoO^ ;
sodium meta-antimonite — NaSbOj ; potassium hydrogen arsenate- — KHjAsOi ; scheelite ;
tin; strychnine sulphate ; anatase ; lead tungstate— PbWO^ ; stolzite ; yttrium niobate
or fergusonite ; pinnonite ; vesuvianite ; urea ; wernerite ; mercury chloride, iodide, and
cyanide ; barium antimonyl tartrate ; phosgenite ; idocrase ; apophyllite ; scapolite ;
braunite ; etc.
Planes
of Symmetry of
the Tetragonal
System.
V. Rhombic system. — Here the crystals have three unequal axes all inclined
620
INOKGANIC AND THEORETICAL CHEMISTRY
at right angles, so that a=j8=y=90° ; and a<.b'^c, Fig. 39 — h is conventionally
taken as unity. The larger of the two lateral axes is called the macrodiagonal-—
fxaKpos, long — and the smaller the brachydiagonal — ^paxvs, short. The crystals
may have three planes of symmetry, Fig. 40 ; and three dyad axes of symmetry.
The crystals are biaxial ; optical extinction is parallel in all three main positions
of the crystal ; and there are three principal indices of refraction — the smallest
index is in the direction of greatest elasticity, and vice versa. This system is
sometimes called the ortJwrhombic, trimetric, or the frismaiic system.
Typical crystals for examination are barytes (Fig. 41) and topaz (Fig, 42).
The crystals are defined by the development forms and the values of the inter-
Pro. 37. — Cassiterite.
Fig. 38.— Nickel Sulphate
Hexahydrate.
•"♦J^^^^s?
Fig. 39. — Crystallographic Axes
of the Rhombic System.
facial angles from which the axial ratios can be calculated. Thus, barytes has
a:h: c=0-815 : 1 : 1-314 ; andc(OOl), o(Oll), and d(\02). Topaz has « : 6 : c=0-529 :
1 : 0-954; and c(OOl), m(llO), w(140), t(223), /(120), /(021), 2/(041), u(ni). Am-
monium sulphate, potassium nitrate, potassium sulphate, zinc sulphate, sodium
nitroprusside, and mercuric chloride also furnish good crystals for examination.
Examples." — Zinc sulphate- — ZnSOj.THaO ; magnesium sulphate— MgSOi.THaO ;
ammonium magnesium phosphate — NH4MgP04.6H20 ; potassium sulphate ; aragonite ;
anhydrous sodium or silver sulphate ; sulphur from solution ; barium, strontium, and
ammonium sulphates ; sodium arsenate ; sodium phosphate — NaH2P04H20 ; iodine ;
potassium nitrate ; tartar emetic ; potassium perchlorate ; potassium permanganate ;
Fig. 40. — Planes of Symmetry
in the Rhombic System.
Fig. 41. — Barytes.
Fig. 42. — One end of a Topaz
Crystal.
topaz ; marcasite ; tridymite ; silver nitrate ; lead carbonate ; silver sulphide ; prehnite ;
calamine ; atacamite ; goslarite ; stephanite ; chrysoberyl ; topaz ; andalusite ; chalcocite ;
acanthite ; hypersthene ; struvite ; tartaric acid ; manganese peroxide ; barium chloride ;
mercuric chloride ; orpiment ; antimonic oxide ; ammonium nitrate ; Rochelle salt ; citric
acid ; iodine ; selenium ; olivine ; cerussite ; strontianite ; redruthite ; bournonite ; hemi-
morphito ; stibnite ; etc.
The physical properties of members of the triclinic, monoclinic, and rhombic
systems vary in all three directions ; for example, heat is conducted at different rates
in all three directions ; again, in the mineral iolite, Al(F20H)Si04, crystallizing in the
rhombic system, light transmitted in the direction of the principal axes often appears
blue, greyish-blue when viewed through the 100 face, and yellow through the 010
CRYSTALS AND CRYSTALLIZATION
621
face. This phenomenon is known as pleochroism. Light is singly refracted in two
directions, and doubly refracted in all other directions, hence the crystals are
optically biaxial.
VI. Monoclinic system. — The name is derived from ju-wo?, one, and k\lv€lv, to
incline — having one oblique intersection in allusion to the fact that the members
of this system can be referred to three unequal axes, of which two, c (the vertical
axis) and a, are inclined to form one oblique angle p, and third lateral axis, h, is
at right angles — to the other two — Fig. 43. The inclined lateral or a-axis is called
the clino-axis or clino-diagonal — KXtVeiv, to incline — and the rectangular lateral
or &-axis is called the ortho-axis or the ortho-diagonal — 6p06<;, straight. Hence,
a=y=90°, and j8^90°, and a^6^c, and there are no closed symmetrical forms,
and accordingly, the crystal must be a combination of different forms. These
crystals may have one plane of symmetry. Fig. 46, and there may be one dyad axis
of symmetry. The crystals are biaxial ; optical extinction is parallel in two positions
and oblique in the third ; and there are three principal indices of refraction. This
system has also been styled the mono-symmetric, clino-rhomhic, or the oblique system.
Typical examples for examination are gypsum (Fig. 44) and ammonium mag-
nesium sulphate (Fig. 45) — the crystals by the development forms, and the values
of the interf acial angles from which the axial ratios can be calculated. Thus gypsum
(Fig. 44) has a : 6 : c=0-690 : 1 : 0-412 ; j8=80° 42' ; and /(111), m(llO), 6(010).
h^
r^b
N/
Fig. 43. — CrystaUographic Axes
of the Monoclinic System.
Fig. 44.' — Gypsum.
Fig. 45. — Ammonium Magne-
sium Sulphate.
Ammonium magnesium sulphate (NH4)2S04.MgS06.6H20 (Fig. 45), has a:b:c
=0-740 : 1 : 0-492=107° 6'; and 6(010), c(OOl), :p(110), /(130), ^(011), /(201),
o(lll), o'(lll), n{121). Potassium chlorate, potassium ferrocyanide, potassium
chloride, barium chloride, and nickel chloride also furnish good crystals for examina-
tion.
Examples.' — ^Borax— Na2B4Oj.l0H2O ; gypsum ; ferrous sulphate- — FeS04.7H20 ;
potassium and sodium carbonates ; felspar- — orthoclase ; sodiimi sulphate-^ — ^Na2SO4.10H2O ;
ammonium magnesium sulphate- — (NH4)2S04.MgS04.6H20 ; potassium chlorate ; potassium
tetrathionate- — K2S40g ; tartaric acid ; sulphur^ — from fusion ; potassium sulphate ; cane
sugar ; arsenic disulphide- — realgar ; ammonium dichromate ; rubidium magnesium sul-
phate ; acid mercuric fluoride ; clinohedrite ; scolecite ; lead chromate ; oxalic acid ;
sodium acetate ; augite ; homblenVle ; vivianite ; epidote ; etc.
Vn. Triclinic system. — Crystals of this system can be referred to three oblique
axes. Fig. 46. There are no closed symmetrical forms, and hence each crystal must
be a combination of different forms. The crystals have no axes nor planes of
symmetry, but they may have a centre of symmetry ; a^h^c, and a^jS^y.
The crystals are biaxial ; the optical extinction is oblique in all three positions ;
and there are three indices of refraction. This system has also been designated
the anorthic, clino-rhomboidal, asymmetric, or the double oblique.
Typical examples for examination are crystals of copper sulphate (Fig.
48) and of potassium dichromate (Fig. 47). The crystals are defined by the
622
INORGANIC AND THEORETICAL CHEMISTRY
development forms, and the values of the interfacial angles from which the axial
ratios can be computed. For instance, potassium dichromate (Fig. 47) has a:h :c
=0-558 : 1 : 0-551 ; a : ^3 : y=82° 0' : 90° 51' : 83° 47' ; a(lOO), 6(010), c(OOl), ^(011),
p(nO); cojyper sulphate, CUSO4.5H2O (Fig. 49), has a : 6 : 0=05715 : 1 : 05575 ;
a:^y=82°16'; 107° 26'; 102° 40'; and ^(011), ^'(011), «(021), «'(02]), 5(121),
5'(121), o(lll). Potassium persulphate, manganese sulphate, and boric acid furnish
good crystals for examination.
Examples.- — Potassimn dichromate ; copper sulphate — CUSO4.5H2O ; calcium thio-
sulphate — CaSgOg.eHjO ; boric acid ; potassiimi ferricyanide ; anhydrous manganese
sulphate ; copper selenate ; anorthite — lime felspar ; cryolite ; chromic phosphate —
CrPOj.GHjO ; labradorite ; chalcanthite ; rhodonite ; albite ; oligoclase ; axinite ;
bismuth nitrate ; etc.
The relation between crystal form and molecular complexity. — J. W. Retgers
(1894) noticed a general relation between the molecular complexity of over nine
Fig. 46. — Crystallographic Axes
of the Trjclinic System.
Fig 4 47. — Potassium Bi-
chromate.
Fig. 48. — Copper Sulphate
Pentahydrate.
hundred substances and their crystalline form. Table I shows the approximate
percentage distribution of compounds of different molecular complexity in the
different crystal systems. The hexagonal system here includes the trigonal and
hexagonal systems.
Table I.- — Proportion of Substances of Different Molecular Complexity
Crystallizing in the Different Systems.
Per cent.
CJomplexity of
Number of
substances
considered.
molecules.
Cubic.
Hexagonal.
Tetragonal.
Rhombic.
Monoclinic.
6
Triclinic.
Elements
40
50
35
1
5 1 5
0
2 -atoms
67
68-5
19-5
4-5
3-0
4-5
0
3 -atoms
63
42
11
19
23-5
3
1-5
4-atoms
20
6
36
5
50
5
0
S-atoms
50
12
38
6
36
6
2
6 -atoms or more
673
5-8
4.6
7 0
27-3
37-3
8
The results indicated in Table II can be expressed differently by including the
cubic and hexagonal systems in one group, and the remaining systems in another
group . From this it follows that substances with complex molecules are more inclined
to crystallize in systems with a low order of symmetry ; and the simpler the mole-
cules of a substance, the greater the probability of its crystallizing in the hexagonal
CRYSTALS AND CRYSTALLIZATION
623
or cubic systems with a high order of symmetry. G. Tschermak ^ has shown that
when 2j 3, or 6 atoms are present in the molecular formula, the compound
usually crystallizes in the rhombohedral or hexagonal system, and the crystals
are characterized by axes of two-, three-, or six-fold symmetry — e.g. Fe203,
FeCls, AgsSbSs, PI3, SrCl2.6H20, etc. — ^those compounds which have the number
Table II.. — Proportion of Substances or Different Molecular Complexity
Crystallizing in the Cubic and Hexagonal Systems.
Crystal ssrstem.
Elements
(per cent.).
Atoms per molecule.
2
3
4
5
6 and over.
Cubic and hexagonal .
Other systems .
85
15
88 53
12 47
40
60
60
60
20-4
79-6
4, but not 6, in their molecular formulae, usually crystallize in the tetragonal system,
and they possess a two- or four-fold, but not a six-fold, symmetry — e.g. ZrSi04.4H20,
etc. Compounds in which the numbers 3 and 4 occur in the molecular formidae,
have a tendency to crystallize in the cubic system the axes of which possess a three-
and four-fold symmetry — e.g. 2KP.ZrF4, Ag3P04, AS4O6, etc.
References.
1 G. D. Liveing, Proc. Roy. Inst., 13. 375, 1891.
2 J. W. Evans, Min. Mag., 18. 324, 1919 ; P. von Groth, Physikalische Krystallogra'phie,
Leipzig, 1876.
3 T. V. Barker, Ann. Eept. Chem. Progress, 12. 256, 1916.
* T. V. Barker, Chem. News, 106. 199, 1912.
6 G. Tschermak, Tschermak's Mitt., 22. 393, 1903.
§ 7. The Growth of Crystals
The very molecules appear inspired with a desire for union and growth.^ — J. Tyndall.
We do not understand the phenomenon of crystallization, nor do we know how
crystals grow. The facts indicated in the preceding sections have made it almost
certain that the space occupied by a crystal is not all matter ; that their structure
is discontinuous ; and that crystals grow by accretion, molecule by molecule, like
bricks in the hands of the builder, and in accord with " an architectural plan more
elaborate and exact than that of any human architect." We are quite ignorant of
the shape of the structural units. C. Huyghens saw that the regularity of crystals
depends on the arrangement of the smallest particles from which they are built, and
he assumed the structural units of calcspar to be rotational ellipsoids. For
convenience, they are usually taken to be spherical, as Robert Hooke did in his
Micrographia (London, 1665), when he said :
All these regular figures that are so conspicuously various and curious, arise only from
three or four several positions or postures of globular particles. . . . And this I have
ad oculum demonstrated with a company of bullets and some few other very simple bodies ;
so that there was not any regular figure which I have met withal of any of these bodies
and I could not with the composition of bullets or globules . . . imitate even almost by
shaking them together.
C. F. G. H. Westfeld (1767) and T. Bergmann (1773),i in his VaricB cristallorum
formce a spato ortcB, also held the view that many crystal forms could be obtained by
laying together little rhombohedra ; and in 1772, J. B. L. Rome de I'lsle pointed
out that the various shapes of natural crystals can be derived from a fundamental
624 * INORGANIC AND THEORETICAL CHEMISTRY
figure or 'primitive form, and that the variety of form which is found in natural
crystals is due to the variety of the secondary faces. About 1801, R. J. Haiiy 2
developed the idea, that all crystal forms, other than the primary ones, could be
exactly imitated by building on the faces of the primary forms, successive layers
of what he called integrant molecules — molecules integrantes — each successive layer
being regularly diminished by the abstraction of one or more rows either parallel
to each edge, or to the diagonals of the faces of the primitive form or in some other
way. R. J. Haiiy's integrant molecules were of three shapes-- the parallelopiped,
tetrahedron, and the trigonal prism. R. J. Haiiy deduced the shapes of his crystal
units from the shapes of the cleavage fragments. W. H. Wollaston,^ like R. Hooke,
suggested the presence of cleavage molecules in crystals, but he also remarked that,
in place of spheres, mathematical points endowed with forces of attraction and
repulsion can be postulated with equal success. In 1831, J. F. C. Hessel * showed
in his book, Krystallometrie oder Krystallonomie und Krystallographie (Leipzig, 1831),
that only thirty-two types of symmetry are possible with a solid bounded by plane
faces conforming to R. J. Haiiy's law of rational indices. At that time compara-
tively few of these thirty-two types were known. In 1865, A. Gadolin independently
made the same discovery as J. F. C. Hessel, and during the next three years V. von
Lang established the external geometry, so to speak, or the symmetry of the thirty-
two classes of crystals.
Speculations were gradually diverted away from the external form of the crystals,
and attention focused more on to the internal orientation of the centres of the
particles ; L. A. Seeber (1824) and G. Delafosse (1843) ^ regarded the crystal structure
as a kind of network of molecular points repeated as identical units throughout
space without regard to their shape or constitution, and thus arose the idea that the
arrangement of the middle points of the structural units resembles a parallelopipedal
network or space-lattice— German, Raumgitter ; French, reseau. Thus, said
G. Delafosse (1843) :
The molecules of the crystal must be situated in a uniform and symmetrical manner,
having their centres of gravity at the points of intersection of a series of parallel planes,
and thus present the picture of a lattice with parallel-figured meshes.
The idea of a space-lattice can be gathered from Figs. 49 to 68 ; it has been
likened to a kind of three dimensional net, in which the particles are situated at the
corners of parallel and equal parallelopipeda, so that the strings of the net represent
the lines of intersection of the planes, and the knots, nodes, or points, their points
of intersection. M. L. Frankenheim ^ examined the different kinds of networks of
points and compared them with the various types of symmetry presented by crystals;
and M. L. Frankenheim's geometrical investigation was supplemented with elegant
proofs by A. Bravais 7 in 1848. Whatever be the actual size and shape of the struc-
tural unit, it is equally certain that all phenomena peculiar to crystals depend upon
the structure or upon the orientation of the constituent molecuiles of the crystals.
The possibility of assigning imaginary crystallographic axes to all known crystals
shows that the molecules must be so related to one another that (1) the structural
imits or molecules of crystals of the same chemical substance, under similar con-
ditions, must be alike in size, and in the distribution of their attractive forces ; and
(2) the relative position of any one molecule must be symmetrical with that of
every other molecule. No other arrangement can be regarded as possible in a
crystal. Hence the study of crystal structure, the orientation of the structural
units, is reduced to the purely geometrical investigation of the possible arrangements
in space of networks of structural units which satisfy these conditions.
Among the problems concerned in elucidating the structure of crystals are :
(1) What is the nature of the structural units ? (2) How are these units arranged
or oriented in space ? and (3) What is the nature and character of the vectorial or
directed forces which fix the relative positions and determine the orientation of the
structural units which make up the crystals ? Great success has been attained with
CRYSTALS AND CRYSTALLIZATION
625
the second problem, which has been studied geometrically quite independently of
the nature of the material. Geometricians have investigated : the nature of the
symmetrical arrangement of particles in space which will confer on matter the
symmetry shown by crystals. As a result, the theory of the homogeneous partition-
ing of space, i.e. of the homogeneous repetition of identical parts in a uniform
structure, has probably reached its final form.
The history of the development of this theory is interesting because it shows
how the testing of the deductions of the mathematician's definitions of homogeneity
by comparison with the morphological properties of crystals from time to time,
compelled the mathematicians to make the definition of homogeneity wider and
wider until it included all known forms of crystals. A. Bravais, for instance, based
Fig. 49.— Simple
Cubic Lattice.
Fig. 50.— Body-
centred Cubic
Lattice.
Fig. 5L— Face-
centred Cubic
Lattice.
his geometrical treatment of the homogeneous partitioning of space on the assump-
tion that if the properties of crystals depend upon the nature and arrangement of
the crystal units, their properties are alike in parallel directions, and that fourteen
types of space-lattice are possible with a symmetry, corresponding with the maxi-
mum symmetry of one or other of the seven systems of crystal architecture.
The fourteen space-lattices are illustrated in Figs. 49-68, where the dots are sup-
posed to represent the positions of the centre of gravity of the atoms. The lattices
belonging to the cubic system are indicated in Figs. 49 to 51. There is : I. The
simple cubic lattice, Fig. 49, with a particle at each corner of a cube. If the atoms
are all nearly equal in size, the symmetry is apparently cubic holohedral — e.g. sodium
chloride, NaCl ; galena, PbS, etc. If two sizes of atoms are present, the symmetry is
plagihedral — e.g. sylvine, KCl ; potassium bromide, KBr ; etc. II. The body-centred
cubic lattice. Here the cube has a particle at each corner
and one at its centre. The symmetry when the atoms
are all alike is cubic holohedral. Fig. 50, e.g. iron, nickel
(in part), and sodium. III. The face-centred cubic lattice.
Here the cube has a particle at each of its corners and one
at the centre of each of its faces. The symmetry is cubic
holohedral when the atoms are all alike. Fig. 51, e.g.
copper, silver, gold, lead, aluminium, nickel (in part), etc.
The space lattice may consist of two or more mutually
penetrating lattices. For example, two interpenetrating
face-centred cubic lattices furnish the double face-cerUred
lattice of the diamond and zinc blende types illustrated in
Fig. 52. The atoms may be all alike when the symmetry
is cubic-tetrahedral in some respects, but holohedral by compensation — e.g. diamond,
silicon, etc. If the atoms are all alike and more than quadrivalent, the symmetry is
hexagonal — trigonal, or rhombohedral — e.g. antimony, bismuth, arsenic, tellurium.
If two kinds of atoms are present, the symmetry will be cubic tetrahedral — e.g. spha-
lerite, ZnS. If three or more kinds of atoms are present, and the atoms are numbered
1, 2, 3, then reading from the top layer downwards, 1, 2, 3, 2, 1. For instance, in
chalcopyrite with Cu, 1 ; S, 2 ; and Fe, 3, alternate layers taken vertically consist
exclusively of each kind in turn. The symmetry is scalenohedral — e.g. chalcopyrite,
CuFeS2 ; stannite, Cu2FeSnS4. If the atoms be supposed in contact, the shapes of
the spaces available for them are illustrated in Figs. 53 to 56, when the constituent
atoms are all equal in size. Atoms of one kind may form a simple cubic lattice
with an interpenetrating symmetrical face-centred cubic lattice containing atoms of
another kind. The symmetry is cubic-holohedral — e.g. calcium fluoride, CaFg. In
VOL. I. - 2 s
Fig. 52.— Double Face-
centred Cubic Lattice.
626
INORGANIC AND THEORETICAL CHEMISTRY
hauerite, MnS2, and pyrite, FeS2, ^^^ symmetry is cubic-pyritohedral, and the space-
lattice is similar to the former, excepting that the simple lattice is distorted. In
cobaltite, CoSAs, the atoms on the distorted simple cubic lattice are of two kinds,
J"- ^
Fig. 53. — Atoms
of Simple Cubic
Lattice.
Fig. 54. — Atoms
of Body -centred
Cubic Lattice.
Fig. 55. — Atoms
of Face-centred
Cubic Lattice.
Fio. 56.- — Atoms
of Double Face-
centred Cubic
Lattice.
and the symmetry is cubic tetartohedral. In cuprite, atoms of copper are on a
body-centred cubic lattice, and the atoms of oxygen are on an interpenetrating
distorted face-centred lattice ; and the symmetry is cubic-gyrohedral.
Fig. 57. — Square
Prism Lattice.
^
!
- — ■*-.
Fig. 58.— Body-
centred Square
Prism Lattice.
Fig. 59.-120°
Prism Lat-
tice.
Fig. 60.— Rhom-
bohedron Lat-
tice.
The tetragonal system is represented by IV. The square prism lattice; Fig. 57 ;
and by V. TJie hody-centred square prism lattice, Fig. 58. The hexagonal and trigonal
or rhombohedral system is represented by VI. The 120° square prism lattice, Fig. 59 ;
VII. The rhombohedron lattice, Fig. 60.
A]^^
1"^-J
-A
■i.
Fig. 61. — Hexagonal
Prism.
Fig. 62.— The Rhombic
Lattice.
Fig. 63.— The Body- Fig. 64.— The Rect-
centred Rhombic angular Prism
Prism Lattice. Lattice — Rhombic.
The hexagonal prism, Fig. 61, may be regarded as a combination of three 120°
prisms, each with a structure like the sixth lattice.
The rhombic system is represented by VIII. The rhombic prism lattice. Fig. 62 ;
IX. The body-centred rhombic prism lattice, Fig. 63 ; X. The rectangular prism lattice,
>-. -
Fig. 65.— The Body-
centred Prism
Lattice — Rhombic.
Fig. 66.— The Mono-
clinic Prism
Lattice.
Fig. 67.— The Mono-
clinic Parallelo-
piped Lattice.
Fig. 68.— The Triclinic
Prism Lattice.
Fig. 64 ; and XI. The body -centred rectangular prism lattice. Fig. 65. The monoclinic
system is represented by XII. The clinorhombic prism lattice. Fig. 66 ; and XIII.
The monoclinic parallelopiped lattice, Fig. 67. The triclinic system is represented
by XIV. The triclinic prism lattice, Fig. 68.
CRYSTALS AND CRYSTALLIZATION 627
In 1869, C. Wiener 8 and C. Jordan extended A. Bravais' assumption by laying
down the principle that " regularity in the arrangement of identical atoms is pre-
sented when every atom has the remaining atoms arranged about it in the same
manner ; thus making homogeneity depend primarily on the continued repetition
through space of the same relation between a unit and the entire structure as un-
limited, instead of laying stress on orientation." L. Sohncke, in his Entwicklung
einer Theorie der Krystallstructur (Leipzig, 1879), followed up the subject, and
assumed that in a homogeneous structure the arrangement about any one point is
the same as that about every other point, so that the aspect from any one point
is the same as that viewed from any other point. He fulfilled this condition, by
selecting one of A. Bravais' fourteen lattices, with one or more lattices identical
with it, and thrusting the lattices into one another so as to satisfy the condition for
homogeneity. This furnished L. Sohncke with 65 systems of points in place of
A. Bravais' 14 ; and these fall into the 32 classes into which observation has classified
all crystals known to exist. In 1887, L. Wulff and F. Haag showed that there
is no provision in L. Sohncke's system for the crystals of the rare mineral dioptase,
nor for the polar or hemimorphic forms of tartaric acid. L. Sohncke then modified
his early theory, which assumed that the units were all of one kind and identically
related to the structure as a whole. A crystal consists of a finite number of inter-
penetrating, regular systems of points ; each separate point system is occupied hy similar
material particles, which may be the saine or different for the interpenetrating partial
systems ivhich form the complex system. Each partial system of units taken by
itself is homogeneously arranged, and all the different units are supposed to possess
identical systems of axes, and to have the same set of translations common to them ;
but L. Sohncke did not stick to the purely geometrical problem, for he introduced
hypotheses as to the physical character of the structural units, which he said can be
rejected " only if they are held to be improbable."
In 1884, P. Curie ^ drew attention to the fact that crystals display not only
identity of parts, but they also exhibit enantiomorphous similarity, and he deduced
L. Sohncke's 32 varieties of external form, but did not pursue the subject further.
Between 1891 and 1894, E. S. vonFederoff,iOA. Schonflies, and W. Barlow, working
by three different methods, and in the order named, showed that L. Sohncke's
definition provided for each unit bbing surrounded by all the other molecules in
like manner where like referred only to identity and not to the resemblance which
obtains between an object and its image in a mirror. In L. Sohncke's system two
identical sets are superposable ; in the extended theory, they are not necessarily
superposable. H. A. Miers n illustrated this by showing that a point in the centre
of a right-hand glove has precisely the same environment as a point in the centre of
a left-hand glove, and yet the two are not superposable, for the one is as it were
the reflection of the other. A. Schonflies admits the principle of reflection across
a plane, inversion about a centre, or a combination of the two as an additional mode
of repetition applicable to a system without changing its aspect. There are 230
possible ways of partitioning space into systems of points as types of homogeneous
structures ; and of these, Sohncke's 65 point-systems and A. Bravais' 14 space-
lattices are special cases. Each of these 230 types of structure can be referred to
one of the 32 classes into which all known crystals can be arranged, and these 32
classes can still further be grouped into the seven-systems of crystal architecture.
" With the establishment of these 230 types of crystal structure," says A. E. H.
Tutton, "the geometrical theory of crystal structure. has attained what in all
probability will prove to be finality." However, directly the space units or cells
are invested with special shapes — the parallelohedra of E. von Federoff and Lord
Kelvin, the Fundamentalbereich of A. Schonflies, or the spheres of influence of
W. Barlow — complications are introduced, and controversial questions arise.
The birth of crystals. — As previously indicated, G. Quincke 12 assumed that the
first stage in the crystallization of a liquid involves the separation of the solution
into two immiscible liquids, one of which is formed in a relatively small quantity,
628 INORGANIC AND THEORETICAL CHEMISTRY
so as to form a kind of emulsion ; and H. F. Link (1839) thought that he could
detect such globules at the moment of separation by using a magnification of 600
diameters. If the nuclear masses of the separating liquid are isotropic,with their
physical characters alike in all directions, the nuclei would form spherical globules.
The spherical globules would afterwards either (i) solidify and serve as nuclei for
the subsequent growth of crystalline particles, or (ii) they would form particles in
the act of solidification. H. Vogelsang (1875) inclined to the former hypothesis
because he and C. Brame (1853) had observed what they considered embryonic
sulphur crystals to separate as globular solids from solutions of sulphur in viscous
solvents. H. Vogelsang called these embryonic crystals glohuUtes.
The term crystallite is applied in several different ways- — e.g. it has been employed for
the structural units of crystals ; for abnormally elongated and branched forms- — crystal
skeletons- — in which the normal faces and angles are not developed ; and for abnormally
developed crystal nuclei which have received various names according to their shape or
appearance — glohuUtes, longulites, margarites, belonites, cumulitea, etc.
The belief that small isotropic globulites first appear as nuclei during crystalliza-
tion, and that only after these globulites have attained a certain size do they assume
the crystalline state, has led W. Ostwald, G. Quincke, and others to believe that in the
development of crystals, the crystal embryos are at first in the state of droplets of
undercooled liquid, and only later become solid as they enlarge into crystals. This
means that the molecules of the primary crystals are not vectorially oriented, a
statement which has not yet been demonstrated.
H. Vogelsang's hypothesis is discredited, because C. Brauns (1899) showed that
what are thought to be solid globules are in reality minute globular masses of under-
cooled solutions of high viscosity, as E. Weiss (1871) had previously supposed. If
the second alternative be true the particles must be sub-microscopic, because T. W.
Richards and E. H. Archibald (1901) found that instantaneous photographs of
crystallizing barium chloride and potassium iodide showed that crystals have a
definite character from the moment they are able to affect a photographic plate.
In his great work On the equilibrium of heterogeneous substances (1878), J. W.
Gibbs showed that the stable form of a crystal is that for which, as a result of
capillary forces, the total surface energy is a minimum. The same conclusion was
drawn by P. Curie in 1885 and called Curie's capillarity theory. Each crystal face
has a specific capillary constant which is measured by the work involved in increasing
its surface face by unit area. Let the areas of the various faces be denoted by
^i> ^2j *3> • • •> ^-nd the respective capillary constants by ctj, a^y 0-3, . . ., then, adds
J. W. Gibbs,i3
On the whole it seems not improbable that the form of very minute crystals in equili-
brium with solvents is principally determined by the condition that E{(T^8^-^C28^-\-cr^8^
+ . . .) shall be a minimum for the volume of the crystal — except so far as the case is modified
by gravity or the contact of other bodies- — but as they grow (in a solvent no more super-
saturated than is necessary to make them grow at all), the deposition of new matter on the
different faces will be determined more by the nature (orientation) of the surfaces and less
by their size and relations to the surrounding surfaces. As a result, a large crystal thus
formed will generally be boimded by those surfaces alone on which the deposit of new matter
takes place least readily, with small, perhaps insensible truncations.
Crystals take the habit which gives them the minimum surface energy, so that
the relative areas of the faces depend on their capillary constants. While the
principle probably operates with microscopic crystals it does not seem to be valid
for large crystals. Droplets of liquid usually assume a spherical shape, correspond-
ing with a minimum surface area per unit volume. A few crystals bounded wholly
or partially by curved faces are known, and others are so richly faceted that they
approximate to a spherical form. The most frequent styles of development are then
plates or fine needles, shapes which approach a maximum area per unit volume.
P. Curie assumed that plane faces are developed in preference to curved faces because
the capillary constants of the former are the lower.
CRYSTALS AND CRYSTALLIZATION 629
T. V. Barker has pointed out that if J. W. Gibbs' theorem be valid, a knowledge
of the capillary constants would enable a prediction to be made of the form of a
crystal. For example, in the case of a cubic crystal of common salt, in which both
octahedral and cubic faces are observed, the crystal could develop only cubic or
octahedron faces according as s(l(X)) : ^(111) is less than 1 : \/3, or greater than
-y/S : 1. S. Berent measured the capillary constant for water and a solution of car-
bamide where in the one case cubic faces are developed and in the other octahedron
faces. The results demonstrated the existence of capillary differences, but they
were unfavourably criticized by F. Pockels. Another deduction from the theory
is that large crystals in favourable circumstances must grow at the expense of small
ones, and this was verified by the work of G. A. Hulett on very minute crystals.
The theory of J. W. Gibbs also leads to the assumption that different faces have
different solubilities. For a crystal departing from the equilibrium shape when
placed in a solution of suitable strength should dissolve from some faces while others
grow. M. le Blanc and G. Elissaf off showed that the only satisfactory way of test-
ing the deduction is to find if a solution can be obtained of such a concentration
that one face grows while another dissolves ; and J. J. P. Valeton proved that it
was not possible to obtain conditions at which the cubic or dodecahedron faces of
alum dissolve while octahedron faces grow, but he did obtain a solution of such
concentration that a variation of temperature even so small as 0*003° sufficed to
transform unmistakable growth into unmistakable solution. . G. Wulff also measured
the relative velocities of growth of the crystal faces of monoclinic Mohr's salt reckoned
from the centre of the crystal. He found the relative rates of growth to be (110),
1-96; (001), 2-25; (111), 2-50; (111), 2-64 ; (Oil), 2-77, when the rate for the (201)
face was taken as unity. S. ToUoczko also found the velocity of solution of gypsum
on the (010), (110), and (111) faces were respectively 1, 1*76, and 1-88 ; and A. Ritzel
found that water dissolved the octahedron faces of a crystal of sodium chloride
faster than the cube faces, but with a dilute solution of carbamide this relation was
reversed. G. Wulff and H. Liebmann also argued that the relative velocities of
growth of the crystal faces are proportional to the capillary constants, but H. Hilton,
G. Friedel, and C. Fastert have shown that the conclusion is faulty ; and A. Ber-
thoud has shown that the differences in the solubilities of the different faces of a
crystal are so minute in comparison with the difference in the rates along different
axes as to be without influence on the crystalline habit.
L. Sohncke (1888) also attempted to establish a relation between the surface
energy of a crystal face and what he called its face-density, meaning by that the
number of mass-points (crystal units) existing in unit area of a crystal face. He
showed that in a face of maximum face-density, the particles can come no nearer
together, and therefore the minimum amount of work remains for the molecular
forces to perform. This means that the surface energy of such a face is a minimum.
Hence, the crystalline face which is most thickly studded with mass-points will
occur most frequently. The density of such points on a face is known as the
reticular density (A. Bravais). L. Sohncke also attempted to find a relation between
the principal cleavage form and the structure of a crystal.
The main factors which determine the crystalline habit are (1) the internal
structure of the crystals ; (2) the degree of supersaturation ; and (3) the nature of
concentration (diffusion) currents in the solution during deposition. A. A. Noyes
and W. R. Whitney (1897) showed that the dissolution of a crystal is governed by
the rate of diffusion of the dissolved molecules across the zone of falling concentra-
tion which, being replenished instantaneously from the crystal, remains saturated.
M. le Blanc applied the theory to the reverse process of crystallization, but
C. L. Wagner and A. Berthoud have pointed out that M. le Blanc's hypothesis takes
no account of the varying rates of crystal growth on different faces. C. L. Wagner
postulated that the thickness 8 of the diffusion zone varies from face to face, but,
as L. Brunner showed, even with the most violent agitation 8 is at least 0"03 mm.,
and the thickness 8 is not likely to be appreciably influenced by forces of molecular
630 INORGANIC AND THEORETICAL CHEMISTRY
magnitude at the crystal surface. A. Berthoud showed that the rate V at which
eqiulibrium tends to establish itself between a given area s of crystal surface and the
solution is proportional to the difference between the concentration Cq of the satu-
rated solution and the concentration Ci in contact with the crystal. If k denotes the
velocity constant of crystallization of the given surface ; K, the diffusion constant ;
and Cy the mean concentration in the diffusion zone, the velocity, V=ks{Ci—Co) ;
or velocity, V=sK(C—Co)l{8-\-Klk). When K/k is very small in comparison
with 8, which occurs when the solution is quite still, the formula reduces to
A. A. Noyes and W. R. Whitney's expression. The more nearly this condition is
attained in practice, the more nearly do the crystal faces approximate to equal
rates of growth, and facets appear which are not seen when the solution is
agitated. A. Ritzel also found the rate of solution of sodium chloride varied with
the degree of under-saturation of the solution.
When a crystal is growing, curiously enough, the liquid in the immediate
vicinity of the growing face is more concentrated, for it contains more of the dissolved
substance per unit volume than the liquid a short distance away from the growing
face. At first sight, it seems as if the growing crystal exerts some kind of attraction
on the molecules of the dissolved substance a short distance away. For instance,
if a saturated solution of zinc silicate in molten lead borosilicate tinted with cobalt
silicate be allowed to crystallize, the crystals of willemite which separate will also
abstract the cobalt eilicate from the solution, and form patches of beautiful " azure
blue " crystals in a colourless matrix. If no crystallization occurs, the matrix will
be uniformly coloured an intense blue. Presumably, the concentration of the
colouring agent at the crystal face is maintained by diffusion from the body of the
liquid. These facts, as well as the phenomenon exhibited by liquid crystals, lend
support to the view that as a liquid nears its crystallizing point, there is a marshalling
of the molecules of a liquid about to crystallize which culminates at the moment of
separation of the solid crystalline nucleus.
Growing crystals of hydrated strontium nitrate take up colouring matter from
a solution coloured with logwood. P. Gaubert i* showed that the different faces of
a crystal have not the same power to absorb colouring matter, e.g. certain faces of the
crystals of lead nitrate or urea nitrate growing in solutions coloured with methylene
blue do not take up the dye ; and picric acid crystals growing in solutions containing
the same colouring agent, have some faces stained blue, while others remain yellow.
R. Marc attributes the power possessed by crystallizing salts to take up organic
dyestuffs to adsorption. It must not be supposed that crystals usually gather up im-
purities from the mother liquid (except by mechanically entangling the mother liquid
with the crystal) ; the converse is often the case. The thrusting aside of impurities
by the tip of a growing crystal can be readily observed under the microscope. The
shape of growing crystals and the movements of the molecules in the solvent against
the resistance exerted by the liquid are controlled by molecular forces which are
not alike in all directions. It also follows that the physical character of the mother
liquid must modify the rate at which the molecules are supplied to the growing
crystal and determined to some extent its habit and form.
The kinetic theory of crystal growth.— The kinetic theory has taught us that
during crystallization, it is probable that a series of exchanges between the molecules
of the crystal and the molecules of the solution are going on all over the surface of
the growing crystal. Molecules of the dissolved substance are attracted to the
surface of the growing crystal, the molecules of the crystal continually pass into
solution again. If the crystal is growing, more molecules are deposited on the crystal
than are lost in unit time ; and if the crystal is dissolving, less molecules are deposited
on the crystal than are lost in a unit of time.
Let Fig. 69 represent, diagrammatically, a growing crystal, one face of which
is incomplete ; and assume that the structural units are spherical molecules. If a
sphere lodges against a completed face, it can touch three other spheres, and whether
or not the molecule leaves the growing crystal will depend upon the force of
CRYSTALS AND CRYSTALLIZATION 631
attraction exerted upon it by the three contiguous molecules. Again, suppose that
a sphere lodges on the little ledge formed by the top layer_of the incomplete face.
It will then touch five instead of three spheres ; and it will be held in place by the
attraction of five contiguous spheres. Obviously, therefore, (1) during the exchange
of molecules between the growing crystal and the solution, those molecules which
have been deposited on the growing face will be retained more tenaciously than those
deposited on a completed face ; (2) as soon as a few molecules happen to be deposited
in juxtaposition on the face of a crystal, subsequent growth on that face will be more
rapid than the sporadic growth elsewhere ; (3) an incompleted layer will rapidly
extend until it covers the entire face of the crystal, etc. These deductions are in
harmony with known facts.
G. D. Liveing's explanation of the phenomena is as follows : The surface tension
at the boundary between a crystallizing solid renders a supply of energy necessary
to generate a surface in the interior of the fluid ; and the supersaturation of air with
water vapour, and the supersaturation of solutions of salts, show that the generation
of a free surface in the interior of a gaseous or liquid fluid is not easy. Similarly,
if a surface is already formed in a fluid, as when a supersaturated solution meets the
air, or the sides of the containing vessel, if the surface energy of either boundary
be less than that at the boundary surface of the crystalline solid and the solution,
energy will have to be supplied in order to produce a new surface — but not so much
as if there were no such surface. Hence, crystals generally form on the top, or on
the sides of the containing vessel. Part of the energy of the change of state from
liquid to solid is generally available for pro-
ducing a new surface ; but when the mass
deposited is small, the energy available will
be correspondingly small — for mass varies as
the cube of the diameter, while surface varies
as the square of the diameter of a solid.
Consequently, the first solid nucleus which
separates from a solution is liable to be
squeezed back into a liquid by its own surface
tension, so as to form a supersaturated solution. ^ ^^ ^ . T^• t n-^™
.J x 11 X i. -1 t -xi- Fig. 69.— Imaginary Diagram of Grow
A deposit will form most easily on a surface with ^„ Crystal.
the same energy as that of the deposit, because
the additional energy required is only needed for the extension of the surface.
This explains the seeding of supersaturated solutions by particles of the same
salt as is in solution ; and also how big crystals grow faster than little ones, for
the ratio of the increase of surface to that of the volume decreases as the crystal
grows.
If one part of a crystal be mutilated or damaged, the injured part may grow more
rapidly than the other parts of a crystal until the injury disappears, and the perfect
crystal is restored.^^ In his study Die Regeneration der Krystalle (Leipzig, 1895-6),
G. Rauber was so impressed by the inherent power of a mutilated crystal to heal
itself that he was led to propound the hypothesis that crystals are controlled by
vital forces. D. N. Artemeeff checked the healing process by exact measurements
with the goniometer. According to T. V. Barker, he found :
In the first period of growtli the sphere exhibits a number of ghttering spots correspond-
ing with the most important faces. The rest of the surface remains matt, but later becomes
covered with tiny crystals in parallel positions, each of which contributes a part to sharp
goniometer reflections of the important forms ; simultaneously, reflections corresponding
with the less important forms appearing as glittering points. As growth proceeds, the less
important faces disappear, the tiny parallel crystals coalesce, and the final result is a hemi-
crystal boimded by common faces.
Further, if a crystal be removed from a solution in which it is growing, it
does not lose its power of growth, for if the crystal be placed in a suitable
environment at any future time, it will continue growing as if there had been no
632
INORGANIC AND THEORETICAL],CHEMISTRY
interruption. These two statements are demonstrated by the so-called caj
quartz in which there is an overgrowth of transparent quartz on an old crystal
covered with a film of clay or other material which has prevented the new growth
adhering to the old so that the capping layer can be sometimes readily detached
from the inner kernel. In ghost quartz, the film of " dust " has not been thick
enough to prevent adhesion, but is sufficient to enable the outlines of the kernel
crystal to be readily seen. In the remark-
able photograph, Fig. 70, some quartz
crystals, grown during some former
geological period, have lost their external
crystalline form by attrition as they
" knocked about the world " — blown
about as sand in the deserts, washed
down the hillsides in streams of water,
etc. — and they were finally deposited as
rounded sand grains along with the moun-
tain limestone from some prehistoric sea.
There, the damaged crystals — sand grains
— met a suitable environment in later
years — probably water percolating through
the limestone rocks, and carrying silicic
Fio. 70.-Growth of Quartz Crystals about ^cid in solution. The damaged crystals
Old Sand Grains. • j in i j •
were repaired, iliach sand gram, now
embedded in each repaired crystal, served
as a foundation for rebuilding the damaged quartz crystals on the original archi-
tectural plan. In the photograph it was impossible to get all the crystals in focus
at the same time. Here, again, we can gaze only in ignorant wonder while the
molecules of the solute deploy their mysterious forces in crystal building,
With rapt admiration we contemplate
Immortal nature's ageless harmony
And how and when, her order came to be.^ — Euripides.
I{,EFEIIENCES.
* C. F. G. H. Westfeld, Miner alogische Abhandlungen, Gottingen, 1767 ; T. Bergmann, Nova
Acta Beg. Soc. Ujpsala, 1, 1773 ; J. B. L. Rome de I'Isle, Essai de cristallographie, Paris, 1772 ;
Cristallographie, Paris, 1783.
2 R. J. Haiiy, Traite de miner alogie, Paris, 1801.
3 W. H. WoUaston, Phil. Trans., 103, 51, 1813.
* J. F. C. Hessel, Ostwald's Klassiker, 88, 89, 1897; A. Gadolin, Acta Soc. FcnnicK, 9. 1,
1867 ; Ostwald's Klassiker, 75, 1896 ; V. von Lang, Lehrbuch der Krystallographie, Wien, 1866.
^ L. A. Seeber, Gilbert's Ann., 76. 229, 1824; G. Delafosse, Mem. Acad. Belgigue,S. 641, 1843.
^ M. L. Frankenheim, Die Lehre von der Cohdsion, Breslau, 1835 ; Nova Acta Acad. Caes.
LeopoldinocarolincB Nat. Cur., 19. 47, 1842.
' A. Bravais, Compt. Rend., 27. 601, 1848; Journ. V J^cole Polyt., 29. 127, 1850; 30. 102, 197,
1851.
* C. Wiener, Die Orundzuge der Weltordnung, Leipzig, 1869 ; C. Jordan, Ann. Mat. pura ap-
plicata, (2), 2. 167, 215, 322, 1869 ; L. Sohncke, Pogg. Ann., 132. 75, 1867 ; Verh. naturwiss. Ver.
Karlsruhe, 7. 1876 ; 9. 1882 ; Wied. Ann., 6. 545, 1879 ; Zeit. Kryst., 13. 209, 1888 ; 14. 417,
426, 1888 ; 20. 452, 1892 ; 25. 529, 1896; L. Wulfif, ib., 21. 253, 1893; 36. 14, 1902.
» P. Curie, Bull. Soc. Min., 7. 89, 418, 1884.
i» E. S. von Federoff, Trans. Russian Min. Soc.,2\. 1, 1885 ; 25. 1, 1888; 26. 454, 1890; Zeif.
Kryst., 17. 610, 1890 ; 20. 25, 1892 ; 21. 679, 1893 ; 24. 210, 1895 ; 25. 113, 1896 ; 36. 209, 1902 ;
38. 321, 1903 ; 40. 529, 1905 ; L. Wulfif, ib., 21. 679, 1893 ; A. Schonflies, Nuchr. GiHt., 483, 1888 ;
239, 1890 ; Krystallsysteme und Krystallstructur, Leipzig, 1891 ; Zeit. phys. Chem., 9. 156, 1892;
10. 517, 1892 ; Zeit. Kryst., 20. 259, 1892 ; W. Barlow, Zeit. Kryst., 23. 1, 1894 ; 25. 86, 1896 ;
Proc. Roy. Dublin Soc, 8. 527, 1897.
1^ H. A. Miers, Science Progress, 1. 483, 1894 ; 3. 129, 1895 ; W. Barlow and H. A. Miers,
B. A. Rep., 297, 1904 ; F. Wallerant, Bull. Soc. Min., 21. 197, 1898 ; A. E. H. Tutton, Crystallo-
graphy and Practical Crystal Measurement, London, 114, 1911; Lord Kelvin, The Molecular
Tactics of a Crystal, Oxford, 1894.
12 G. Quincke, Ann. Physik, (4), 7. 631, 1902 ; (4), 9. 1, 1902 ; (4). 18. 1, 1905 ; H F. Link,
CRYSTALS AND CRYSTALLIZATION 633
Pogg. Ann., 46. 258, 1839 ; M. L. Frankenheim, ih., 110. 1, 1860 ; C. Brame, CompL Rend., 36.
463, 1853 ; H. Vogelsang. Die Krystalliten, Bonn, 13, 1875 ; E. Weiss, Pogg. Ann., 142. 324,
1871 ; C. Brauns, Neuea Jahrb. Min. B. ^.,13. 39, 1899 ; T. W. Richards and E. H. Archibald,
Proc. Amer. Acad., 36. 341, 1901 ; W. Ostwald, Lehrbuch der allgemeinen Chemie, Leipzig, 1. 1040,
1903 ; 0. Biitschli, Untersuchungen iiber Strukturen, Leipzig, 1898.
" J. W. Gibbs, Trans. Connecticut Acad., 3. 343, 1878 ; Scientific Papers, London, 1. 320,
326, 1906 ; P. Curie, Bull. Soc. Min., 8. 145, 1885 ; E. Brunner, Zeit. phys. Chem., 47. 56, 1904 ;
51. 95, 1905 ; C. L. Wagner, ib., 71, 401, 1910 ; L. Bruner and S. Tolloczko, ib., 35. 283, 1900 ;
Zeit. anorg. Chem., 28. 314, 1901 ; 35. 23, 1903 ; 37. 455, 1903 ; K. Drucker, ib., 29. 459, 1902 ;
S. Tolloczko, Bull. Acad. Cracow, 209, 1910 ; A. Ritzel, Zeit. Kryst., 49. 152, 1911 ; L. Sohncke,
ib., 13. 221, 1888 ; F. Novak, ib., 47. 421, 1905 ; H. Danneel, Zeit. Elektrochem^, 10. 41, 1904 ;
W. Ostwald, Lehrbuch der allgemeinen Chemie, Leipzig, 1. 1040, 1903 ; G. A. Hulett, Zeit. phys.
Chem., 37. 385, 1901 ; A.A. Noyes and W. R. Whitney, ib., 23. 689, 1897 ; Journ. Amer. Chem. Soc,
19. 930, 1897 ; F. Pockels, Naturwiss. Rund., 14. 383, 1899 ; S. Berent, Zeit. Kryst., 26. 529, 1896 ;
G. Wulff, ib., 34. 385, 1901 ; H. Liebmann, ib., 53. 171, 1914 ; T. V. Barker, Annual Reports on
the Progress of Chemistri). London, 14. 248, 1918 ; J. J. P. Valeton, Ber. Sachs. Ges. Wiss., 67. 1,
1915 ; M. le Blanc and G. Elissafoff, ib., 65. 199, 1913 ; M. le Blanc and I. I. Andreeflf, Zeit. phys.
Chem., 77. 635, 1911 ; A. Berthoud, Journ. Chim. Phys., 10. 624, 1912 ; G. Friedel, ib., 11. 478,
1913 ; C. Fastest, Neues Jahrb. Min. B. B., 33. 265, 1912 ; H. Hilton, Centr. Min., 573, 1901 ;
Mathematical Crystallography, Oxford, 105, 1903.
1* R. Marc, Zeit, phys. Chem., 61. 385, 1908; 67 470, 1909 ; 68. 104, 1909 ; 73. 685, 1910; 75.
710, 1911; P. Gaubert, Recherches recentes sur le fades des cristaux, Paris, 1911.
" J. W. Judd, Proc. Roy. Inst., 13. 250, 1891 ; G. D. Liveing, ib., 13. 376, 1891 ; D. N- Arte-
meefF, Zeit. Kryst., 48. 417, 1910 ; T. V. Barker, Annual Reports of the Progress of Chemistry,
London, 14. 246, 1918.
§ 8. Analysis of the Structure of Crystals by X-rays
All effects are exactly proportional to their causes, therefore, unless their mutual re-
lations be examined by accurate trials, theory must be lame and imperfect.- — T. Bergmann.
The works of nature which seem most desirous to escape bur scrutiny are sometimes
those which have most to show us.— R. J. Hauy (1801).
It is shown in the text-books on physics that when a b^ani of light strikes against
a series of very fine lines regularly ruled on the surface of a metal or glass plate,
each line acts as a fresh centre from which a secondary train of light waves is
diffracted. The diffracted waves enhance some of the normal light waves and damp
down others, with the result that the beam of light is analyzed into a series of spectra ;
the diffracted waves quench the coloured waves of normal light in the order and
proportion of their wave-lengths.
B. Walter and R. Pohl found that the diffraction effects produced by the passage
of X-rays through fine slits indicate that the wave-length of these rays is of the order
10~9 cm., a value but little less than the estimated distance between contiguous
molecules in a crystal. In a paper On the diffraction of short electromagnetic waves
hy crystals (1912),i M. von Laue argued that a crystal must form a natural kind of
grating on account of the regular disposition of the structural units. The units
of a crystal are, however, so small in comparison with the wave-length of ordinary
light that the crystal behaves as if it were a continuous medium when exposed to a
ray of ordinary light ; on the other hand, the wave-length of the X-rays is so short
— about i^;^th of that of light — that the structural units of the crystal form a series
of widely separated and regularly arranged particles each of which should diffract
a small proportion of the energy of the incident X-rays ; each structural unit should
be a centre of diffraction from which a secondary pulse of wavelets is diffracted
producing interference effects somewhat analogous with the effect of a diffraction
grating on ordinary light.
W. Friedrich and P. Knipping (1912) tested M. von Laue's hypothesis by allowing
a primary pencil of X-rays to pass through a crystal, and afterwards impinge on a
photographic plate. When the plate was developed the result with a crystal of zinc
blende, ZnS, when the X-rays were parallel to the diagonal axis through the centre
of the cubic crystal, was remarkable, four series of spots were formed symmetrically
grouped about a central image, as illustrated by half a photograph, Fig. 71, also
634
INORGANIC AND THEORETICAL CHEMISTRY
called a rorUgenogram or X-rayogram, or a radiogram. Several hours' exposure were
needed to produce good results because the greater proportion of the rays are not
deflected by the crystal. The dark central spot represents the undeflected pencil of
rays, while the smaller dark spots— called Laue's spots — symmetrically ranged about
the central spot, represent secondary deviated beams due to diffraction or reflection
effects of the internal planes of the crystal. The X-rays have presumably been
diffracted by the structural units of the crystal en route, so that the secondary
wavelets passed along and produced interference maxima.
M. von Laue supposed that the X-rays are electromagnetic radiations which set
up vibrations in their passage through the crystal so that each structural unit be-
comes the centre of a wave disturbance. The resulting waves undergo interference,
and a spot is produced in the diagram where a set of vibrations are so close in phase
as mutually to reinforce each other. When a crystal is placed in the path of a beam
of X-rays, the rays are partially reflected from the planes of the crystal which contain
a relatively large number of atoms, but not from planes taken at random which do
not contain many atoms. Each spot represents a partial reflection of the primary
beam of X-rays by a plane rich
in atoms. The general equations
governing the interference of a
three-dimensional grating have been
developed. W. Friedrich and E.
Wagner showed that the radiograms
are the result of a continuous spec-
trum, and not of a monochromatic
beam. It thus appears as if the
diffracted rays should cause a general
darkening of the whole of the photo-
graphic plate, and not produce well-
defined spots. P. Debye got over
the difficulty by assuming that the
structural units at ordinary tem-
peratures are not stationary, but
possess some vibratory or oscillatory
movement; and he showed that, in
consequence, the intensity of the
diffracted rays from most of the
planes must be reduced to very
low values, and only in the planes
characterized by a fairly dense
packing of particles will the effect
survive this weakening. In support of P. Debye's assumption that the atomic
vibrations are accelerated with a rise of temperature, M. von Laue and J. S. van der
Lingen found that with mica, secondary spots are faintly visible at ordinary tempera-
ture under conditions where at 400° they do not appear ; there is no trace of spots
with rock salt at 620° ; the reflected rays are weakened in intensity with a rise of
temperature ; and the mean distance apart of the atomic planes is augmented by
the expansion of the crystal as indicated by a decrease in the glancing angle of the
reflected rays. G. Friedel showed that the 32 classes of crystal symmetry can
yield a total of eleven types of radiogram ; and that the radiogram will not decide
whether or not a crystal is endowed with a centre of symmetry. In the cubic system,
for example, the holohedral, holoaxial, and tetrahedral classes all yield holohedral
patterns, but the tetartohedral and pyritohedral classes give pyritohedral patterns.
J. Stark tried to explain Laue's spots by assuming that the X-rays are corpuscular, and
that the corpuscles travel most easily in certain avenues in the crystal, each set of avenues
giving rise to a spot. G. Wulff showed that this view is not tenable since there are many
wide avenues not represented by spots in the radiogram. L. Mandelstam and H. Rohman
Fig. 71.-
-M. von Laue's Spots for Zinc Blende with
W. L. Bragg 's Projection.
CEYSTALS AND CRYSTALLIZATION 635
suggested that the spots are due to reflections at the surface of cleavage cracks, which must
be so fine as to escape detection by ordinary optical means, but M. von Laue showed that
this view is untenable.
W. H. Bragg 2 (1912) found that when an incident beam of X-rays falls on a crj^stal
face, the beam is reflected from the face itself ; he further showed that the law
of equality of the angles of incidence and reflection applies to the beam of X-rays.
As a matter of fact, the reflecting plane is not merely the geometrical surface of the
crystal, because the rays probably pass through a whole series of planes of molecules
parallel to the face before an appreciable absorption occurs, and a small amount of
energy must therefore be reflected by each of^hese planes. Thus E. Hupka
roughened the surfaces of quartz and gypsum so that they scattered ordinary
light completely, and found that the intensity of the reflected beams of X-rays
was not appreciably influenced.
Let F2, F3, ... be planes of atoms (or molecules) parallel to the crystal face Fj.
Suppose a parallel beam of X-rays LiL^, falls on a crystal face Fj. The incident
beam will be reflected by each atom, and the various atoms on the face will be centres
of propagation of the reflected beam Li'L'^. The same will be true for the succeed-
ing planes F2, F3, ... If hS (Fio;. 72) be perpendicular to the incident beam
LiL-2^, and aS perpendicular to the reflected beam L^Lc;^, the difference in the path
travelled by a ray reflected from the plane Fi, and that reflected from the plane
F2, will be bP-\-Pa, but bP=Pa, and this is a projection of the distance I between
the two consecutive planes Vi and F2 upon the direction of the incident and emergent
beam. If 6 denotes the glancing angles of the
reflected beam, bP or Pa=l sin 6 ; and the
whole difference of phase will be 21 sin 6, that
is, the trains of wavelets from each plane of
the crystal will follow one another at intervals
21 sin 6, and an interference maximum can
occur only when this distance is equal to the
wave-length A or to a multiple of A, say 2A,
3A, . . . Consequently, if the wave-length A is
such that nX=2l sin 6, where n is an integral
number, the waves will augment one another and produce a maximum inter-
ference. Consequently, if the incident beam contains rays of every possible
wave-length, the crystal will appear to select the rays with those particular
wave-lengths which follow the 21 sin 6 rule, and produce maximum interference.
The angles of reflection can be measured, and by using rays of the same wave-
length, the distance / can be compared in different crystals, and with different faces
of the same crystal.
A diagrammatic representation of Bragg's apparatus — called an X-ray spectrometer- —
is shown in Fig. 74, and a perspective drawing in Fig. 73 with corresponding lettering.'
The X-ray tube is placed in a lead box fitted with slits A and B, about a millimetre wide,
to allow a fine pencil of X-rays to pass on to the crystal C placed on a little table with its
axis passing through the crystal face and which can be rotated by the vernier V and scale
SS ; the ionization chamber / with its electrode E turns about the same axis. A second
vernier Fg indicates the angle at which the chamber / has been set in order that the beam
of X-rays reflected by the crystal may fall upon and be admitted by the adjustable slit D.
The ionization current of the chamber / with the electroscope E and reading microscope M,
indicates the strength of the reflected beam of X-rays for each angle of incidence.
P. P. Ewald has shown that although apparently so different, M. von Laue's
and W. H. Bragg's interpretations really amount to the same thing. M. von Laue
worked with transmitted rays, W. H. Bragg with reflected rays. Bragg's method
gives a rapid survey of the general structure of a crystal, and in the simpler cases
it may furnish all that is required, but the more complex cases may require to be
supplemented by Laue's radiograms, which introduce greater precision in the finer
details. Laue's radiograms, if used alone, may be inconclusive and give erroneous
results.
636
INORGANIC AND THEORETICAL CHEMISTRY
In P. Debye imd P. Soherrer's method,* a narrow beam of X-rays is allowed to traverse
an aggregate of small crystals, and the resulting diffraction pattern is photographed. The
disposition of the crystaJs is assumed to be perfectly irregular, and A. W. Hull insures this
^^S
Fig. 73. — 'X-ray Spectrometer.
by reducing the substance to a fine powder, and rotating the glass tube containing the
powder while it is being exposed. In the latter case, the diffracted rays fall on photographic
plates for some hours, and concentric bands are obtained which represent the X-rays reflected
from all the important layers of atoms instead of from one at
a time. M. de Broglie mounted the crystal, to be exposed in
the path of the X-ray pencil, upon a rotating stage.
The measurements of the ionization current are
usually plotted vertically while the angles of inci-
dence are plotted horizontally. The curve is called an
X-ray spectrum. This is done for potassium chloride
and sodium chloride crystals in Fig. 75. There is a
comparatively small reflection of the rays for all angles
_ of incidence, but a very much larger reflection for
Fio. 74. Diagrammatic Plan '^P^cial angles. This is shown by the peaks in the
of Bragg's X-ray Spectro- curves, Fig. 75. These peaks recur again for angles
^^^^^- whose sines are twice those of the former angles, they
recur again at triple these values, and so on. Since
the wave-length A and spacing I of the crystal planes parallel to a given face are con-
nected by the equation nX=2l sin 6, the second peak of the curve gives respectively
for potassium and sodium chlorides 2^1 sin J(10-43°)=A, and 21^ sin J(11-8°)=A,
where l^ and I2 respectively denote the spacings for the (lOO)-plane of the
crystals of potassium and sodium chlorides. Hence, Zi=5-48A and 12=4: SbX. This
CEYSTALS AND CRYSTALLIZATION
637
shows that while the crystals of the two salts probably have a similar structure,
the molecule of potassium chloride is more voluminous than that of sodium chloride,
This agrees with the observed molecular volumes, for obviously, if Mi and M^
denotes the respective molecular weights of potassium chloride (37*8), and sodium
chloride (27 "8), and Dj and Dzthe corresponding densities, then Z^^ ; Z2^=Mol. vol.
KCl : Mol. vol. NaCl=lfi/Di : M2ID2 ; and hence l^D/M should be a constant
for different members of this series of salts. The computed value for potassium
chloride is 5'48'^(l-97-;-75'5)A=l*63A; for sodium chloride, 1*62A; and for potassium
bromide, 1 '63 A. This is taken to mean that the structure of these salts is analogous.
r —
Potassium
chloride
— (100)-
Sodium
chloride.
—(100)—
10"
^UL
15° 20° 25° 30 5 10° 15° 20° 25°
Glancing Angle. Glancing Angle.
Fia. 76. — X-ray Spectra of Crystals of Potassium and Sodium Chlorides.
30
The curve in which the intensity is plotted against the glancing angle of in-
cidence with the (100) -face of rock salt exhibits three peaks Ai, B^, and Cj at glancing
angles 13"* 48', 11° 30', and 10° respectively. These three peaks are produced with
a platinum anticathode in the X-ray tube, and they have been found to correspond
with three monochromatic beams of X-rays with wave-lengths A^=l*316xl0~8
cm. ; A£=l-095xl0-8 cm. ; and Ac=0-96xlO-» cm. The Ap Bi, Ci peaks are
repeated by a less pronounced system of peaks A2, B^, and C2. with glancing angles
respectively 27° 36', 23° 30', and 20° ; and these again by a third system B^, and
O3 of still smaller intensity and with the respective angles 35° 50' and 30° 48'. For
the ^-peaks, therefore, sin 13° 48' : sin 27° 36'=0-238 : 0*463 ; for the B-peaks
sin 11° 30' : sin 23° 30' : sin 35° 50'=0199 : 0*399 : 0*585 ;
and for the O-peaks, sin 10° : sin 20° : sin 30° 48'=0173 :
0"342 : 0'512. These ratios correspond closely with 1:2:3.
For corresponding maxima on the (100), (110), and the
(111) planes for sylvine, KCl, the angles are respectively
5° 13', 7° 18', and 9° 3', and their sines are in the proportion
1 : \/2 : ^/2^. The same ratios occur with the sines of the
angles for corresponding maxima on these faces with rock
salt, although their absolute values are different. Let
OBFAEGBC, Fig. 76, represent a simple cubic space
lattice, the (100) -planes are parallel to OBDC, and their distance apart Iiqq—OA ;
the (llO)-planes are parallel to CBFE, and their distance apart is l(uo)'=OP ; and
finally, the (111) -planes are parallel to ABC, and their distance apart is l(iu)=OQ.
Geometrically, therefore,
Fig. 76.
(100)
1
^(110)
1^
^(111)
OA
OP'OQ
or 1 : V2 : a/3
Experiment shows that for the crystals of sylvine
^(100)
^(110) ^(111)
= sin 5° 13' : sin 7° 18' : sin 9° 3'
which is very nearly that required for the simple cube lattice 1 : '\/2 : \/3.
ratios with the three types of cubic space lattice (Figs. 49, 50, and 51) are
111
These
^(100) * ^(110) ^(111)
Simple cubic lattice
Body-centred cubic lattice
Face-centred cubic lattice
J2
n/2
^3
n/3
W3
638 INORGANIC AND THEORETICAL CHEMISTRY
The measurements for both sodium and potassium chlorides are thus in agreement
with the measurements for the simple cubic lattice.
Thus, the various glancing angles 6i, 6^, ^3, at which a plane of crystal units serve
as efficient reflectors, are determined. The wave-lengths of the rays emitted by
anticathodes of platinum, rhodium, tungsten, or palladium are known. The results
are then interpreted by means of the equation nX=2l sin d. The term n is unity
with a reflection produced by the mutual reinforcement of pulses provided by
successive chemically and crystallographically identical planes — first order reflection
at angle ^1 ; similarly, the co-operation of the second plane gives a second order
reflection at an angle 6^ ; likewise also with the co-operation of the third plane,
the third order reflection at an angle ^3 is obtained ; and so on. The intensity of
the reflection diminishes regularly in passing up the orders excepting in cases
where successive planes of crystal units differ as to composition, or distances apart,
so that the even orders may appear stronger than the odd orders, while certain
orders may vanish altogether. With rock salt, the relative intensities of the
reflections from the cubic and octahedron planes are :
1st order.
2nd order.
3rd order.
4th order.
Cube planes, face (100)
100
18-7
6-25
• —
Octohedral planes, face (111)
16-5
24-4
3-10
4-2
The regular decrease of intensity on ascending the scale of orders is interpreted
to mean that the units are chemically identical, and the spacing of the successive
cube planes is the same ; on the other hand, the periodic
sy''^^^t^^<' '^< ^^^^ ^^^ ^^^ ^^ intensity on the octahedral planes indicates
^_'_f._j_^^^fej-^' [ J some structural peculiarity ; otherwise expressed, the
|\j I ; p^(;ni\ i j reflecting units in successive cube planes are identical
I ^^^'[""^'^^^W^^::^ ill all respect ; but, assuming that the individual atoms,
^: 1 r^KI-jf? .Vjl^^S. ! and not the molecules, are reflectors, it follows that the
I i ^dBJ i : i%' reflecting units in the octahedral (111) planes are alternate
j I Jm^/l^ji^^\..\jy% layers of sodium and chlorine atoms. This gives the
! -'<^~3^£^*^^^^' arrangement shown in Fig. 77, where the solid circles
represent chlorine atoms and the open circles sodium atoms.
Crystals ^of'^the Afkar -^^^ patterns of the Laue spots for sodium and potas-
Halides. sium chlorides, potassium bromide and potassium iodide
are not identical. Potassium chloride gives the pattern
characteristic of a simple cube lattice with points at each of the four corners ; potas-
sium bromide and iodide give patterns characteristic of the face-centred lattice ;
while sodium chloride gives a pattern which seems to be intermediate between the
other types. It is known that the intensity of amplitude of the waves of the secon-
dary radiation produced by X-ray impulses is nearly proportional to the masses of
the atoms. It was therefore inferred that while in potassium chloride crystals the
atomic weights of the potassium and chlorine are sufficiently close to make the two
atoms almost equally active as centres of difiraction, the difference between the
atomic weights of sodium and chlorine is sufficient to complicate the simpler
potassium chloride pattern ; while with potassium bromide or iodide, the difference
between the atomic weights of the respective elements is so great that the effect
produced by the lighter atom is overpowered by the heavier one. If the atoms of
sodium be represented by open circles and the chlorine atoms by solid circles —
Fig. 77 — the space lattice must have an equal number of both kinds of spots, and
the arrangement of the black and white points at the corners of the elementary
cube will represent the effects produced by sodium chloride ; while if the black
spots are alone considered, the effective centres of diffraction will appear to be
located at the corners and face centres of an elementary cube. In fine, the space
lattice of sodium chloride may be regarded as being composed of two interpenetrat-
ing face-centred cubic lattices with the sodium atoms arranged on the one, and
the chlorine atoms on the other. The structure of potassium chloride is much the
CRYSTALS AND CRYSTALLIZATION 639
same, the two interpenetrating space lattices — one chlorine and one alkali metal —
are so intercalated that the chlorine space lattice is shifted over a distance of half
the edge of the cube of the metal space lattice, so that each chlorine atom falls
midway between two consecutive metal atoms, and similarly one metal atom falls
between two consecutive chlorine atoms.
According to the kinetic theory, there are 6*06x1023 molecules per gram.
The density of sodium chloride is 2167; and therefore there are 2'167x6"06
Xl023-|-58-46=2-3xl023 molecules or 4-6x1023 atoms per c.c. The atoms of
sodium are arranged cubicall^, and consequently the average distance apart of
these molecules is the cube root of the reciprocal of 4*6x1023, or 3*5 XlO"^ cm. —
very nearly one hundred-millionth of an inch.' This means that along the edge
of a crystal of sodium chloride there are nearly a hundred million atoms per inch.
The absolute dimensions of a space lattice can be calculated from the glancing
angle 6 — the angle of reflection of X-rays from crystals — which can be measured
with great accuracy, and the known wave-length A, by Bragg's equation nX=2d sin 6,
where d denotes the spacing of the crystal planes parallel to the crystal face. If
V denotes the volume of the elementary cell ; M, the molecular weight ; D, the
density ; n, the number of molecules in the cell ; and N the number of molecules
in a gram-molecule (6*06x1023), nM—NvD. Consequently, the atomic weight of
an element can be calculated from the observed data when the atomic weight of
the other elements in combination with it are known.
These experiments give direct proof that the structural units of crystals are
arranged in space-lattices. There is little indication of the way these units are
united to form the so-called chemical molecule. Indeed, the crystal molecule as a
structural unit seems to have lost its significance. (1) There is no evidence of
chemical combination ; (2) a sodium atom is no more closely attached to one chlorine
atom than it is to any other ; (3) there is no sign of a molecular structure in the
chemical meaning of the term ; and (4) each atom is an integral part of the whole
crystal, and is not connected with any particular group of atoms which form its
chemical molecule. The whole crystal endlessly extended in all directions seems to
form one gigantic crystal molecule. It must be added that the scattering of X-rays
is a purely atomic effect, consequently, although the exploration of crystal structure
by these rays may reveal the mean positions of the atoms, yet, from the very nature
of the case, it cannot throw direct light on the existence or non-existence of molecules
in the crystalline condition.
The view that the radiograms demonstrate that no chemical molecules exist in
crystals is not generally accepted. F. Rinne ^ claims that groups corresponding
with molecules can be often recognized. A. L. W. E. van der Veen, and A. Smits
and F. E. C. Scheffer claim that since the distances between the atoms in the solid
state are small compared with their diameters, the atoms belonging to the same
molecule can be but slightly closer than those belonging to different ones, and that
the difference cannot be detected by the X-rays. A. Fock adds that even though
the radiograms of sodium chloride show that six chlorine atoms surround one
sodium atom this does not prove that one chlorine atom is not combined with one
sodium atom. Isomorphism, electrical conductivity, and the fact that crystals and
solutions of the same substance give the same molecular weight is taken by
A. Fock to demonstrate the continued existence of molecules. P. Groth also claims
that interatomic connections must remain even though the chemical molecule as
such has lost much of its significance from a crystallographic point of view. The
special nature of the symmetry elements in some crystals is closely related to the
atomic structure of the chemical molecule itself, so that this cannot have any signifi-
cance if there are no chemical molecules in the crystal. J. Beckenkamp holds that
the ultimate structures of crystals are triclinic, and by submicroscopic twinning
systems of higher symmetry are produced ; and that the radiograms merely give
the average positions of the atoms. A. C. Crehore mathematically investigated the
mechanical forces between the atoms in a space lattice, and states that the essential
640
INOKGANIC AND THEORETICAL CHEMISTRY
difference between crystal and molecular structures is that in the former, the atoms,
under certain mutual restrictions, may revolve about non-parallel axes, while in the
latter, all atomic rotations are necessarily parallel.
W. L. Bragg found the results with zinc blende, ZnS ; fluorspar, CaF2 ; and
calcite, CaCOs, to be almost identical when the X-rays are taken diagonally through
the centre of the cube. Each point in a space lattice is situated with respect to
its neighbours like every other point ; and, in the case of the three compounds
just indicated, it is possible to satisfy the conditions only by assuming that each
molecule acts as a single point in that it contains one atom heavier than the
others which is responsible for the observed diffraction pattern. Hence, it is
probable that single atoms are associated with each diffracting unit.
Still further, the value /(D/ikf)* is nearly constant for crystals of potassium
chloride, KCl ; sodium chloride, NaCl ; zinc blende, ZnS ; fluorspar, CaF2 ; and
pyrite, FeS2. Hence, it is inferred that the number of molecules associated with
each diffraction centre is the same ; and since the crystals are so differently consti-
tuted, it is probable that one and only one molecule is associated with each diffract-
ing centre. The volume of unit parallel opided of the space lattice for potassium
chloride is about one-eighth that of the other crystals, because both its atoms, having
nearly the same atomic weights, are equally effective as centres of diffraction, whereas
with the others, only the heavier atom is effective.
W. H. and W. L. Bragg 6 found that in the double face-centred space lattice of
the diamond, each carbon atom (black and shaded spots. Fig. 78) is surrounded by
four other carbon atoms at equal distances away,
and which are related to it like the four apices of a
tetrahedron are related to the centre. If series of
adjacent space lattices be examined, it will be
found that the atoms appear to be arranged in a
series of rings of six (heavy dotted lines. Fig. 78).
This recalls the benzene, CgHg, ring. The diamond
has also been investigated by L. Foppl, W. Barlow,
A. L. W. E. van der Veen, and P. P. Ewald. W. L.
Bragg found the space lattice of zinc hlende, ZnS,
to be similar to that of the diamond, if sulphur
atoms (circles, Fig. 78) and zinc atoms (black
spots, Fig. 78) be alternately substituted for carbon
atoms. The zinc atoms now occupy the corners and
a cube with a sulphur atom in the centres of the alternate
Fig. 78. — Space Lattice of Zinc
Blende (and the Diamond).
face-centres of
small cubes so that each sulphur atom is surrounded by four symmetrically
placed zinc atoms ; and each zinc atom is surrounded by four symmetrically
placed sulphur atoms. The structures of zinc blende and of the diamond
are thus based on the so-called double face-centred cubic lattice formed by
the interpenetration of two face-centred cubic lattices. The space lattice for
fluorspar, CaF2, resembles that of zinc blende with the calcium atoms occupy-
ing the corners and face-centres of a cube, and fluorine atoms in the centres of all
the small cubes, instead of in alternate cubes as was the case with the sulphur
atoms of zinc blende. Each fluorine atom is now surrounded by four symmetrically
placed calcium atoms, but each calcium atom has eight fluorine atoms arranged
around it and related to the central calcium atoms as the eight corners of a cube
are related to the centre. The crystals of magnetite, Fe304, show that the space
lattice is fundamentally the same as the diamond with groups Fe304 taking the place
of carbon atoms. Two out of the three atoms of iron are surrounded by four oxygen
atoms arranged at the corners of an imaginary tetrahedron about each atom of iron
as centre. The other atom of iron is arranged so that it is surrounded by six oxygen
atoms belonging to the ox}^gen tetrahedra. It is therefore thought that the two
iron atoms are probably bivalent, and one iron atom is tervalent representing a
structural formula ¥e2"^Fe^^^0^. Quite similar results are obtained with spinel,
CRYSTALS AND CRYSTALLIZATION
641
Mg2A104, in which magnesium takes the place of bivalent iron in magnetite, and
aluminium the place of ferric iron. L. Vegard reported that in zircon, ZrSi04, the
atoms of zirconium and silicon are arranged alternately in a space lattice of the
face-centred tetragonal type ; each silicon and zirconium atom appears to be as-
sociated with two oxygen atoms forming Si02- and Zr02-groups as structural units
of the space lattice. The oxygen atoms appear to be closer to the silicon atoms
than to those of zirconium, probably because of the greater affinity of silicon for
oxygen. Space lattices of crystals of rutile, Ti02, and of cassiterite, S11O2, resemble
those of zircon, in which silicon and zirconium atoms are replaced by identical atoms
of titanium or tin.
The space lattices of crystals of copper, silver, gold, and lead show that the
structural units are arranged like a face-centred cube, Fig. 51. Each unit is sur-
rounded by twelve equidistant units. W. L.
Bragg found the space lattice of pyrites, reS2,
is rather complex ; it shows that the atoms of
iron are arranged at the corners of a face-centred
cube (circles. Fig. 79), and each iron atom has
four equidistant sulphur atoms (black spots,
Fig. 79) around it, and others at a slightly greater
distance away. Each sulphur atom has three
iron atoms arranged around it, with other iron
atoms at a slightly greater distance away. W. L.
Bragg also found that hauerite, MnS2, ullman-
nite, NiSbS, and cohaltite, CoAsS, have a similar Fig.79.— Space Lattice of Pyrites,
structure. Cuprite, CU2O, belongs to a similar class.
W. H. and W. L. Bragg found that the space lattice of calcite, CaCOs, Fig. 80,
shows that the carbon and oxygen atoms occur on triangular planes perpendicular to
the crystal axis. The calcium atoms lie in planes just above and below the carbon
and oxygen planes, so that each carbon atom is surrounded by six equidistant
oxygen atoms. In Fig. 80, the large black dots represent carbon atoms, the small
black dots oxygen atoms, and the circles calcium atoms. For the sake of clearness,
the oxygen atoms are omitted from the upper part of Fig. 80, and the arrangement
^Y
Fig. 80. — Space Lattice of Calcite.
of the oxygen atoms is shown in the lower part of the diagram as a series of layers
perpendicular to the trigonal axis. The position of the different planes is obvious
from the lettering. Each carbon atom is associated with three oxygen atoms, while
the latter are only associated with one carbon atom, and the distance between the
calcium and carbon atoms is greater than the distance between the oxygen and
carbon, or the oxygen and calcium. Each CO3 group thus appears as a unit equi-
distant from the six calcium atoms. The space lattices of rhodochrosite, MnCOs,
siderite, FeCOs, ^^^ sodium nitrate, NaNOg, are similar to that of calcite. The change
in the valency of the sodium and nitrogen in sodium nitrate from the basic element
and carbon in the carbonates is noteworthy since it appears to make no difference
in the general arrangement of the atoms in the crystals. In dolomite, MgCOs. CaCOs,
the structure is similar to that of the carbonates with alternate atoms of calcium
VOL. I. 2 T
642 INORGANIC AND THEORETICAL CHEMISTRY
and magnesium in place of calcium. The structure of hcematite, Fe203, belongs to the
calcite class with the carbon atoms removed and each calcium atom replaced by two
iron atoms arranged like a dumb-bell parallel to the c-axis.
J. Herwig ' investigated the space lattice of gypsum ; S. Nishikawa investigated the spinel
minerals — magnetite, Fe304 ; ruby-spinel, MgAl204 ; W. L. Bragg investigated magnetite ;
F. M. Jager and H. Haga have examined d- and /-sodium chlorate ; ammonium-iron and
pota-ssiura-chromium alums ; d- and Z-triethylenediamine cobaltic bromide ; beryl, apatite,
ethylsulphates of the rare earths, nephelene, calcite, dolomite, phenacite, tourmaline,
quartz, cinnabar, aragonite, topaz, anhydrite, cordierite, hambergite, hemimorphite, struvite,
sodiiun ammoniima d-tartrate, Z-asparagine, zinc sulphate, and benitoite ; F. Rinne has
examined cyanite, diopside, epidote, scolecite, sucrose, anhydrite, aragonite, calcite, dolo-
mite, quartz, carborundum, beryl, and cuprite. L. Vegard has studied silver, gold, lead,
anatase, ammoniiun iodide, tetramethyl ammonium iodide, rutile, cassiterite, zircon,
xenotime, and thorite. C. M. Williams has also studied the rutile group, and his results are
not always in agreement with those of L. Vegard. Copper was examined by W. L. Bragg ;
scheelite and vmlfenite by R. G. Dickinson; iron, silicon, aluminium, sodium, lithium,
nickel, magnesium, graphite, and the diamond by A. W. Hull ; chalcopyrite, by C. L.
Burdick and J. H. Ellis; barium, strontium, and lead nitrates, by S. Nishikawa and
K, Hudinuki; garnet, by S. Nishikawa; different forms of silica, by S. Kyropoulos; and
white and grey tin, by A. J. Bijl and N. H. Kolkmeijer.
L Langmuir's theory o! solids and liauids.— I. Langmuir believes that the work
of W. H. and W. L. Bragg shows that in all probability crystals are not built of
molecular units in the ordinary sense of the term. With potassium chloride, for
instance, each atom of potassium is surrounded by six equidistant atoms of chlorine
arranged as if they were placed at the corners of an octahedron ; each chlorine atom
is similarly surrounded by six equidistant potassium atoms. The identity of the
molecules of potassium chloride, KCl, thus appears to be lost, unless the whole
crystal itself be regarded as itself forming one molecule. Each atom is united
chemically with all the adjacent atoms, and these in turn are similarly united with
those beyond. Consequently, the ordinary conception of valency no longer holds
good, each atom appears to be united with far more atoms than corresponds with
the normal valency. The valency of potassium, for instance, equally divided be-
tween six chlorine atoms, and the valency of chlorine between six potassium atoms.
Considerations like these led I. Langmuir ^ (1916) to elaborate the definition of a
molecule. A molecule, said he, is a group of atoms held together by atomic forces ;
the gas molecule is defined in terms of Avogadro's hypothesis, while a continuous
liquid or solid mass is called a solid or liquid molecule.
No structural formula consistent with the primary valency of the constituent
atoms can be employed to represent the structure of crystalline solids like potassium
chloride. I. Langmuir (1916), assuming that the primary valencies hold good for
gaseous molecules of potassium or sodium chloride ; and that if there were no residual
affinity or secondary valencies developed when the temperature is lowered or the
pressure raised, these molecules could not condense to form a liquid or solid, supposes
that in the solid state, what Werner calls the secondary valencies, altogether supplant
the efEects of primary valency. He also bases a similar assumption for molecules
formed from bivalent and tervalent atoms upon Bragg's space lattices for zinc
blende, ZnS ; fluorite, CaFg ; pyrites, FeS2 > hauerite, MnS2 ; magnetite, FcgOi ;
and spinel, MgAl204. In the case of the diamond, each carbon atom appears to be
surrounded by four others equidistant from and arranged around the central atom
much as the four corners of a regular tetrahedron are related to the centre. Conse-
quently, the primary valencies of the quadrivalent carbon atom seem to exert some
influence on the formation of the solid crystal. In cubic crystals of methane, CH4,
the carbon atom is probably surrounded by four hydrogen atoms held by primary
valencies, and the crystal is held together by secondary valencies, so weak indeed
that methane melts and boils at very low temperatures. Langmuir represents
the constitution of a methane crystal diagrammatically as in Fig. 81, where each
hydrogen atom is associated with a particular carbon atom ; each carbon atom is
CRYSTALS AND CRYSTALLIZATION 643
associated with four hydrogen atoms ; and all other hydrogen atoms besides
these four are combined with a different carbon atom. Langmuir calls aggregates
H H H
HCHHCHHCH
H H H
H H H
HCHHCHHCH
H H H
Fig. 81.
of atoms of this kind group molecules, because the atoms in the group may be
distinguished from those outside the group. A crystal of zircon, ZrSi04, likewise
contains group molecules Zr02 and Si02. Similar remarks apply to the quadri-
valent atoms of titanium and zirconium in zircon, Zr02, rutile, Ti02, and cassiterite,
Sn02, where rontgenograms by L. Vegard (1916) show that the molecular groups
Zr02, Ti02, or Sn02 form the structural elements in the space lattices.
When the vapour of potassium chloride is condensed to a solid and re- vaporized,
it is probable that when a potassium atom escapes from the surface of the solid, it
takes away one of the four adjacent chlorine atoms to form a gaseous molecule of
potassium chloride, and judging from rontgenograms of the crystals, the chances are
against the two partners being the same as were previously united before the con-
densation of the vapour. On the other hand, it is probable that if methane were
treated similarly, the same four atoms of hydrogen would remain united to the same
carbon atom in both the solid and gaseous states. I. Langmuir calls the largest
aggregates of atoms which may pass from the gaseous to the solid or liquid phase
and back again to the gaseous phase without exchanging atoms with other aggre-
gates, a fixed molecule.
Still following I. Langmuir, since the secondary valencies of inorganic solids
usually supplant the primary valencies exhibited by the substances in the gaseous
state, (i) the composition of a solid should give little or no information about the
primary valencies ; and (ii) it should be possible to make more solid compounds
than accord with the rules for primary valencies. If the arrangements of the atoms
are regular so as to form a space lattice, the resulting solid should satisfy the tests
for a chemical compound, even though it exhibits no relationship with the primary
valencies of the constituent atoms. A large number of compounds are known in
the solid state which do not accord with the ordinary doctrine of valency. For
example, G. Tammann (1906-7) ^ obtained compounds corresponding with AgMgs,
AgMg, AuZn, AugZng, CugAl, CuAl, CuAl2, Mg4Al3, AlSb, ZuyFe, NaZni2, NaCdg,
etc., and he found that about 26 per cent, of the binary metal compounds which
he investigated have formulaj in accord with those based on the primary valencies
of the elements. In the case of minerals, and particularly the silicates, the number
of exceptions to the valency doctrine is greater than with alloys. In these compounds,
as also with potassium chloride, the atoms are held together by secondary valencies,
and they have a definite composition because the constituent atoms are arranged
as a space lattice. The reason solids so frequently accord with the ordinary rules
for valency is that they are usually formed from solutions or from gaseous phases.
Even in the case of alloys formed by solidification from a fused mixture, certain
restrictions are necessarily imposed by the very method of formation. Hence,
I. Langmuir argues that if methods for the preparation of solid compounds at
sufficiently low temperatures could be devised, there is no conceivable limit to the
number of possible compounds formed by secondary valencies.
If the units A and B in a space lattice could be replaced in an irregular manner
by another unit C, the resulting crystal would not have a definite composition, and
would not therefore be recognized as a chemical compound, but would rather be
said to have formed a solid solution or mixed crystals. There is no reason to suppose
that the forces holding the structural units together are any different in kind in the
644 INOKGANIC AND THEOKETICAL CHEMISTRY
two cases ; nor is there any reason, other than mere definition, to suppose that if
one combination is a chemical compound the other is not a chemical compound.
Again, a solid body built up from units — either atoms or group molecules — arranged
irregularly in space would not form a space lattice, but it would be called an amor-
phous substance or a glass. The case of thorite is curious. L. Vegard lo found that
tetragonal crystals of thorite, ThSi04, are isotropic in polarized light, and show no
indications of an optic axis ; and Rontgen ray analyses show that while the crystal
preserves the outward form of a tetragonal crystal, the original lattice resembling
zircon has in the course of time been completely broken down and only the outer
frame remains to indicate the original orientation of the atoms. The evidence thus
seems to indicate that the internal structure is the same as that of an amorphous
solid, and this in spite of its external crystalline form.
It is very doubtful if the nature of the forces holding together the units of amor-
phous substances or glasses are any different in kind from those in crystals. This
recalls F. Wald's assumption that the composition of chemical compounds is variable,
and out of all the possible variations which actually occur, chemistry reserves the
term compounds for those of constant definite composition. This, says F. Wald, is
quite an arbitrary choice. F. Wald also argues that the law of multiple proportions
as well as the other stochiometrical laws are really founded on similar conventions. n
References.
1 W. Friedrich, P. Knipping, and M. von Laue, Sitzber. Munckener Akad., 303, 1912 ; M. von
Laue, Ann. Physik, (4), 41. 971, 1913 ; Bar., 50. 8, 1917 ; Ann. PhysiJc, (4), 50. 433, 1916 ; Phys.
Zeit., 14. 421, 1075, 1913 ; G. Wulff, ib., 14. 217, 1913 ; Zeit. Kryst., 54. 59, 1914 ; M. von Laue
and F. Tank, Ann. Phy.nk, (4), 41. 1003, 1913 ; B. Walter and R. Pohl, ib., (4), 25. 715, 1908 ;
J. Stark, Phys. Zeit., 13. 973, 1912 ; L. Mandelstam and H. Rohman, ib., 13. 220, 1912 ;
G. Friedel, Com'pt. Rend., 157. 1533, 1913; J. Olie and A. J. Bijl, Proc. Acad. Amsterdam, 19.
920, 1917; A. Schonflies, Zeit. Kryst, 54. 545, 1915; 55. 321, 1915.
2 W. Friedrich, Phys. Zeit., 14. 1079, 1913 ; M. von Laue and J. S. van der Lingen, ib., 15.
76, 1914; E. Wagner, ib., 14. 1232, 1913; P. P. Ewald, ib., 14. 465, 1913; 15. 399, 1914;
P. Debye, Ber. deut. phys. Ges., 15. 857, 1913 ; E. Hupka, ib., 15. 369, 1913 ; W. L. Bragg, Proc.
Cambridge Phil. Soc, 17. 43, 1913 ; Nature, 90. 410, 1912 ; W. H. Bragg, ib., 90. 572, 1912 ; W. H.
and W. L. Bragg, Proc. Roy. Soc, 88. A, 428, 1913 ; 88. A, 277, 1913 ; W. L. Bracg, ib., 88. A,
248, 1913 ; W. H. Bragg, ib., 88. A, 246, 1913 ; Phil. Mag., (6), 27. 881, 1914.
* I am indebted to Prof. W. H. Bragg for permission to use his diagrams of space lattices ;
and to Messrs. Pye & Co. (Cambridge) for the use of Fig. 73.
* P. Debye and P. Scherrer, Phys. Zeit., 18. 291, 1917 ; A. W. Hull, Phys. Rev., (2), 10. G61,
1917 ; G. Wulflf and N. Uspensky, Phys. Zeit., 14. 785, 1913 ; T. Terado, Proc. Math. Phys. Soc,
Tokyo, 7. 60, 1913 ; M. de Broglie, Jonrn. Phys., (5), 4. 101, 1914.
» F. Rinne, Neues Jahrb. Min., ii, 47, 1916 ; Naturw. Rund., 4. 221, 233, 1916 ; Zeit. anorg.
Chem.y 96. 317, 1916; A. Fock, Centr. Min., 392, 1916; J. Beckenkamp, ib., 97, 1917;
A. L. W. E. van der Veen, Proc. Acad. Amsterdam, 25. 993, 1917 ; A. C. Crehore, Phil. Mag., (6),
26. 25, 1913 ; (6), 29. 750, 1915 ; (6), 30. 257, 613, 1915 ; A. Smits and F. E. C. Scheffer, Proc. Acad.
Amsterdam, 19. 432, 1916 ; W. Voigt, Phys. Zeit., 17. 76, 128, 152, 1916 ; P. Groth, Ber., 47.
2063, 1914; Zeit. Kryst., 54. 65, 1914; P. Niggli, Zeit, anorg. Chem., 94. 207, 1916; P. Pfeiffer,
ib., 97. 161, 1916.
« W. H. Bragg, Phil. Trans., 215. A, 266, 1916 ; W. L. Bragg, Phil. Mag., (6), 28. 355, 1914 ;
W. H. Bragg and W. L. Bragg, X-rays and Crystal Structure, London, 1915 ; Proc. Roy. Soc , 89.
A, 277, 1913 ; W. H. Bragg, ib., 89. A, 246, 1913 ; W. L. Bragg, ib., 89. A, 248, 468, 1913 :
W. Barlow, ib., 91. A, 1, 1914; L. Foppl, Phys. Zeit., 15. 191, 1914; P. P. Ewald, ib., 15.
399. 1914 ; Ann. Physik, (4), 44. 267, 1914 ; A. L. W. E. van der Veen, Zeit. Kryst., 51.
645, 1913.
' J. Herwig, Phys. Zeit., 14. 417, 1913 ; S. Nishikawa, Sug. But. Kizi Tokyo, (2), 8. 199, 1915 ;
F. M. Jager, Proc. Acad. Amsterdam, YJ. 1204, 1915 ; H. Haga and F. M. Jager, ib., 17. 430,
1916; F. Rinne, Ber. Sachs. Qes. Wiss., 67. 303, 1915; L. Vegard, Phil Mag., (6), 31. S3,
1916 ; (6), 32. 65, 505, 1916 ; (6), 33. 395, 1917 ; C. M. Williams, Proc. Roy. Soc, 93. A, 418, 191 7 ;
W. L. Bragg, Phil. Mag., (6), 28. 355, 1914 ^ R. G. Dickinson, Journ. Amer. Chem. Soc, 42. 85,
1920; C. L. Burdick and J. H. Ellis, ib., 39. 2518, 1917; A. W. Hull, Phys. Rev., (2), 10. 661,
1917; S. Nishikawa, Proc Tokyo Math. Phys. Soc, (2), 9. 194, 1917; S. Nishikawa and
K. Hudinuki, ib.. (2), 9. 197, 1917; S. Kyropoulos, Zeit. anorg. ('Item., 99. 197, 1917; A. L. Bijl
and N. H. Kolkmeijer, Proc Acad. Amsterdam, 21. 494, .501, 1919.
8 I. Langmuir, Journ. Amer. Chem. Soc, 38. 2220, 1916.
' G. Tammann, Lehrbuch der Metallographie, Leipzig, 1914.
i<» L. Vegard, Phil. Mag., (6), 32. 93, 1916.
CRYSTALS AND CRYSTALLIZATION 645
11 F. Wald, Zeit. phys. Chem.y 18. 337, 1895 ; 19. 607, 1896 ; 22. 253, 1897 ; 23. 78, 1897 ; 24.
315, 1897 ; 25. 525, 1898 ; 26. 77, 1898 ; 28. 13, 1899 ; W. Ostwald, Journ. Chem. Soc, 85. 506,
1904 ; The Fundamental Principles of Chemistry (London, 1909) ; M. Planck, Wied. Ann., 57.
72, 1896 ; P. Volkmann, ib., 61. 196, 1897 ; L. Boltzmann, i6., 57. 39, 1896 ; 58. 695, 1896.
§ 9. Liquid Crystals ; Crystalline Liquids ; or Anisotropic Liquids
The very name seems to be a self-contradiction. How can a liquid be a crystal, and
how can a crystal be a liquid ? — H. A. Miebs (1896).
In 1876, 0. Lehmann i found that at temperatures above 146°, silver iodide can
flow like a viscous solid, and that although it is actually in the liquid condition,
it still exhibits several properties characteristic of crystals. Further investiga-
tions, by F. Reinitzer (1888), on cholesteryl benzoate; by L. Gattermann
(1890), on ^-azoxyanisole and ^-azoxyphenetole ; and by 0. Lehmann himself on
ammonium oleate, etc., have shown that the phenomenon is not uncommon ;
rigid solidity is not an essential characteristic of crystals. If the temperature of
these substances be gradually raised, while .they are on the stage of a microscope —
called a crystallization microscope — it will be observed that double refraction indicates
that the molecules have a definite alignment at temperatures above their melting
point when the crystals, if touched with a needle, wobble like jellies, for they are
then soft, compressible, elastic, more or less viscid, turbid, anisotropic liquids. The
term liquid crysisiis—flussige Kristalle — was therefore proposed by 0. Lehmann for
substances which have the characteristic properties of crystals — excepting solidity
and geometrical form. In order to avoid the hypothesis implied in the cognomen
liquid crystals, some prefer the term anisotropic liquids, or birefringent liquids.
The molecules of a crystalline solid are arranged quite regularly, and they are
retained more or less rigidly in position by elastic forces. Liquid crystals have the
usual properties of liquids, but unlike ordinary liquids they also show : (i) double
refraction ; and (ii) interference colours in polarized light. It seems as if the directive
cohesive forces which bring the molecules together, at the softening temperature,
are not sufficient to fix them so rigidly about their centre of gravity as to prevent
the mass wobbling. In consequence, the optical properties of liquid crystals show
that (i) they have an internal structure which in some respects is characteristic of
crystals, but that (ii) their external faces are more or less indistinct and mobile.
I. F. Homfray ^ found the solubility of carbon dioxide in the liquid crystals is 18,
and in the isotropic liquid, 26. The optical properties of liquid crystals have been
studied by 0. Lehmann, D. Vorlander, E. Dorn, and F. Wallerant.
C. Mauguin found that the liquid crystals of 2?-azoxyanisole take up a definite
orientation on a fresh cleavage surface of muscovite mica ; and F. Grand] ean found
the same result obtains, in ninety cases out of a hundred, with a number of such
liquids on fresh cleavage faces of talc, muscovite, phlogopite, brucite, blende, orpi-
ment, pyrophyllite, rock-salt, sylvine, and leadhillite. The orientation is sometimes
independent of temperature, and in other cases it changes continuously or discon-
tinuously with temperature. The continuous variation with temperature is taken
to show that this property is not necessarily due to the alignment of the molecular
axes of the liquid on a row of structural particles in a space lattice, but is an equi-
librium property dependent on capillarity.
The surface tension of a liquid tends to make the surface occupy the smallest
possible area ; and a growing crystal likewise tends in the same direction, viz.
minimum surface area. The molecules, however, during crystallization are also
under the influence of opposing directive forces which make the crystal assume its
characteristic geometrical form. If the surface tension were the stronger force, the
crystal would assume a spheroidal form. The smaller the volume of a given mass,
the greater the relative effect of surface tension, and conversely. With thin films, the
effect of surface tension is very pronounced, and in 1857, M. Faraday ^ showed that
when thin films of gold or silver on glass are heated, the mirror loses its reflecting
646
INORGANIC AND THEORETICAL CHEMISTRY
power ; the metal, under the influence of surface forces, and in spite of the directive
crystalline forces, collects itself into globular aggregates just as occurs when a
thin film of oil on the surface of water collects itself into globular aggregates.* The
surface tension of the gold prevents the crystalline forces developing a characteristic
geometrical shape.
When the temperature of a small portion of a crystalline solid is raised,the internal
molecular motions are presumably augmented, and this weakens the directive
forces which produce crystallization ; surface tension is acting the whole time ;
ultimately, the directive forces yield to surface tension, and the crystals assume a
more or less globular form. The solid is then said to have melted. It is claimed that
in liquid crystals, the directive forces of crystallization are not completely over-
powered by surface tension, although the two are almost balanced, for the fluid
crystals are more or less rounded as illustrated in Figs. 82 and 83. The effects of
surface tension are also seen when two round liquid crystals are brought into
contact ; union takes place, and a single rounded crystal is formed — the surface
area of the single crystal is less than the sum of the surface areas of the two parent
Fia. 82. — Liquid Crystals of Am-
monium Oleate.
Fig. 83. — Parazoxyanisole — crossed
Nicols (0. Lehmami).
crystals. The elongated liquid crystals of potassium oleate unite only when the long
axis of the one is nearly parallel with that of the other, and not if the two are in
contact with their long axes at right angles to one another. 5
G. Friedel and F. Grandjean here show that the shapes of the liquid crystals of
ammonium oleate have no similarity with the crystal forms, but are figures of revolu-
tion of great complexity ; and C. Mauguin found that in some of the more mobile
liquid crystals there is no definite shape, but a continuous internal movement in
the smallest globule that can be isolated for observation. The movement increases
in intensity as the temperature rises, but no definite regularity could be detected.
0. Lehmann frequently emphasized his opinion that the globules are not liquid
crystals but rather aggregates of such ciystals.
Under certain conditions, some crystals can be so affected by heat, pressure, etc.,
that they pass suddenly into a new system more stable under the altered conditions
— much as a half-opened pocket-knife closes with a snap. There are many examples
of substances which pass abruptly from one solid modification to another without
any transitional liquid state. For instance, H. A. Miers has shown that a section
of boracite under the polarizing microscope appears to be traversed by doubly
CRYSTALS AND CRYSTALLIZATION 647
refracting lamellae, and when the section is warmed to 265°, a cloud seems to pass
over the crystal and it becomes dark — the twin lamellae reappear on cooling down
below the same temperature ; and W. J. Pope has shown that molten chloral hydrate
on a microscope slide cools to a film of uniaxial needle-like crystals, and these on
standing gradually pass into biaxial lamellar crystals.
There is a temperature at which a crystalline solid loses its elasticity and becomes
a turbid anisotropic liquid, which at a higher temperature loses its turbidity, clears
and forms an isotropic liquid. With ^-azoxyanisole, for example, there are two
transition temperatures corresponding with these changes : Solid crystals->liquid
crystals->ordinary liquids ; and L. Gattermann found
116° 134°
^-Azoxyanisolesoiidcrystais^-Azoxyanisoleiiquidcrystais^-Azoxyanisoleuquid
A comparison of H. B. Roozeboom's diagram (1900) ,6 Fig. 84, with the correspond-
ing diagram for water, will show the conditions under which H. B. Roozeboom thinks
that the liquid crystals are related on the one hand to the solid crystals, and on the
other to the liquid and vapour. There are two triple or transition points instead of
one. At one triple point 0, solid, liquid crystals, and vapour are in equilibrium,
and at the other triple point 0', liquid crystals, liquid and vapour are in equilibrium.
The change from an anisotropic to an isotropic liquid is characterized by a small
heat absorption. E. Bose and F. Courat found the energy change to be very small
in passing from the liquid to the liquid crystal phase, in
comparison with that which obtains in passing from the
liquid crystal to the solid phase ; with anisaldazine,
the former was about one calorie, the latter 20 cals.
C. de Kock and R. Schenck found the molecular latent
heat of 23-azoxyanisole to be 0*68 cal. C. Tubandt and
E. Lorenz also found that purified silver iodide forms no
plastic or liquid phase at 550°, two degrees below its
melting point. Similar remarks apply to the chloride
and bromide of silver, and the three thallium halides emperature
which H. Stoltzenberg and M. E. Huth once believed ^'curvet^JlTow^ "^LTuid
to form liquid crystals. All these substances there- Crystals (Hypothetical),
fore must be definitely deleted from the list of liquid
crystals. The point involved is not whether these crystals are so soft that it would
be a misnomer to call it a solid, but rather whether the turbid double refracting
liquids are restricted to organic compounds of some complexity, or whether
representatives are to be found among the simple binary compounds. P. N.
Pawloff, G. Wulfi, and W. Voigt have also discussed the nature of liquid
crystals.
G. Quincke (1894) suggested that the effects obtained by 0. Lehmann were
produced by a trace of oil in the substance he examined ; and G. Tammann (1905) '^
argued that the turbidity of liquid crystals in contrast with the clearness of solid
crystals shows that the former are emulsions or suspensions and not homogeneous
substances, but 0. Lehmann replied that the turbidity of liquid crystals is not in
evidence when they are examined under the microscope, and that the apparent
turbidity is a secondary effect due to the aggregation of a large number of crystals
differently oriented, just as marble appears opaque when observed en masse, even
though it is really composed of a mass of transparent crystals of calcite. All attempts
to separate the alleged emulsion into its constituent parts by G. Bredig and G. von
Schukowsky and A. Cohn by electrostatic or centrifugal processes have been futile.
No clear proof of heterogeneity has been obtained. The temperature of liquefaction
is constant, and is affected by pressure and admixture with foreign substances j ust
as it is in the case with ordinary crystals. 0. Lehmann (1910) studied the segrega-
tion of impurities during the formation of liquid crystals. Attempts have been made
to explain some of the phenomena which occur in the vicinity of the melting point of
648 INORGANIC AND THEORETICAL CHEMISTRY
metals by the existence of a liquid crystal phase.® Some metals assume a plastic
condition at some distance below their recognized melting points.
The cause of the anisotropy of crystals is referred by A. Bravais ^ to the arrange-
ment of molecules in parallel planes — ^that is, in certain privileged directions ; more
usually it is referred to the orientation of dissymmetrical molecules. For example,
the theory suggested by W. Weber in 1850 to explain magnetization assumes that
the molecules are dissymmetrical in having two magnetic poles— ^one positive and one
negative — and a body is magnetic only when a larger proportion of the constituent
magnetic molecules are disposed with like poles in one direction. When the molecules
of a body occupy random positions, so that the positive or negative charges are
averaged in all directions alike, the body appears neutral. If an external stress
comes into play and the molecules possess a certain degree of mobility, they will
gradually turn in the same direction, and, little by little, one side will become positive
and the other negative. The greater the controlling stress the more marked is the
change. If the controlling force be withdrawn, the molecules gradually return
more or less to their former position and the body remains more or less magnetized
accordingly. Glass and many liquids were found by J. Kerr to become optically
anisotropic or temporarily doubly refracting when acted upon by the poles of a
high voltage electric machine — J . Kerr's phenomenon. A parallel phenomenon is
produced by magnetic forces. Q. Majorana found that colloidal solutions of iron
salts exhibited this phenomenon, and A. Cotton and H. Mouton obtained similar
results. The explanation of the phenomena turns on the assumption that electric
charges are accumulated on the molecules which, in consequence, possess electric
poles so that the electric or magnetic field produces a slight change in the
direction of the molecules, and the optical effect is the secondary result of the
orientation of the molecules.
Again, when certain crystals are heated or cooled they exhibit positive and nega-
tive electrical charges — pyro-electricity — and the strength of the charge is propor-
tional to the variation in temperature. R. J. Hauy was much impressed with the
far-reaching consequences of this phenomenon. He said :
I do not know whether there is anything better calculated to excite the interest of
physicists than these minute electrical instruments built up by crystallization and confined
within the compass of a crystal not more than 2 mm. in thickness.
Similarly when certain crystals are compressed or expanded, they undergo a change
of form, and also exhibit positive and negative electrical charges — piezo-electricity
— and the strength of the charge is proportional to the degree of elastic deformation.
It is here assumed that a variation in the distance apart of the molecules caused by
changes of temperature or pressure disturbs the orientation of the polar molecules,
and the slight mobility of the molecules enables a small excess of them to orient them-
selves temporarily in a particular direction. If the temperature be raised sufficiently
high, the resulting increase in the movements of the molecules stirs them up, so to
speak, and tends to destroy the temporarily established order. Thus, a magnetized
body may be demagnetized, and piezo- and pyro-electricity may disappear. In
P. Lenard's opinion, also, the film next to the free surface of a liquid is a double
layer of positive and negative charges which he explains by assuming that all the
molecules in the surface present their electrical poles of the same kind towards the
exterior, like, adds A. Perrier, the quills of a porcupine. The hypothesis here
assumed is that the constituent particles of a body are dissymetrical in possessing
electrical and magnetic poles, and that these particles are capable of revolving in
such a way that they can occupy more or less parallel positions in particular direc-
tions ; as a result, the body acquires special properties in these privileged directions.
Dissymmetry, said P. Curie, is necessary for the production of the phenomenon.
Si cette dissymetrie n'existe pas, le phenornene est impossible.
In conformity with the above, 0. Lehmann's liquid crystals can be regarded
as microscopic systems in which the optical and magnetic properties show that
CRYSTALS AND CRYSTALLIZATION 649
certain groups of the constituent molecules have oriented themselves spontaneously
in certain definite directions. According to E. Bose's swarm theory,io if two or more
elongated molecules approach so closely that the mean distances of their centres of
gravity are less than half the length of the molecule, all free rotation must cease
except about the direction of elongation, and the molecules will tend to orient
themselves in parallel formations. A swarm of such molecules disposed in a parallel
direction, and in a perfectly fluid condition, without any suspicion of a space-lattice
arrangement, will possess the symmetry of a figure of rotation and behave as a
uniaxial crystal. Each swarm of such molecules will appear as a clear transparent
liquid crystal ; and the turbidity of a large mass of liquid is simply due to the
reflection and diffusion- of light at the mutual boundaries of the swarms. The
average size of the swarms will decrease on heating, and the point at which the
swarms become smaller than the wave-length of light will be the clearing point ;
above this temperature the liquid will appear singly refracting. The idea is illus-
trated by Figs. 85 and 86. D. Vorlander prepared a number of complex organic
compounds which when melted on a glass slide yielded a clear mass of a doubly
refracting liquid which in converging light gave a normal uniaxial interference
figure, Fig. 87, and, if the substance has an enantiomorphous molecular configura-
tion, the interference figure shows rotatory polarization. If thicker than 0*3 mm.
the masses may become turbid. The indices of refraction for the ordinary and
extraordinary rays have also been determined for ethylbenzylideneamine a-methyl-
and a-ethyl- cinna mates by E. Dorn and W. Lohmann. The change from a
solid crystal to an anisotropic liquid is regarded by H. W. B. Roozeboom as a
polymorphous transition point, and the clearing point as a true melting point ;
nil
Fig. 85.^ — ^Molecules disposed Fig. 86. — Molecules swarming Fig. 87. — Molecular Arrange-
in all Directions, into Parallel Groups. mentforD. Vorlander' sUni-
axial Interference Figures.
if the liquid crystals are merely liquids with groups or swarms of molecules
aligned in parallel, H. W. B. Roozeboom's transition point must be regarded as a
true melting point ; and his melting point as the temperature at which the average
size of the swarms becomes smaller than the wave-length of light.
L. Pucciante and R. Schenck found the viscosity of the crystalline liquid to be
usually less than that of the isotropic liquid. E. Bose found that the viscosity of an
anisotropic liquid increases quite normally with a fall of temperature down to the
point where the turbidity appears, and then decreases very sharply to a minimum
at 2° below the clearing point. The viscosity then begins to suffer the normal
increase with a fall of temperature, but not to the extent of the isotropic liquid.
The lower value of the viscosity of an anisotropic liquid as compared with that of an
isotropic liquid has been called the viscosity anomaly. E. Bose and F. Courat
showed that if the form of the molecule be that of an elongated ellipsoid of rotation,
the viscosity of a swarm can theoretically fall to two-thirds the value of the same
substance in an isotropic condition, the ratio of the two viscosities for anisaldazine
is anisotropic : isotropic=0*65. The change from an isotropic to an anisotropic
is always attended by a moderate increase in density, corresponding with the
closer packing possible with elongated molecules.
It will thus be observed that the question whether liquid crystals really have a
crystalline structure turns on the definition of a crystal. 0. Lehmann, in his
Veher die Definition des Begriffes Kry stall (1890), regards the regular arrangement
of the particles into a homogeneous anisotropic solid to be an accidental and not an
essential feature. It is highly probable that the molecules of liquid crystals are
grouped in swarms in each of which there is a definite alignment ; but if a crystal
650 INORGANIC AND THEORETICAL CHEMISTRY
is a homogeneous mass of material, arranged on a space-lattice pattern, the term
liquid crystal is probably a misnomer. Hence, wrote H. A. Miers :
It will be wise to retain the names crystal and crystalline in their old signification, rather
than to extend them so as to include the birefringent liquids whose existence has been
established by Lehmann.
Several of the characteristic tests for a crystalline structure are quite inapplicable
— e.g. elasticity and cleavage. D. Vorlander says that no biaxial liquid crystals
are known, while some fifty are definitely known to be uniaxial, and these are all
complex organic compounds. Nearly all compounds with a complex molecular
structure crystallize in a biaxial systems. These facts, said T. V. Barker, are of
superlative importance inasmuch as they present a statistical proof that the structure
of the so-called liquid crystals is not crystalline.
The whole of the firmly established properties — dichroism, and the invariable straight
extinction when the " crystal " is resting on a " prism " face, the apparent absence of
double refraction in parallel light, and the perfect uniaxial figure in convergent light when
the " crystal " is resting on its base — are in complete harmony with a structure analogous
to that of an even-grained piece of wood.
The effect of an electromagnetic field on liquid crystals is said by W. Nernst n to
furnish a decisive proof of the correctness of the swarm theory. 0. Lehmann first
noted that a magnetic field clears the droplets of jo-azoxyanisole ; and E. Bose
noted that the effect can be perceived with the application of 600 Gauss units of force,
and with a few thousand units, the layers of liquid 4 mm. thick are immediately
cleared when the liquid is viewed along the lines of force ; if the current be cut off,
the liquid again becomes turbid. The phenomenon thus interpreted recalls Weber's
theory of magnetization. H. von Wartenberg and C. Mauguin extended the work,
and found that when a homogeneous film is exposed to a transverse magnetic
field, the optic axis is gradually deflected in the plane containing the lines of force,
but on releasing the force, the optic axis immediately returns to its normal position.
References.
1 F. Reinitzer, Sitzber. Akad. Wien,9^. 719, 1886; 97. 167, 1888; Monatsh., 9. 42.1, 1888;
L. Gattermann and A. Ritsehke, Ber., 23. 1738, 1890 ; 0. Lehmann, Zeit. KrysL, 1. 120, 492,
1877 ; 18. 457, 1890 ; Zeit. phys. Chem., 4. 462, 1889 ; 5. 427, 1890 ; 18. 90, 1895 ; Wied. Ann.,
40. 401, 1890 ; 41. 525, 1890 ; 53. 632, 1894 ; 56. 771, 1895 ; Ann. Physik, (4), 2. 629, 1900 ; (4),
5. 236, 1901 ; (4), 8. 908, 1902 ; (4), 12. 311, 1903 ; Zeit. Kryst., 18. 457, 1890 ; 52. 597, 1913 ;
Flussige Krystalle, Leipzig, 1904; Die neue Welt der fliissigen Krystalle, Leipzig, 1911;
T. M. liOwry, Annual Reports of the Progress of Chemistry, 6, 1910 ; D, Vorlander, Kristallinisch-
flussige Substanzen, Stuttgart, 1908 ; Phys. Zeit., 15. 141, 707, 1914 ; R. Schenck, Kristallinisch
Flussigkeit und flussige Kristalle, Leipzig, 1905 ; Ber., 41. 2033, 1908 ; H. A. Miers, Science Progress,
6. 119, 1896 ; E. Bose and F. Courat, Phys. Zeit., 9. 169, 1908. I am indebted to Prof. O. Lehmann
for permission to reproduce Figs. 82 and 83.
2 I. F. B.omira.y, Journ. Chem. Soc, 97. 1669, 1910 ; 0. Lehmann, Ann. Physik, (4), 21. 181,
1906 ; (4), 2. 661, 1900 ; (4), 18. 796, 1905 ; D. Vorlander, Ber., 41. 2033, 1908 ; E. Dorn, Phys.
Zeit., 11. 777, 1910 ; F. Wallerant, Compf. Bend., 143. 605, 694, 1906 ; 148. 1291, 1909 ; 1). Vor-
lander and W. Kasten, Ber., 31. 2033, 1908 ; E. Dorn and W. Lohraann, Ann. Phjsik, (4), 29.
533, 1909; F. Giesel, Phys. Zeit., 11. 192, 1910; T. Rotarsky, Ber.,^1. 1094, 1908; C. Mauguin, Compt.
Bend., 156. 1246, 1913, 1916 ; F. Grandjean, Bull. Soc. Min., 39. 164, 1916.
3 M. Faraday, Phil. Tran^., 147. 145, 1857.
* G. T. Beilby, Proc. Roy. Soc, 72. 226, 1903 ; Journ. Soc. Chem. Ind., 22. 1166, 1903 ; Trans.
Optical Soc, 22. 1907 ; J. C. M. Gamett, Phil. Trans., 203. A, 385, 1904 ; 205. A, 237, 1905.
^ G. Tammann, Zeit. phys. Chem., 18. 91, 1895 ; G. Friedel and F. Grandjean, Bull. Soc Min.,
33. 192, 409, 466, 1910 ; C. Mauguin, Compt. Rend., 154, 1359, 1912.
® H. W. B. Roozeboom, Die heterogenen Gleichgewichte vom Standjmnkte der Phasenlehre,
Braunschweig, 1. 152, 1901.
' G. Tammann, Ann. Physik, (4), 4. 524, 1901 ; (4), 8. 103, 1902 ; (4), 19. 421, 1905 ; O. Leh-
mann, i6., (4), 50. 599, 1916; (4), 48. 725, 1915; (4), 51. 353, 1916; (4), 52. 445, 736, 1917; (4), 56.
81, 1918; (4), 56. 321, 1918; (4), 57. 243, 1918; (4), 58. 631, 1919; Arch. Sciences Geneve, {4), S2. 5,
1911 ; Verh. deut. phys. Ges., 20. 63, 1918 ; Sitzber. Akad. Heidelberg, 29, 1911 ; Phys. Zeit., 17. 241,
1916; 19. 73, 88. 1918; G. Friedel and F. Grandjean, Compt. Rend., 151. 327, 1910 ; Bull. Soc
Min., 40. 69, 1917 ; A. E. H. Tutton, Crystallography, London, 931, 1911 ; G. Bredig and G. von
Schukowsky, Ber., 37. 3419, 1904 ; A. Cohn, Zeit. Elektrochem., 10. 866, 1904 ; R. Schenck, Ann.
CRYSTALS AND CRYSTALLIZATION 651
Phyaik, (4), 9. 1063, 1902 ; Zeit. phys. Chem., 25. 337, 1898 ; C. de Kock, ib., 48. 129, 1904 ;
G. Quincke, Wied. Ann., 53. 013, 1894 ; D. Vorlander, Ecr., 31. 2033, 1908 ; O. Lehmann, Phys.
Zeit., 11. 44, 1910 ; E. Bose and F. Courat, ib., 9. 169, 1908 ; H. Stoltzenberg and M. E. Huth,
Zeit. phys. Chem., 71. 641, 1910 ; C. Tubandt and E. Lorenz, ib., 87. 627, 1914 ; P. N. Pawloff,
Journ. Russian Phys. Chem. Soc, 41. 685, 1909 ; G. WuifF, Zeit. Kryst., 46. 261, 1909 ; W. Voigt,
Phys. Zeit., 17. 76, 128, 152, 305, 1916.
8 H. le ChateUer, Rev. Met., 3. 105, 1906 ; H. C. H. Carpenter with C. A. Edwards, Proc. Inst.
Mech. Eng., 164, 1907.
» P. Curie, Bull. Soc. Min., 7. 89, 418, 1884 ; Journ. Phys., (3), 3. 393, 1894 ; (Euvres, Pans,
56, 78, 118, 1908 ; A. Perrier, Scient. Amer. Suppl, 87. 18, 46, 1919 ; P. Lenard, Ann. Physik, (4),
47. 463, 1915 ; A. Cotton and A. Mouton, Ann. Chim. Phys., (8), 11. 145, 289, 1907 ; (8), 20. 194,
1910 ; Les ultramicroscopes et les objets ultramicroscopiques, Paris, 1906 ; Q. Majorana, Rend.
Accad. Limei, 11. i, 374, 463, 531, 1902 ; 11. ii, 90, 139, 1902 ; W. Weber, Abhand. Sachs. Ges., 1.
485, 1852 ; J. Kerr, Phil. Mag., (5), 3. 321, 1877 ; (5), 5. 161, 1878 ; A. Bravais, Journ. VEcole
Pohjt., 19. 127, 1850 ; 20. 102, 197, 1851 ; R J. Hauy, Traite de mineralogie, Paris, 2. 343, 1801.
i» E. Bose, Phys. Zeit., 8. 347, 513, 1907 ; 9. 708, 1908 ; 10. 32, 230, 1909 ; 12. 60, 1911 ;
E. Bose andF. Courat, ib., 9. 169, 1908 ; T. V. Barker, Annual Reports of the Progress of Chemistry,
11. 260, 1915 ; 14. 239, 1918; Y. Bjornstahl, Ann. Physik, (4), 56. 161, 1918.
11 W. Nernst, Zeit. Elekirochem., 16. 702, 1910 ; E. Bose, Phys. Zeit., 12. 60, 1911 ; H. von
Wartenberg, ib., 12. 837, 1230, 1911 ; C. Mauguin, Compt. Rend., 152. 1680, 1911.
§ 10. Isomorphism— Mitscherlich's Isomorphic Law
The laws of nature represent design ; they are embodied design.- — P. Cakus.
Dans les substances cristallisea, la forme des molecules integrantes, et, par suite, des
cristaux, depend du nombre et de la position respective des atomes dont les molecules sont
composees.' — A. M. AmpIire (1814).
According to E. Wohlwill,i Basil Valentine prepared mixed crystals of the
vitriols, and the mixed crystals were mentioned about the middle of the eighteenth
century by J. F. Henkel and by A. G. Monnet. In 1772, J. B. L. Kome de I'lsle
noticed that a mixture of copper and iron sulphates furnishes crystals in the form
characteristic of iron not copper sulphate, and in 1787, N. Leblanc 2 made the same
observation. This appeared to be an exception to D. Guglielmini's generalization
or Haiiy's law — that the angles between similar faces of the crystals of a given
substance are characteristic of one definite compound. Analogous results were
observed with crystals from mixed solutions of iron and aluminium potash and
ammonia alums ; many minerals also are almost identical in crystalline form
through possessing a different chemical composition. In 1801, N. Leblanc con-
firmed J. B. L. Rome de I'lsle's observation and found many other examples.
He noted that the aluminium of alum could be replaced by iron without altering
the crystalline form. In 1797, L. N. Vauquelin also noted that the potassium in
alum could be replaced by ammonium, without changing the crystalline form ;
and in 1816, J. L. fray Lussac found that crystals of very different composition
could be obtained from mixed solutions of different alums. A. Bernhardi (1809)
investigated mixed crystals, and found that different substances can have the same
crystalline form — e.g. magnesium and zinc vitriols, and he observed that in crystal-
lizing a mixture of copper and iron vitriols the one can so influence the other that
the two salts crystallize in the same form, and C. F. Bucholz proved by analysis
that there must be at least 13 parts of copper vitriol to 87 of zinc vitriol in order
to give to the mixed crystals the peculiar form of the former salt. F. S. Beudant
made important contributions to the subject about 1817. W. H. WoUaston investi-
gated the mixed crystals of zinc and copper vitriols in 1818, and a year later, B.
de Villiers, in his De la cristallisation (Strasbourg, 1819), also discussed this question.
The analyses of M. H. Klaproth, L. N. Vauquelin, P. Berthier, and others showed
that the variable composition of mixed crystals applies not only to laboratory
preparations but also to numerous minerals. J. N. von Fuchs (1815) also showed
that certain constituents of a compound can be replaced by other so-called vicarious
constituents — vicariende Bestdndtheile — without altering its general character ;
for example, gehlenite is essentially calcium aluminium silicate, 3CaO.Al203.2Si02,
652 INORGANIC AND THEORETICAL CHEMISTRY
and yet the calcium can be more or less replaced by magnesium or ferrous iron,
and the aluminium by ferric iron, without changing the general physical properties
of the mineral. In the BerthoUet-Proust controversy, C. L. BerthoUet argued
from such examples that chemical compounds may have a variable composition,
while J, L. Proust argued that these crystals are really mechanical mixtures ;
R. J. Haiiy maintained that when two substances crystallize together in this manner
the dominant crystalline form will be characteristic of the component which is in
excess. R. J. Haiiy explained the results by postulating that one constituent
might determine the crystal-form of a substance even though present in very small
amounts, while the other constituent remained without influence on the crystal
form.
E. Mitscherlich's investigations opened up the subject in a most interesting
manner. His work is recorded in memoirs : On the relations between chemical
composition and crystalline form, published in Sweden between 1818-1821.3 While
making preparations of the arsenates and phosphates of potassium and ammonium,
he noticed that the crystals were so like each other as to be indistinguishable by
simple inspection ; and a closer examination led E. Mitscherlich to conclude :
(1) That bodies of different chemical composition may have the same crystalline
form; (2) substances of similar constitution have the same crystalline form.
E. Mitscherlich wrote :
The same number of atoms combined in the same manner produce the same crystaUine
form ; the crystalline form is independent of the chemical nature of the atoms, and is
determined solely by their number and mode of combination.
This relation is now known as Mitscherlich's law. He noticed that the acid
arsenates and phosphates of potassium, sodium, or ammonium crystallize in similar
tetragonal forms, Fig. 7 (left), that one element or groups of elements may be
exchanged for another which appears to act in an analogous manner. Thus arsenic
may be exchanged for phosphorus, and potassium for ammonium without affecting
the form of the crystal. In Mitscherlich's words :
Every arsenate has its corresponding phosphate, composed according to the same
proportions, combined with the same amount of water of crystallization, and endowed
with the same physical properties : in fact, the two series of salts differ in no respect,
except that the radicle of the acid in the one series is phosphorus, while in the other it is
arsenic.
It is not difficult to understand how atoms of different elements may be so related
that they can be mutually interchanged without altering the crystalline form and
general character of the compound. The idea is illustrated by an old simile : the
lines of a tesselated pavement are not altered if the blue tiles are replaced partly
by red or by green ones, so long as the different-coloured tiles retain the original
size and shape of those they replace.
Besides the phosphates and arsenates, Mitscherlich observed that a certain
group of mineral carbonates — calcite, CaCOs ; dolomite, CaMg(C03)2 I siderit^
or chalybite, FeCOs '> calamine or smithsonite, ZnCOs ; and dialogite or rhodocrosite,
MnCOs — 3,11 form isomorphous crystals in the trigonal system (Fig. 88), and
an application of the X-ray spectrum has enabled W. L. Bragg (1914) to demonstrate
the structural similarity of rhodochrosite, chalybite, and dolomite. Again, the
mineral sulphates— barytes, BaSO^ ; celestine, SrS04 ; and anglesite, PbS04—
all form similar rhombic crystals ; while aragonite, CaCOs ; witherite, BaCOa ;
strontianite, SrCOs ; and cerussite, PbCOs, form isomorphous rhombic crystals.
Numerous other examples could be quoted. E. Mitscherlich applied the term
isomorphism — from la-o^, equal ; li-op^jiiq, shape — to connote the fact that ana-
logous elements can replace one another without affecting the apparent shape
of^ the crystals. It is therefore inferred that in a crystalline solid, each con-
stituent atom occupies a certain domain or portion of the space occupied by
the whole molecule.
CRYSTALS AND CRYSTALLIZATION 653
E. Mitscherlich's law of isomorphism, as well as the phenomena of polymorphism,
appear to contradict R. J. Hauy's law, and there was some reluctance in France to
accept E. Mitscherlich's conclusions. For instance, in some Reflexions sur le
memoire de M. Mitscherlich qu'on etait recueillies dans une conversation avec M. Haiiy
'par un de ses eleves,^ it is said that Haiiy considered : Si la theorie de M. Mitscherlich
etait juste, la mineralogie serait la plus pitoyahle des sciences. R. J. Haiiy, ^ however,
specially remarked on the crystallographic resemblances between certain minerals
like barytes and strontianite by saying :
There is almost an identical primitive form- — noyau- — in each, and the crystals themselves
furnish ocular resemblances, so that they may be compared with what botanists term
family resemblances — air defamille.
E. Mitscherlich, no doubt, had in mind absolute identity of crystal form as the
basal principle of isomorphism, but later investigations have shown that the
crystals of isomorphous substances are nearly but not absolutely identical, but
only similar in form ; and thus confirmed the earlier observations of W. H. WoUaston
(1812) 6 that the rhombohedral cleavage angles of the native carbonates of the
calcite series are nearly but not quite the same ; and E. L. Malus that :
Des recherches posterieures des nous enseigneront comment cette loi g^nerale sera
modifi6e par la petite difference qui se trouve quelquefois dans les angles des combinaisons
isomorphes.
There axe small but real differences in similar interfacial angles of the members
of an isomorphous series of compounds. For example, the corresponding angles
P (Fig. 88) of the following isomorphous carbonates of the calcite series are far
from identical, even if they are approximately similar :
Calcite, Dialogite, Chalyblte, Magnesite, Smithsonite,
CaCOg MnCOg FeCOg MgCO, ZnCO,
Angle . . . 105° 5' 106^ 51' 107° 0' 107° 20' 107° 40'
The idea will perhaps be clear from Fig. 88, where the change in the interfacial
angle P in passing from calcite to smithsonite is shown in section. With the
carbonates of the aragonite series, the angles between the prism faces are :
Aragonite,
Strontianite,
Witherite,
Cerussite,
CaCOg
SrCOg
BaCOg
PbCOg
Angle
. 116° 10'
117° 18'
117° 48'
117° 18'
Axial ratio a :b: c
; 0-623: 1:0-721
0-609 : 1 : 0-724
0-595 : 1 : 0*741
0-610: 1 : 0-723
Specific gravity
. 2-95
3-74
4-32
6-60
Specific heat
0-1992
0-1445
0-1078
0-0814
Molecular heat .
19-66-20-18
21-31
21-34
21-73
and for the isomorphous sulphates of the barytes series, the angle of the rhombic
prisms are respectively 101° 46', 103° 48', and 104° 11' with barytes, BaS04, anglesite,
PbS04, and celestine, SrS04. With the isomor-
phous sulphates, ZnS04.7H20, MgSO^.THgO, and
NiS04.7H20, the angles are respectively 89°
22', 89° 26', and 89° 56'; and with the trigonal
series : potassium platinate, K2Pt{011)^, stannate,
K2Sn(0H)6, and plumbate, K2Pb(0H)6,the angles
are respectively 74° 18', 75° 14' and 75° 19'.
A. E. H. Tutton ^ found that in the isomor-
phous selenates and sulphates of potassium, rubi- ^^^ 88.— Diagramatic Representa-
dium, and caesium, specific chemical replacements tion of the Variation In the Angle
are accompanied by clearly defined changes in the P of the Isomorphous Carbonates,
crystal structure along specific directions. Thus,
when the basic element, say, potassium, in an alkaline sulphate or selenate is
replaced by another of the same alkali family group, rubidium or caesium, the
greatest alteration occurs in the crystal angles corresponding with an elongation
654 INORGANIC AND THEORETICAL CHEMISTRY
of the vertical axis ; and when the acid-forming element sulphur is replaced by
selenium, its family analogue, the greatest expansion takes place along the horizontal
axes of the crystals. A. E. H. Tutton's diagram. Fig. 89, shows in an exaggerated
manner, the effect of replacing potassium in potassium sulphate oj selenate by the
basic elements rubidium and caesium.
Ehitropic series, — According to A. E. H. Tutton, in a strictly isomorphous series,
where the interchangeable elements belong to the satne family group of the periodic
classijicaiion^ the whole of the properties of the crystals — 7norphological, optical, thermal,
and physical — -are, in general, functions of the atomic weights of these elements, and
for the purpose of emphasizing the closeness of the relations connecting the several
members, it is called a eutropic series — evrpoiras, well nourished. Thus, thallium
sulphate and selenate, and ammonium sulphate are isomorphous with the potassium,
rubidium, and caesium sulphates and selenates, because the radicle thallium or
ammonium can replace the alkali metal without causing angular or structural
changes greater than those produced by an interchange of the same family of
elements. All the salts bear some definite chemical analogy, and crystallize in
the rhombic system in forms whose angles rarely differ by more than 3°. In
addition, the members of a eutropic series are not only isomorphous, but the inter-
changeable radicles belong to the same family group — e.g. the thallium, ammonium,
potassium, rubidium, and caesium sulphates from an isomorphous series, but the
two former are not included in the eutropic series formed by the three latter ;
aragonite, strontianite, witherite, and cerussite form an iso-
y ^ morphous series, but the last is excluded from the eutropic
series.
F. M. Jager ^ similarly investigated the hexagonal crystals
of the isomorphous rare earth ethyl sulphates of the type
R'"(EtS04)3.9H20, in which R'" denotes yttrium, Ian-
thanum, cerium, praseodymium, neodymium, samarium,
europium, gadolinium, dysprosium, erbium, thulium, and
neo-ytterbium. The variation in passing from one member
of the series to another is but a few minutes, so that the
probable value of c : a for the whole series is c : a=0"5062
Fig. 89. ±0'0012. The molecular volumes are distinctive. A. E. H.
Tutton has shown that F. M. Jager's results might have
been anticipated if the factors operating towards the extreme closeness of the
angular values bp considered : (i) Small variations in the atomic weights of the
elements concerned ; (ii) the mass effect of the remainder of the molecule ; and
(iii) the high symmetry of the hexagonal system.
The ratios in an isomorphous series whose members are not related eutropically,
do not stand as an arithmetical or harmonic series of integral numbers. G. Linck ^
found empirically that if F denotes the crystal volume, D the specific gravity, and
M the molecular weight, the quotients VD/M of the members of a eutropic series
are related with one another as an arithmetical or harmonic series. This is
illustrated in Table III.
Morphotropic series. — ^According to the structural theories of E. von Federcff,
A. Schonflies, and W. Barlow space is partitioned into space units, space lattices,
or elementary cells, Fig. 90, which E. von Federoff called polyhedra, A. Schonflies,
FundamerUalbereich, and W. Barlow, spheres of influence. Further, following H. A.
Miers, analogous portions of matter are supposed to be distributed in each space
unit. No hypothesis is made as to the characteristics of these portions of matter :
nor of the arrangement of the atoms in the molecules. Each space lattice is
considered to be made up of units or points which represent either the centres of
gravity of the constituent molecules, or the centres of rest about which those
centres of gravity oscillate. The physical properties of crystals make it clear
that whatever be the nature of the vibratory motions of the molecules, the move-
ment does not take place outside a certain imaginary ellipsoidal domain or sphere
CRYSTALS AND CRYSTALLIZATION 655
of influence. Consequently, the molecules can then be discussed as if they were
arranged like a system of points at rest. On this view, crystals are regarded as
aggregates of ellipsoids or spheres, piled up in such a way that the corresponding
axes are arranged in accord with some definite geometrical plan. Each molecule
then appropriates to itself a space equal to one space unit— illustrated by the
heavier lines in Fig. 90. It can then be assumed that the volume of each space
Table III. — Eutbopic Series of Crystals.
Salts.
Axis ratios,
a:b:c
Crystal vol.
V
Specific
gravity,
D
Molecular
weight,
M
VD/M
Ratio.
K2SO4
CS2SO4
Rhombic sulphates.
0-5727 : 1 : 0-7418
0-5723 : 1 : 0-7485
0-5712 : 1 : 07531
0-4248
0-4284
0-4302
2-666
3-615
4-246
174-4
267-1
361-9
0-006496
0-005798
0-005048
9
8
7
Ca(N03)2
Sr(N03)2
Ba(N03)2
Cubic nitrates
1
1
1
2-6440
2-9857
3-2435
164-08
211-68
261-48
0-016116
0-012404
0-012404
K2Mg(S04)26H20.
Rb2Mg(SOj26H20
Cs2Mg(SO4)26H20
Monoclinic magnesium sulphates.
0-7413 : 1 : 0-4993
)8 = 75° 12'
0-7400 : 1 : 0-4975
^ = 74° 1'
0-7279 : 1 : 0-4946
jS = 72° 54'
0-35784
0-35391
0-34410
2-028
2-382
2-670
402-9
495-4
590-6
0-00180121 36
0-0017017 34
0001556
31
unit can be represented by the quotient of the molecular weight by the molecular
volume. W. Muthmann,io F. Becke, and A. E. H. Tutton attempted to determine
the structure or rather the relative distances between homologous points in the
space lattices of known crystals by measuring the relative distances of the crystal
molecules from each other along different directions in a series of isomorphous
crystals. P. Groth had previously emphasized the fact that on substituting a
univalent atom or radicle in place of hydrogen, a change in the form of the crystal
may take place in a particular direction, and he
called the phenomenon morphotropy. A related
phenomenon was noticed by A. Laurent (1840),
and by F. de la Provostaye (1870), when chlorine
was substituted for hydrogen in certain organic
compounds — e.g. naphthalene, etc. The subject
also received the attention of L. J. Wallmark,
W. G. Hankel, L. Bodart, J. Nickles, T. von o3
Alth, L. Pasteur, etc. Numbers representing the ^^
relative dimensions of the space units can be Fia. 90. — Space Lattice,
derived from measurements of the crystallo-
graphic axes, etc. The axial ratios a:b:c are used in describing the form of a
crystalline substance, and they usually change, more or less, in passing from one
substance to another ; but since the axial ratios of a substance represent ratios of
the actual dimensions of the corresponding homogeneous structure, they do
not indicate the change in dimensions which occurs on passing from one sub-
stance to the other. It has therefore been found convenient to link up the
axial ratios a: h : c with the molecular volume V of the substance so as to
furnish ratios x» ^> ^> which are proportional to the ratios a:h:c, and
656 INORGANIC AND THEORETICAL CHEMISTRY
which represent the linear dimensions of the elementary cell, Fig. 90. Let x, «A>
and o) represent the lengths of the sides of a unit cell ; and V its molecular
volume ; further, for the sake of simplicity, suppose the cell be rectangular, and
let the directions of the sides coincide with the crystallographic axes. Then
F=x^a>=i¥/Z), where M denotes the molecular weight of the substance, and D
its specific gravity. Hence x? ^, and co may be taken to represent molecular
intervals along the edges of the cell. Let a, h, and c represent the crystallographic
axes coincident with the edges of the rectangular cell, then, by a well-known
theorem in trigonometry,
X ilf (X) ^ 3/ V
Consequently,
c2F
h
Consequently, the relative dimensions of the space units in a series of related crystals
can be calculated from measurements of the crystallographic axes, etc. The values
X, j/f, CO so determined are called the topic parameters, or topic axes, or
molecular distance ratios, and for a series of related substances, they represent the
changes in the molecular magnitude of corresponding translations of the homogeneous
structure common to the substances which occur in passing from one member of
the series to another. n
The topic characters are calculated in an analogous way for crystals belonging to other
systems. If the angle between the lines w and iff he a; between cu and x be j8 ; and
between x ^^^ ^ ^^ y, then, for a triclinic crystal, F = x^co sin a sin ^ sin y ; for^ cubic
crystal, 0= 6= c, and a=j8=y = 90° and sin 90° is unity, therefore x=^=^=Vl^; and
for a monoclinic crystal, a=')/ = 90°, so that V=x^^ ^^^ ^«
It is sometimes convenient to represent the cubic capacity- — crystal volume — of a
solid calculated from the axes ratios a, 6, c and the angles a, j8, y of a crystal. In the
cubic system, where a=b=c, and a = jS = y = 90°, the crystal volume is unity; in the
rhombic system, where a<^6^c, and a=j3=y = 90°, the crystal volume is ac if 6 = 1 ; db
if c = l ; and be if a — 1 ; in the tetragonal system, where a=6$c, and a = jS = y = 90°, the
crystal volume is a^, if c = l, and c if a = l ; in the hexagonal system,, where a = b^c, and
a=j3 = 90°, and y = 60°, the crystal volume is ialJS, if c = l ; and ^c ^JS, if a = l ; in the
monoclinic system, where a^&^c, and a = y = 90°, and j8^90° ; the crystal volume is
ac sin j3, if 6 = 1 ; db sin ^, if c = 1 ; and be sin ^, if a = l ; ~ and in the triclinic system,
where a^b^c, and a^^^y, the crystal volume is 2acsj^, if 6 = 1 ; 2ab ^Jl, if c = l ; and
2bc^, if a = l, where /x is put in place of sin 5. sin (s — a).sin (s — jS).sin (s — y).
In his Beitrdge sur Volumetheorie der Jcristallisierten Kdrper,W. Muthmann (1894) 12
calculated the topic axes of the acid phosphates and arsenates of potassium and
ammonium, and also of the alkali permanganates. The relation between the axial
ratios and the molecular distance ratios of the latter are :
KMnO^ .
RbMnO* .
C8Mn04 .
NH4Mn04.
Hence, while the axial ratios measure only the relative distances of translations,
in homogeneous structure in the case of one substance, because one axial dimension,
h, is taken as unity, the topic axes indicate the relative dimensions of corresponding
translations in the several members of an isomorphous series ; W. Muthmann con-
cluded that with the alkali permanganates the differences between corresponding
molecular distance ratios of the various salts indicate that the unit of crystalline
structure is composed of four chemical molecules ; but T. V. Barker's results with
the alkali perchlorates, isomorphous with the permanganates, did not agree with
a :
: c
M
X '• "^ '. 0
0-7972 :
: 0-6491
58-526
3-8554
: 4-8360 :
: 3-1390
0-8311 :
: 0-6662
63-228
4-0322
: 4-8517;
: 3-2312
0-8683 ;
: 0-6853
70-042
4-2555
: 4-9009 :
: 3-3584
0-8164
: 0-6584
62-126
3-9767
: 4-8711
: 3-2071
CEYSTALS AND CKYSTALLIZATION 657
W. Muthmann's conclusions. A. E. H. Tutton likewise calculated the topic para-
meters of the alkali sulphates and selenates ; J. A. le Bel and A. Ries, of the
substituted ammonium chloroplatinates ; G. Mez, of the derivatives of carbamide ;
etc.
If Xi' ^i5 ^1 > X2> 'A2> ^2 5 • • • ^® t^® topic axes of an isomorphous series in
which it is assumed that the molecules are similar, that the arrangement of the
molecules is similar, and that the crystals have the same elementary parallelo-
pipedal cells which vary slightly in dimensions in passing from one series to the
other, then, xv ^i? ^i give the relative molecular intervals along three directions
in the crystals of the one substance ; x2j ^2» ^2j ^^^ molecular intervals along
three corresponding directions in the second substance. The ratio ipi : ijj^
represents the relative increased or decreased separation along the given direction
in molecules of the two different substances, owing to the replacement of one element
or radicle by another in the series. In illustration, W. Muthmann found that the
molecules of the tetragonal phosphates separate almost uniformly in all directions
when the atom P is replaced by an atom of As. When the K atom is replaced by
the NH4 radicle in either the tetragonal phosphate or arsenates, the molecules
are again separated, but almost entirely in the direction of the principal axis.
Hence, concludes W. Muthmann, the metallic elements occupy such a position in
the molecule that the line uniting them to the acid radicles are parallel to the point
axis.
If the symmetry of the crystal molecules be tetragonal like that of the crystal
each physical molecule will be a complex cluster of at least eight chemical molecules,
say KH2PO4, and he assumes that in this complex, eight P0(0H)2 radicles are
arranged at the corners of two superposed horizontal squares and a KG radicle is
attached above or below each P0(0H)2 radicle. A. E. H. Tutton employed analogous
reasoning for the rhombic crystals of the alkali sulphates, and he considered the
accord justifies the assumption that the crystal elements of the alkali sulphates
are situated at the corners of a rectangular rho^lbic prism so that each cluster
consists of four chemical molecules arranged in a definite symmetrical manner.
The molecule of a crystal may thus include several chemical molecules, and G. J.
Stoney ^^ proposed to call the former macromolecules to distinguish them from the
latter. W. Muthmann predicted that thallium and rubidium sulphates would
have almost identical forms, and this prediction was later verified by A. E. H.
Tutton. F. Slavik i^ has calculated the topic axes of the morphotropic series
ammonium iodide, NH4I, tetramethyl ammonium iodide, N(CH3)4l, tetraethyl
ammonium iodide, N(C2H5)4l ; and tetrapropyl-ammonium iodide, N(C3H7)4l,
and fouid on writing Me for CH3 ; Et for C2H5 ; and Pr for C3H7 :
Mol. vol.
a-.hi
c
x:^':'-
NH4I (cubic) .
57-51
1:1;
: 1
3-860 : 3-860 :
: 3-860
NMe4l (tetragonal) .
. 108-70
1: 1
: 0-7223
5-319:5-319;
: 3-842
NEt4l (tetragonal) .
. 162-91
1: 1
: 0-5344
6-648 : 6-648 ;
; 3-686
NPrJ (rhombic)
. 235-95
0-776 : 1
: 0-6283
6-093 : 7-851 ;
: 4-933
Hence, while the value of co is almost the same in the first three cases, the values
of X and xjj are increased by substituting four-methyl groups in place of hydrogen,
and still more if four of the heavier ethyl-groups be introduced ; with the still
heavier propyl-groups more drastic changes take place in the spatial arrangement
of the atoms, and a very marked change in the molecular volume is that the propyl-
compound no longer possesses the same crystal symmetry.
C. A. Kenngott,i5 A. Schrauf , and F. Pfaff measured the relation between the
hardness and specific gravity of isomorphous bodies, and found the mean hardness
of the crystal faces to be related less definitely with the chemical composition than
other physical properties. K. R. Koch measured the elastic constants of the two
main alkali chlorides.
The cleavage of crystals is connected with their internal cohesion, and
G. Tschermak i^ and A. Sadenbeck have given a number of examples showing the
VOL. I. 2 u
658 INORGANIC AND THEORETICAL CHEMISTRY
analogy in the cleavages of isomorphous substances. H. Baumbauer, G. Tschermak,
and F. Becke ^^ have examined the corrosion figures of isomorphous compounds.
E. Jannettaz is has investigated the thermal conductivity of isomorphous crystals,
and he concludes that :
Les clivages les plus faciles ou leurs r^sultantes sont parall^les aux axes les plus grands
de conductibilite th^nnique, et inversement, m^ine, dans les cas de trois clivages.
The thermal expansion of some isomorphous substances has been measured by
H. Fizeau and F. PfafE.i^ The magnetic properties of isomorphous substances have
been studied by J. Grailich and V. von Lang,20 and some analogies were observed.
Isomorphous bodies show close resemblances in their optical properties, although
there are some irregularities ; they have been compared with respect to the position
and length of their optical elasticity axes ; the index of refraction ; double
refraction, and dispersion. The pioneer work was done by H. de Senarmont,2i
J. Grailich, V. von Lang, and H. Topsoe and C. Christiansen. In spite of the fact
that H. de Senarmont found
The mechanical causes which determine the geometrical form are of a different order
from those which determine the optical properties, inasmuch as the form remains the
same in an entire series of isomorphous substances, whereas the optical properties show
not only fundamental variations, but a complete inversion in their relative magnitude ;
and J. Grailich and V. von Lang
Different substances cannot enter into the molecule without changing the form of the
crystal, but the optical qualities are more deeply affected the greater the change in the
constitution of the molecule. . . . There is no direct relation between the optical properties
and such properties as cleavage, hardness, and magnetic susceptibility which rest on the
different arrangement of the molecules ;
there are, however, many analogies between the optical properties— -double re-
fraction, index of refraction, etc. — of isomorphous bodies. The similarity in
external form is one sign that there is an analogy of structure, so that isomorphous
substances usually exhibit not only close chemical analogies but also close analogies
in their physical properties. It is, however, possible that compounds of very
different chemical composition have the same structure, and there is a risk in
using isomorphism in the attempt to establish chemical relations where none exist.
The law of mixed crystals. — Extended observations have multiplied examples
of substances which possess a similar chemical constitution and a similar crystalline
form ; but at the same time the observations have also brought into prominence
the fact that substances which crystallize in similar or identical forms — particularly
in the cubic system — may exhibit wide divergencies in chemical constitution.
The converse of E. Mitscherlich's law does not, therefore, hold good. Similarity
of chemical composition or similarity in crystalline form are not adequate tests
for isomorphism. E. Mitscherlich also stated that " while substances of different
crystalline form cannot combine other than in fixed proportions, substances of the
same crystalline form can crystallize together in all proportions." F. S. Beudant 22
was the first to suggest that " mixed crystals " are melanges chimiques ou associations
non mecaniques en proportion indefinie. He said :
I have adopted the expression melange chimique in order to distinguish by a specific
term a chemical association of bodies which has characteristics different from other
chemical associations to which the name comhinaisona chimiques is applied. I do not,
however, seek to imply that the components which can be associated in the melanges
chimiques in an infinitude of proportions, are really united chemically or simply mixed.
The m,elange8 chim^iques could be designated comhinaisona indefinies in which the product
always possesses, more or less, the properties of one or other of the components.
Homogeneous crystals containing two salts mixed in indefinite proportions,
and formed in solutions containing a mixture of both salts, were called Misch-
krystalk — i.e. mixed crystals — by H. W. B. Roozeboom in 1899 ; they have also been
called isomorphous mixtures, but both terms are liable to misconception because
CKYSTALS AND CRYSTALLIZATION 659
the mixed crystals are mixtures only in the sense that ordinary homogeneous solutions
are mixtures, and hence some prefer the older term solid solution, used by J. H.
van't HofE 23 i^ 1890, and regard crystals as homogeneous phases and not an aggre-
gate of two or more phases. L. de Boisbaudran claims that he first applied the idea
of solid solutions, dissolvents solides, in papers on fluorescence between 1886 and
1890, and that in an unpublished note — on supersaturation — before the Academie
des Sciences in Paris in 1866, he stated :
I do not hesitate to attribute the same cause, (i) to the solution obtained by dissolving
one soHd in another (isomorphisme de Mitscherlich) ; (ii) to the solution of liquids in
another one ; and (iii) to the solution of vapours in one another.
The formation of apparently homogeneous crystals is not accepted by chemists
as a decisive test of the individuality of a chemical species, since, in crystals, certain
elements may replace one another indefinitely without altering the form of the
crystals. When mixed crystals of lead and barium nitrate are treated with a
saturated solution of barium nitrate, the lead nitrate is dissolved out and a skeleton
of barium nitrate remains which is not doubly refracting although the original mixed
crystals exhibit this quality. The failure of mixed crystals to satisfy the law of
constant composition has led chemists to agree arbitrarily that mixed crystals are
mixtures and not definite compounds, although the phase rule regards mixed crystals
as homogeneous single phases.
.The colour of mixed crystals is usually intermediate between the colour of their
component salts — thus, yellow caesium chloroplumbate, Cs2PbCl6, and deep blue
caesium chloroantimoniate, Cs2SbCl6, give mixed crystals of a green colour ; on
the other hand, the two yellow salts, caesium chloroplumbate, Cs2PbCl6, and
caesium chlorotellurate, Cs2TeCl6, give mixed crystals of an orange-red colour.
J. W. Retgers' colour test for mixed crystals.' — Saturated solutions of the two salts
which differ in colour are placed side by side on a microscopic slide and brought together
with a glass rod. The crystals which form on evaporation are examined under a microscope.
If the salts are isomorphous, the colour of the crystals varies gradually from one side to
the other, the crystals of the pure compounds being visible on the extreme edges. If the
two salts are not isomorphous, they do not mix, and near the centre, where the two solutions
have been brought together, distinct crystals of each compoimd can be seen owing to
their difference in colour. If the two salts have the same colour, J. W. Retgers uses a
third salt of a different colour from the other two. If both the same coloured salts form mixed
crystals with the salt of a different colour, the two salts under examination are isomorphous.
J. W. Retgers (1889) considers all important the property of forming mixed
crystals in all proportions such that " if the percentages of one constituent of the
mixture be plotted as abscissae, and the corresponding magnitudes of the physical
properties be plotted as ordinates, the different points lie in a continuous line."
Two substances are really isomorphous only when the physical properties of their
mixed crystals are continuous functions of their chemical composition ; or the
physical properties of isomorphous mixtures are continuous functions of the per-
centage composition — Retgers' law. Physical properties here include geometrical,
optical, thermal, elastic, and electrical properties. This agrees with F. W. Kiister's
statement that all the physical properties of isomorphous mixtures which have
been hitherto investigated are purely additive, and are continuous functions
of their percentage composition. For instance, the refractive index curve of
isomorphous mixtures of potassium and thallium alums lies on a straight line.
H. de Senarmont, A. des Cloizeaux, H. Dufet, G. WyroubofE, E. Mallard, R. Brauns,
A. Fock, T. Hiortdahl, etc., have shown that " the difference between the indices of
refraction of a mixture of two isomorphous salts and those of the components is
inversely as the number of the equivalents of the two salts in the mixtures," so
that if /i be the index of refraction of the mixtures, /xj and /X2 the indices of refraction
of the component salts of molecular weight Mi and ilf 2 respectively, then :
Ml/Xl+M2/X2
^-" M1+M2
660 INORGANIC AND THEORETICAL CHEMISTRY
G. Bodlander also obtained analogous results with the circular polarization of
mixed crystals of lead and strontium dithionates. The solubilities of isomorphous
mixtures have been investigated by C. F. Rammelsberg, C. von Hauer, and
F. Riidorfi. H. W. B. Roozeboom's work is indicated in the chapter on solutions, and
he adds that if the osmotic pressure of a saturated solution of mixed crystals with
an increasing content of one of the constituents, increases or diminishes, so is the
proportion of this constituent in the solution greater or less than in the mixed
crystal. Again, the melting points of isomorphous mixtures of albite and anorthite
furnish a series which is almost a straight line representing albite by Ab and anorthite
by An:
An AbAn. AbAn2 AbAn Ab2An AbgAn Ab
Melting point . 1532° 1600^ 1465° 1419° 1367° 1340° —
The melting point of albite has not been determined accurately within 150°. These
results are in agreement with the much-discussed generalization of F. W. Kuster
(1891) from observations on the melting points of mixtures of organic compounds.
He found a simple linear relation, sometimes called Kuster's rule : The solidiiying
point of an isomorphous mixture lies on a straight line connecting the melting
points of the individual components, and it can be calculated from the percentage
composition of the mixture. This simple linear relation is supposed to represent
perfect isemorphism. Imperfect isomorphism is assumed to be the cause of the
slight concavity or convexity usually observed with these curves. On practical
grounds, J. W. Retgers considers specific gravity, or the reciprocal of specific
gravity — the specific volume — to be the most suitable property for investigation.
This property was found by G. Tschermak, 0. Pettersson, R. Brauns, J. W. Retgers,
etc., to be a function of the specific gravity of the components such that if D and v
respectively denote the specific gravity and specific volume of the mixture, D^
and D2, and i^i and V2, the specific gravities and specific volumes of the components
when a denotes the volume percentage and j3 the weight percentage of the second
component of the mixture, then :
An example by J. W. Retgers is indicated in Fig. 91, where the specific volume of
mixed crystals of potassium and ammonium aluminium sulphates are plotted.
The continuity of the curve shows that the specific gravity or specific volume and
chemical composition of the mixed crystals are isomorphous. A similar continuous
curve is obtained with mixtures of magnesium and manganese pyrophosphates.
The curves sometimes show a break, as is the case with the dihydrogen phosphates
of potassium and ammonium. Isomorphous mixtures are formed only when not
more than 20 per cent, of the second constituent is present. Intermediate mixtures
do not form homogeneous crystals. This is illustrated in J. W. Retgers' curve,
Fig. 91, showing that the two substances are not miscible in all proportions, but
the two portions of the curve are parts of one straight line corresponding with the
isomorphous character of the two salts. Otherwise expressed, the solubility of
each substance in the other may be limited like the solubility of many salts in water,
or if the one in the other may dissolve the other in all possible proportions.
There has been some differences of opinion as to whether the formation of
mixed crystals with isomorphous substances must take place in all proportions,
or if gaps may occur ; in other words, if the term isomorphous be applied only
to those substances which form a continuous series of mixed crystals, and excluded
from those which form only partial series of mixed crystals. In opposition to
J. W. Retgers, W. Stortenbecker (1903) considers that if the substances are truly
isomorphous, no gaps will occur, while B. Gossner (1906) also considers that
isomorphous salts may exhibit gaps, and this the more, the greater the difference
in the molecular volume (molecular weight divided by the specific gravity) of the
CKYSTALS AND CRYSTALLIZATION
661
respective compounds. This, in turn, is possibly conditioned by the relative sizes
of the structural units of the two salts. According to B. Gossner (1907), also, if
the molecular volume of the isomorphous substances be nearly the same — e.g. nickel
or zinc fluosilicates and double alkali sulphates — the salts will form a continuous
series of mixed crystals ; while if the molecular volumes are very different —
e.g. copper or cobalt fluosilicates and double alkali sulphates — there wiU be a break
in the series of mixed crystals. If the continuous curve (Fig. 91) or broken curve
(Fig. 92) are not in the same straight line, the two salts, even if perfectly miscible
0-60
0-50
0-40
0-30
1:
oL56'37
^
y^
^
y^
*^
y
X
,^
y
^
37
51
P^rcent\{H\\^\^
'^~
«o 0-8
> 0-6
NO rh/A?S(.
—
1 —
—
«^ 0-4
—
"Crysta/s
1
4 0-2
:
1.
J
1
1
Q^O
1
20 40 60 80 100
Fig. 91. — Specific Volumes of Mixed
Crystals of Ammonium and Potassium
Aluminium Sulphates.
. KHjPO^. O 20 40 60 80 \00 percent.
[NH4)H2P0<4. 100 80 60 40 20 0 percent.
Fig. 92.— Specific Volumes of Mixed Crystals of
Ammonium and Potassium Dihydrogen Phos-
phates.
in all proportions, would not, according to J. W. Ketgers' definition, be called
isomorphous. For instance, ammonium and ferric chloride are not isomorphous,
although octahedral ammonium chloride forms coloured mixed crystals by taking
up a small amount of ferric chloride.
The formation of mixed crystals is conditioned or favoured by substances of
analogous structure, and this phenomenon is therefore regarded as strong evidence
of isomorphism, but the formation of mixed crystals is not a sufficient criterion of
isomorphism, since there is quite a large
number of cases of their formation by sub-
stances of different form ; nor is the forma-
tion of mixed crystals a necessary criterion of
isomorphism, since the isomorphous salts may
interact forming another chemical individual
instead of producing mixed crystals. In other
words, double compounds may be formed
which interfere with the application of J. W.
Retgers' rule — e.g. J. W. Retgers found with a
mixture of rhombic silver nitrate and potas-
sium nitrate, each can dissolve a little of the ^
other still forming rhombic crystals, but the
two salts crystallize together forming a double Fig. 93.— Specific Volumes of Mixtures of
salt, KAg(N03)2, which is monoclinic and has Silver and Potassium Nitrates,
a specific volume 0*31. If the double salt
were an isomorphous mixture, it would have a specific volume 0'38. The point
corresponding with this mixture thus lies off the straight dotted line, as shown in
Fig. 93. Similar results are obtained with mixtures of potassium and sodium
sulphates, the point corresponding with 3K2S04.Na2S04 is a long distance from
the line connecting the two components ; similarly, mixtures of magnesium and
calcium carbonates show a deviation for dolomite, MgCOs.CaCOs ; and potassium
chloride and cupric chloride— CUCI2.2H2O— form 2KCl.CuCl2.2H2O.
Overgrowths. — In 1816 J. L. Gay Lussac^* found that a crystal of alum continues
\J-iJ\J
i
0-47fl-2^
a45
,•••'
/"
0»40
x--
.•••^
035
.y
0-30
..••■'
•)C
'^\(
>6
..•••■
1
1
n-'
■
0-25
/■'..
1
!
0-20
02299^
1 1
Percknt^c
JNO3 1
20 40
60 80
62-73
100
662 INORGANIC AND THEORETICAL CHEMISTRY
to grow when placed in a solution of another alum. If a crystal of dark violet
chromium alum be placed in a saturated solution of ordinary potassium alum, a
transparent colourless overgrowth — lame de superposition — of potassium alum is
deposited as a crust over the dark-coloured chromium alum as a nucleus. Similarly,
a crystal of colourless zinc sulphate — ZnS04.7H20 — can be coated with an over-
growth of green nickel sulphate — NiS04.7H20 — and vice versa ; crystals of sodium
nitrate grow on Iceland spar ; and a pale amethyst triclinic crystal of manganese
sulphate — MnS04.5H20 — can be coated with blue copper sulphate — CUSO4.5H2O
— and vice versa. The parallel overgrowths formed in this way have been called
episomoiphs. C. von Hauer grew episomorphs of magnesium sulphate on magnesium
chromate or nickel sulphate ; of potassium magnesium sulphate on the corresponding
cobalt and nickel salts. In nature also episomorphs of potash and soda felspars
are common. H. Kopp (1873) stated that this power of forming overgrowths, as
well as the power of forming mixed crystals, enables isomorphism to be detected
even when no particulars about the crystalline form or about the chemical composi-
tion are available. There are some exceptions to the test for isomorphism — trigonal
potassium sulphate can be coated with a layer of hexagonal potassium sodium
sulphate — KNaS04 ; etc.
Isomorphous substances were found by F. M. Jager and H. Haga 25 to yield similar
radiograms although the relative intensities of the spots were often different.
C. Viola concluded that mixed crystals of magnesium and zinc sulphate are formed
in layers composed alternately of each salt, and therefore differ radically from
solid solutions. L. Vegard and H. Schjelderup have also examined the X-ray
spectra of mixed crystals of potassium chloride and bromide, and of potassium and
ammonium bromide. They found that the crystals behaved as single entities and
there was no indication that the crystals were composed of thin homogeneous
laminae. G. Tammann examined the effect of reagents which attack one of the
components of binary mixed crystals of gold and copper, and gold and silver. The
action is not proportional to the amount of the soluble constituent ; indeed, there
are limits of composition between which the resistance to attack is very great.
Strictly speaking, all substances with a similar crystalline form are isomorphous.
The similarity of external geometrical form is prima faxiie evidence of the similarity
of internal structure, so that those crystalline substances are isomorphous whose
structure is analogous (E. Mitscherlich) . If the molecular volumes are sufficiently
close, isordorphous substances usually form (i) homogeneous mixed crystals (J. W.
Retgers) ; (ii) they form parallel overgrowths on each other (H. Kopp) ; (iii) they
are mutually active in inducing crystallization when a supersaturated solution of
the one is inoculated with a small fragment of the other ; and (iv) are generally of
analogous chemical constitution. None of these tests is an infallible criterion, and
here, as is so often the case, a conclusion can be drawn only after carefully balancing
the available circumstantial evidence. C. Hlawatsch recognizes degrees of iso-
morphism, and he has discussed the nature of isomorphism, and classified substances
according to their degree of isomorphism on the lines of the following scheme :
( 1 ) The substances exhibit no chemical analogy, but show similarities in certain zones
which frequently grow parallel. (2) The substances show analogies in their angles, but
do not exhibit the same cleavages or habit. This may be termed isogonism. (3) The
substances form mixed crystals, but have not analogous structure. (4) The last case is
not to be confused with that presented by isopolymorphous substances when the two
modifications possess very different stability. (5) The substances show like structure
expressed, not merely by similarity of form, but by like cleavage, twinning, and habit.
(6) The substances have similar crystal structure, and may form mixed crystals, but do
not belong to the same crystal sub-class. (7) The substances possess similar structure
with identical symmetry, and form mixed crystals, but are not chemically analogous.
(8) The substances show chemical analogy in addition to the other characters. (9) Lastly,
they possess chemical analogy, form mixed crystals, have similar structure, and angular
relations which are functions of the atomic weights of the interchangeable elements.
According to G. Tammann (1907), chemically analogous elements are usually
CRYSTALS AND CRYSTALLIZATION
663
isomorphous. Elements in the same groups in periodic system usually form mixed
crystals and not compounds, while many elements not in the same group but
chemically similar also form mixed crystals. In binary alloys, elements of high
melting point usually form mixed crystals with those of low melting point.
A. Arzruni, in his Die Beziehungen zwischen Krystallform und chemischer Zusammen-
setzung (Braunschweig, 1898), arranges sixty-eight of the elements in ten isomorphous
series (isomorphe Reihen) :
Series /.~H, K, Rb, Cs, NH^, Tl, Na, Li, Ag. Series //.—Be, Zn, Cd, Mg, Mn, Fe,
Os, Ru, Ni, Pd, Co, Pt, Cu, Ca, Sr, Ba, Pb. Series III.^L&, Ce, Di, Y, Er. Series /F.— Al,
Fe, Cr, Co, Mn, Ir, Rh, Ga, In (Ti). Series F.— Cu, Hg, Pb, Ag, Au. Series F/.— Si, Ti,
Ge, Zr, Sn, Pb, Th, Mo, Mn, U, Ru, Rh, Ir, Os, Pd, Pt, Te (?). Series VII.— N, P, V,
As, Sb, Bi. Series F///.— Nb, Ta. Series IX.S, Se, Cr, Mn, Mo, N, Te (?), As, Sb.
Series X.—F\, CI, Br, I, Mn, Cy.
A given element may appear in different isomorphous series. In illustration,
manganese, in its different states of oxidation, belongs to different classes. There
are also certain regularities exhibited by the members of a sub-group in Mendeleeff's
periodic arrangement.
There is undoubtedly a profound connection between the similarity of crystalline
form and the similarity of chemical structure. The thousand and one known cases
typified by the isomorphism of potassium sulphate and selenate were supposed to
be a result of the chemical similarity of the replaceable elements, and isomorphous
replaceability was found to be a periodic function of the elements since the elements
belonging to the same sub-group in the periodic system usually gave isomorphous
compounds. T. V. Barker 26 has compiled a number of examples of what he calls
" unusual types of isomorphism," in which there is no similarity of valency structure.
The following are taken from T. V. Barker's list, where the number of known analogues
of any particular compound is indicated in brackets :
Monoclinic system.
CuTiFj.4H20 (2) .
CuCbOF5.4H20 (1)
CUWO3F4.4H2O (1)
KsHSnFg (1)
K3HCbOF7 (0) .
MnCl2.4H20 (1) .
BeNagF^ (0) .
(NHJ^SeO, (2) .
CsgHgl, (0) .
Rhombic system,
K3SnCl4.2H20 (3).
KaFeClg.HaO (4) .
0-7471 :
0-7627 :
0-7648 :
0-6277 :
0-6279 :
1-1525:
1-9913 :
1-8900:
1-3155:
0-5564 ;
0-5629 ;
0-5629 ;
0-4928 ;
0-4900 ;
0-6445 ;
0-6929 ;
1-1987;
0-9260 :
104° 9'
103° 20'
103° 14'
93° 0'
93° 14'
99° 25'
99° 20'
115° 29'
110° 4'
Rhombic system.
KCIO4 (8)
BaS04 (5)
KBF4(1)
K2SO4 (15)
K2BeF4 (4)
0-7817:
0-8152:
0-7898 :
0-5727 :
0-5708 :
a:b:c
0-6852 : 1 : 0-7586
0-6911 : 1 : 0-7178
[N(CH3)j2HgCl4(l) 0-5766
Tetragonal system.
Xenotime, YPO4 (0) 1
Zircon, ZrSi04 (1) 1
Cassiterite,Sn02orSnSnOa(4) 1
KI04(3) . . 1
CaWO, (7) . . 1
KOsOgN (1) . . 1
KRUO4 (0) . . 1
: 1-2792
: 1-3136
: 1-2830
: 0-7418
: 0-7395
: 0-7893
a:c
: 0-6177
: 0-6400
: 0-6726
: 1-5534
: 1-5268
: 1-6319
: 1-6340
Rhombic crystals of aragonite, CaCOs, and nitre, KNO3, have axial ratios
respectively a:h: 0=0*622 : 1 : 0721 and 0-591 : 1 : O'TOl ; both salts have a similar
crystalline form but a very different chemical constitution. Similarly with rhombo-
hedral calcite, CaCOs, and sodium nitrate, NaNOs, with the respective rhombohedral
angles of 74° 55' and 73° 27', and axial ratios a : c=l : 0-854 and 1 : 0-8297. The
constitutional formulae of calcium carbonate (calcite) and sodium nitrate are
represented respectively by
0=C<^>Ca
0
>N-ONa
so that if these formulae really indicat-e internal structures the observed isomorphism
must be due to some obscure accidental cause. The facts indicated a similarity in
crystal structure ; the valency theory indicated complete dissimilarity. As a
result, some denied the apparent isomorphism of these compounds, and narrowed
the definition of isomorphism so that a special name homomorphism or isogonism
664 INORGANIC AND THEORETICAL CHEMISTRY
was devised for the phenomena presented by substances which differ in chemical
constitution, but have a similar crystalline form. Attempts to evade the difficulty
presented by calcite and sodium nitrate where the total sum of the valencies is in
case twelve, by assuming that the sexavalent groups CaC and NaN replace each
other isomorphously. The explanation is futile, because it fails to account for the
replaceability of the alkali metals by the ammonium radicle, where the valency
summations are respectively one and nine. Hence, adds T. V. Barker, " the equality
of valency summation has nothing to do with isomorphism."
The X-ray spectrum has enabled W. L. Bragg (1914) to show a close structural
relationship between the crystals of calcium carbonate and potassium nitrate, which
are not considered to be isomorphous. Tetragonal zircon, ZrSi04, and thorite,
ThSi04, are isomorphous with rutile, Ti02, and cassiterite, Sn02. The difference
in structure is not so apparent if the constitutions be represented : Zr02Si02,
Th02Si02, (TiO)2. In virtue of isogonism, the supposed relations which have been
traced between chemical composition and crystalline form are often quite accidental.
In the triclinic system, isogonism is usually an indication of isomorphism, but its sig-
nificance becomes less and less as the crystals increase in symmetry. T. V. Barker
claims that it is no longer advisable to limit the term isomorphism to cases of
chemical and crystal symmetry by interpreting the chemical similarity in terms of
the older valency hypothesis. It is true that the dissimilarity in the constitution
of these substances with similar crj^stals is so great that it might be hazarded that
" isomorphism may be totally independent of chemical structure ; "but this view
is untenable, for it is the valency structure which is at fault. If the constitution of
these compounds be interpreted on A. Werner's co-ordination structures, there are
indications that the compounds with apparently dissimilar structure will prove to
have analogous structures. For example, the molecular formulae of the first three
compounds in T. V. Barker's list are not very similar, but when expressed according
to A. Werner's system, the case is somewhat different :
rTiF6]cu+4H20 ; Tob^ lcu+4H20 ; and [w^2lou+4H20
Again, J. C. G. de Marignac's compounds KsHSnFg and K3HCbOF7 appear closely
related when represented by A. Werner's formulae :
(FK)3
OCbFH
F
r (FK)3-
and SnFH
L F2 .
There is no analogy in the chemical formulae of the compounds 2KCl.SnCl2.2H2O
and 2KCl.FeCl3.H2O, either in this molecular form or when expressed by the complex
salt formulae K2SnCl4.2H20 and K2FeCl5.H20. The co-ordination formulae, how-
ever, show a close analogy :
}^Ko),W''^^[Kioh
Again, [Mn(H20)4]K2 and [BeF4]Na2 are similar if it be assumed that the molecules
of water which are co-ordinated can be replaced isomorphously by halogen atoms.
The isomorphism of Znl2.4NH3 with potassium beryllium fluoride leads tothe assump-
tion that the internal structures are similar in this sense : [Zn(NH3)4]l2 and [BeF4]K2,
where the positive ion of one compound is analogous with the negative ion of the
other. A. Werner also represented KOSO3N and KBF4 respectively by the co-
ordination formulae [OsOsNJK and [BFJK. Hence, adds T. V. Barker, co-ordination
evidently supplies a medium in which analogy of chemical composition in isomor-
phous compounds formerly classed as homomorphic or isogonic, comes strongly
into the foreground.
Isodimorphism — In 1829, J. F. W. Johnston 27 drew attention to the fact that
CRYSTALS AND CRYSTALLIZATION
665
CoAsj
NiAso
a plumbiferous calcite, which he named plumbocalcite, and which not only contained
calcium carbonate but also the lead carbonate. Hence, the latter must crystallize
in two forms, so that in addition to rhombic cerussite, there must also be a
rhombohedral form of lead carbonate which can crystallize with the corresponding
form of calcium carbonate. Hence, both lead and calcium carbonate were called
isobimorphs. Such a form of lead carbonate has not yet been discovered, but
J. F. W. Johnston's reasoning was sound. Some years later M. L. Frankenheim
made a similar discovery with respect to calcium carbonate and potassium nitrate.
Many examples are now known in which there are two independent series of
isomorphous salts, and the phenomenon is called isodimorphism. The pyrite
and marcasite families of minerals form two independent series of isomorphous
crystals. The following were compiled by J. P. Iddings (1906) :
Cubic. Rhombic,
pyrite marcasite
sraaltite saffrolite
chloanthite rammelsbergite
(Co, Fe) (S, As)2 cobaltite glaucodote
NiS2.Ni(Sb, As) a . . . . corynite wolfachite
Each of the sulphates RSO4.7H2O (where R may be Mg, Zn, Ni, Co, Fe, Mn)
is dimorphous, forming rhombic and also monoclinic crystals. The rhombic crystals
of al] the salts form one isomorphous series, and the monoclinic crystals of all
the salts form another isomorphous
series. The isodimorphism is here
limited to certain proportions of the
constituents. For example, mixtures
of iron and magnesium sulphates
give homogeneous monoclinic mixed
crystals if less than 54 per cent, of
magnesium sulphate, MgS04.7H20,
be present, and rhombic mixed cry-
stals if more than 81 per cent, be
present. The specific volume curve
of the mixed crystals does not there-
fore lie in one straight line. This is pIIq^Jh^oioo so
illustrated by J. W. Retgers' diagram, * ^
Fig. 94. Fig. 94. — Specific Volumes of Mixed Crystals of
Magnesium and Ferrous Sulphates — Limited
SimQarly, a mixed solution of silver Isomorphism,
and sodium chlorates gives mixed cubic
crystals if the sodium chlorate be in excess, and mixed tetragonal crystals if the silver chlorate
be in excess. J. W. Retgers found that copper sulphate soliitions crystallizing in the
presence of a small proportion of zinc sulphate (0 to 7*98 per cent.) furnishes triclimc
mixed crystals of CUSO4.5H2O and ZnSO^.SHaO ; and if a larger proportion of zinc sul-
phate be present, 65*59 to 83 '35 per cent., monoclinic crystals of a mixture of CUSO4.7H2O
and ZnSO^.THgO ; while if a relatively large proportion of zinc sulphate, 97*68 to 100 per
cent., be present, rhombic crystals of a mixture of CUSO4.7H2O and ZnS04.7H20 are
formed. Similar results are obtained with mixtures of copper and magnesium sul-
phates. H. W. Foote (1902) has shown that tetragonal beryllium sulphate, BeS04.4H20.
forms mixed tetragonal crystals with beryllium selenate, BeSe04.4H20, provided the mole-
cular proportions of the respective salts in solution does not exceed 7*33 : 1 ; while if the
solution contains a less proportion of the sulphate, a series of rhombic crystals can be
prepared when the molecular proportions of the beryUium sulphate to selenate lie between
4 : 1 and pure beryllium selenate.
The composition of mixed crystals of one salt with a maximum proportion of
the other is called the mixing limit. Thus, tetragonal mixed crystals of berj^llium
sulphate and selenate have reached the mixing limit when their composition has
BeS04.4H20 : BeSe04.4H20=7-33 : 1. J. H. van't Hoff (1898) suggested an
interesting analogy between mixed crystals and ordinary solutions. Pairs of salts
which crystallize together are likened to perfectly miscible liquids like alcohol and
I
0-6
r—
] —
—
—
f)'h05
n
1^
,••*
.-••"■
..-•
^^'^
...-•
■■■\
...-•
,.'■'''
y
y
,y*
iC
1
-••'
k
W
)ll'
1
1
in^
I
!
0-5
1
...
40
60
60
40
80 100 percent
20 O percent
666 INORGANIC AND THEORETICAL CHEMISTRY
water, while salts of the second class, whose isomorphism is limited, are likened to
partially miscible liquids like aniline and water. The analogy has been pushed still
further. At temperatures exceeding 165°, aniline and water mix in all proportions,
while below that temperature the two liquids are but partially miscible ; hence, it
is inferred that unless other changes intervene, salts which are but partially miscible
as ordinary temperatures may be perfectly miscible at more elevated temperatures.
According to the phase rule, tetragonal BeS04.4:H20 will contain a maximum
proportion of BeSe04.4:H20 when the solution from which it is deposited is saturated
with regard to the mixed crystals of BeSe04.4H20 and BeS04.4H20. In such
systems there are four phases — vapour, solution, and two solids — and three
components — the two salts and water ; the system is accordingly univariant.
Accordingly, if one of the possible variables — temperature, pressure, or concentra-
tion of phase — be changed, the others must be fixed and unalterable. For instance,
if the temperature be fixed, the concentration of each phase and the vapour pressure
must be fixed ; and if the temperature be changed, another variable must be
changed. Hence, the composition of mixed crystals at the mixing limit will change
with change of temperature, and this is in agreement with observations.
Two or more compounds which, judged by all analogies, might be expected to
be isomorphous may exhibit pronounced differences in crystalline form ; but, by
suitably altering the conditions, they may furnish a second form, so that the isomor-
phism of the series is established. Monoclinic felspar — orthoclase — usually contains
some sodium ; while triclinic soda felspar — albite — contains some potassium.
Hence P. Groth (1874) 28 inferred that this is a case of isodimorphism, and that two
pure varieties — monoclinic and triclinic soda and potash felspars — should exist. The
prediction was verified two years later by A. des Cloizeaux's discovery of microcline,
the triclinic form of potash felspar; and later, by P. Barbier discovery of
barbierite, the monoclinic form of albite. Sodium phosphate forms two distinct
crystals — rhombic and monoclinic. The arsenate appears in only one of these
forms. Hence it is inferred that a monoclinic sodium arsenate isomorphous with
rhombic sodium phosphate remains to be discovered.
References.
^ E. Wohlwill, Ueher isomorphe Mischungen der selensauren Sake, Gottmgen, 1860 ; Liebig's
Ann., 114. 181, 1860 ; A. G. Monnet, TraiU d^ la vitriolisation et de Valunation, Paris, 1769 ;
J. F. Henkel, Rhine mineralogische und chemiscJie Schrifien, Dresden, 1744-69.
* J. B. L. Rome de I'lsle, Essai de Criatallographie, Paris, 67, 1772 ; N. Leblanc, Journ.
Phys., 55. 300, 1802; J. N. Fuchs, Schweigger's Journ., 15. 377, 1815; F. S. Beudant,
Ann. Chim. Phys., (2), 4. 72, 1817; L. N. Vauquelin, ih., (1), 22. 258, 1797; J. L. Gay
Lussac, ib., (2), 2. 176, 1816 ; P. Berthier, ib., (1), 58. 149, 1806 ; F. S. Beudant, Ann. Mines,
(1), 2. 1, 1817; (1), 3. 239, 289, 1818; A Bernhardi, GeUen's Jonrn., 8. 360, 1809; W. H.
WoUaston, Ann. Phil., 11. 283, J818 ; C. L. BerthoUet, Essai de statique chimique, Paris, 1. 442,
1803 ; M. H. Klaproth, Beitrdge zur chemischen Kenntniss der Min£ralk6rper, Berlin, 2. 16, 239,
1797 ; 5. 131, 1810; 0. F. Bucholz, Gilbert's Ann., 9. 434, 1801.
» E. Mitscherlich, Handl. Akad. Stockholm, 4, 1821 ; Ann. Chim. Phys., (2), 14. 172, 1820;
(2), 19. 350, 1821 ; Oesammelte Schriften, Berlin, 1906.
* Ann. Chim. Phys., (2), 14. 305, 1820,
^ R. J. Haiiy, Trait e de mineralogie, Paris, 1801.
« W. H. WoUaston, Phil. Trans., 102. 159, 1812; 92. 385, 1802; J. B. Biot, Ann. Chim.
Phys., (2), 14. 192, 1820 ; E. L. Malus, ib., (2), 19. 377, 1821.
' A. E. H. Tutton, Journ. Chem. Soc, 63. 337, 1893 ; 65. 628, 1894 ; 69. 344, 496, 1896 ;
71.846, 1897; 87. 1183, 1905; Proc. Boy. Soc., 6Q. 248, 1900; 68. 322, 1901; 83. A, 211,
1910 ; Crystalline Structure and Chemical Constitution, London, 1910 ; Crystals, London,
1911 ; Phil. Trans., 216. A, 1, 1915.
8 F. M. Jager, Bee. Trav. Chim. Pays-Bas, 33. 343, 1914; A. E. H. Tutton, Phil. Trans.,
216. A, 1, 1915.
* G. Linck, Grundriss der Kristallographie, Jena, 1913.
1° W. Muthmann, Zeit. Kryst., 22. 497, 1894; F. Becke, Sitzber. Akad. Wien, SO. 204,
1893; A. E. H. Tutton, Journ. Chem. Soc, 65. 628, 1894; P. Groth, Ber., 9. 449, 1870;
A. Laurent, Compt. Bend., 11. 876, 1840 ; 14. 818, 1842 : 15. 350, 1842 ; F. de la Provostaje, ib.,
11. 635, 1840; L. J. Wallmark, Journ. prakt. Chem., (1), 31. 169, 1844; W. G. Hankel, Pogg.
CRYSTALS AND CRYSTALLIZATION 667
Ann., 55. 479, 1842; L. Bodart, CompL Bend., 27. 321, 1848; J. Nickl^s, ib., 27. 244, 1848;
T. von Alt, Sitzber. Akad. Wien, 12. 664, 1854 ; L. Pasteur, C(ympt. Rend., 26. 635, 1848.
" F. Becke, Sitzber. Akad. Wien, 30. 204, 1893 ; W. J. Pope, Annual Reports of the Progress
of Chemistry, 5. 258, 1909.
12 A. E. H. Tutton, Journ. Ohem. Soc, 65. 688, 1894; 87. 1183, 1906; J. A. le Bel and
A. Ries, Zeit. Kryst., 36. 321, 1902 ; 39. 49, 1904 ; G. Mez, ib., 35. 242, 1902 ; W. Muthmann,
ib., 22. 497, 1894 ; T. V. Barker, ib., 43. 529, 1907.
13 G. J. Stoney, B. A. Rep., 988, 1885.
" F. Slavik, Zeit. Kryst., 36. 268, 1902.
i"* C. A. Kenngott, Jahrb. Geol. ReichsansL, 3. 104, 1862 ; A. Schrauf, Lehrbiich der Physikal-
ischen Chemie, Wien, 2. 69, 1868 ; Pogg. Ann., 134. 422, 1868 ; F. Pfaff, Sitzber. Akad. Munchen,
255, 1884 ; H. R. Koch, Verh. Nat. Ges. Freiburg, (2), 8. 1, 1881.
1 « G. Tschermak, Sitzher. Akad. Wien, 45. 604, 1862 ; 0. Sadenbeek, Ueber die Theilbarkeit den
KryMalle, Berlin, 1876.
1' H. Baumbauer, Pogg. Ann., 138. 563, 1869 ; 139. 349, 1870 ; 140. 271, 1870 ; 142.
325, 1871 ; 145. 459, 1872; 150. 619, 1873; Neues Jahrb. Min., 411, 1873; Zeit. Kryst., 1.
54, 1877; Sitzber. Akad. Berlin, 863, 1887; 447, 1890; G. Tschermak, Tschermak's Mitt., 4.
99. 1882 ; F. Becke, ib., 11. 224, 1890.
18 E. Jannettaz, Ann. Chim. Phys., (4), 29. 6, 1873 ; Compt. Rend., 75. 1501, 1872 ; 114.
1352, 1892 ; Bull. Soc. Geol., (3), 5. 410, 1877 ; Bull. Soc. Min., 2. 104, 1879 ; V. von Lang,
Sitzber. Akad. Wien, 54. 163, 1866; T. Liebisch, Physikalische Krystallographie, Leipzig, 148,
1891.
18 H. Fizeau, Ann. Bur. Longitudes, 562, 1890 ; F. Pfaff, Pogg. Ann., 107. 148, 1869.
2» J. Grailich and V. von Lang, Sitzber. Akad. Wien, 32. 43, 1858 ; 33. 439, 1858.
21 H. de Senarmont, Ann. Chim. Phys., (3), 33. 391, 1851 ; J. Grailich, Krystallographisch-
optische Untersuchungen, Wien, 1858; J. Grailich and V. von Lang, Sitzber. Akad. Wien, 27.
3, 1867 ; 31. 85, ]858 ; 33. 369, 1858 ; 34. 135, 1859 ; H. Topsoe and C. Christiansen, Vidensk-
SeM. Nat. Math. Kjobenhavn, 9. 625, 1873 ; Ann. Chim. Phys., (5), 1. 5, 1874 ; A. Arzruni,
Zeit. Kryst., 1. 165, 1877 ; A. des Cloizeaux, Mem. Acad., 18. 512, 1867.
22 F. S. Beudant, Ann. Chim. Phys., (2), 4. 72, 1817 ; (2), 7. 399, 1817 ; (2), 8. 5, 1818.
23 H. Ambronn and M. le Blanc, Ber. Sachs. Ges. Wiss., 173, 1894 ; J. H. van't Hoff, Zeit.
phys. Chem., 5. 322, 1890 ; L. de Boisbaudran, Compt. Rend., 113. 832, 1891 ; 142. 196, 1906
G. Bodlander, Neiues Jahrb. Min. B. B., 12. 62, 1898 ; Ueber das optische Drehungsvermogen
isomer pher Mischungen aus den Dithionaten des Bleis urid des Strontiums, Breslau, 1882 ; J. W
Retgers, Zeit. phys. Chem., 3- 497, 1889 ; 4. 693, 1890 ; 5. 436, 1890 ; 6. 193, 1890 ; 8. 6, 1891
9. 267, 385, 1892 ; 10. 629, 1892 ; 11. 328, 1893 ; 12. 583, 1893 ; 14. 1, 1894 ; 15 529, 1894
16. 577, 1895; 20. 481, 1896; B. Gossner, Ber., 40. 2373, 1907 ; Zeit. Kryst., 38. 110, 1903
39. 381, 1904; 43. 130, 1907; A. Fock, ib., 6. 163, 1882; W. Muthmann, ib., 17. 336, 1890
W. Stortenbecker, Zeit. phys. Chem., 43. 629, 1903; F. W. Kuster, ib., 5. 601, 1890; 8. 577
1891 ; H. de Senarmont, Ann. Chim.. Phys., (3), 33. 413, 1851 ; A. des Cloizeaux, Ann. Mines.
(6), 11. 321, 1858 ; (5), 14. 366, 1858 ; E. MaUard, ib., (7), 10. 176, 1876 ; G. Wyrouboff, 5mZZ. Soc.
Min., 7. 8, 1884 ; 2. 91, 170, 1879 ; E. MaUard, ib., 3. 3, 1880 ; H. Dufet, ib., 1. 58, 1878 ; 2,
140, 1879 ; 3. 180, 182, 1880 ; Compt. Rend., 86. 880, 1878 ; 91. 286, 1880 ; 99. 990, 1884 :
C. Soret, ib., 99. 867, 1884 ; Arch. Sciences Geneve, (3), 12. 553, 1884 ; R. Brauns, Neues. Jahrb
Min., ii, 72, 1886 ; ii, 12, 1891 ; A. Fock, Zeit. Kryst., 4. 583, 1880 ; 5. 598, 1881 ; T. Hiortdahl
Vidensk. Selsk. Forh. Christiana, 7, 1882; C. F. Rammelsberg, Pogg. Ann., 91. 321, 1854;
F. Riidorff, ib., 148. 454, 555, 1873 ; Sitzber. Akad. Berlin, 356, 1885 ; C. von Hauer, Journ. prakt
Chem., (1), 98. 137, 1866 ; (1 ), 103. 114, 1868 ; Sitzber. Akad. Wien, 53. 221, 1866 ; G. Tschermak
ib., 50. 566, 1864 ; H. W. B. Roozeboom, Zeit. phys. Chem., 8. 604, 1891 ; 0. Pettersson, Ber.
9. 1676, 1876 ; C. Hlawatsch, Zeit. Kryst., 51. 417, 1912 ; G. Bruni, Feste Losungen und Iso-
morphismus, Leipzig, 1908 ; H. W. B. Roozeboom, Zeit. phys. Chem., 30. 386, 1899.
2* J. L. Gav Lussac, Ann. Chim. Phys., (2), 2. 176, 1816 ; C. von Hauer, Sitzber. Akad. Wien,
53. 226, 1866 ;'^H. Kopp, Ber., 12. 914, 1879 ; 15. 1653, 1882 ; G. Tschermak, Tschermak's Mitt.,
4. 99, 1881 ; E. Mallard, Bull. Soc. Min., 9. 117, 1886 ; J. W. Retgers, Zeit. phys. Chem., 3. 603,
1890 ; 5. 460, 1890 ; Neues Jahrb. Min., i, 147, 1891.
25 F. M. Jager and H. Haga, Proc. Acad. Amsterdam, 18. 1357, 1916; L. Vegard and
H. Schjelderup, Phys. Zeit., 18. 93, 1917 ; C. Viola, Atti Accad. Lincei, (6), 25. ii, 286, 1916 ;
G. Tammann, Nachr. Gott., 199, 1916; A. Arzruni, Die Beziehungen zwischen KrystaUform und
chemischer Zusammensetzung, Braunschweig, 1898.
26 T. V. Barker, Journ. Chem. Soc, 101. 2484. 1912.
27 J. P. W. Johnston, Edin. Phil. Journ., 6. 79, 1829 ; M. L. Frankenheim, Pogg. Ann., 40.
447, 1837 ; 92. 354, 1854 ; J. P. Iddings, Rock Minerals, New York, 19, 1906.
28 P. Groth, Tabellarische Uebersicht der Mineralien, Braunschweig, 106, 1874; A. des
Cloizeaux, Ann. Chim. Phys., (5), 9. 433, 1876; P. Barbier, Compt. Rend., 146. 1330. 1908;
W. T. Schaller, Amer. Journ. Science, (4), 30, 358, 1910.
668 INORGANIC AND THEORETICAL CHEMISTRY
§ 11. The Rectification of Atomic Weights by Isomorphism
When one body is isomorphous with another whose molecule contains a known number
of atoms, then the number of atoms per molecule of the other body is also known because
isomorphism is a mechanical consequence of the identity of atomic structure.- — J. J.
Berzelius(1833).
While perhaps not accepting J. J. Berzelius' dictum tvithout modifying the mean-
ing of " number of atoms " to allow for cases of isomorphism where a radicle like
NH4 containing five atoms can take the place of one atom of the alkali metals in some
isomorphous salts, yet J. J. Berzelius' statement of the law of isomorphism can be
used as a control in deducing the chemical composition of a salt ; and also in atomic
weight determinations for deciding between two numbers which are multiples
of a common factor. The method is restricted to crystalline compounds ; and it
is only applicable in conjunction with other methods of atomic weight determinations,
since at least one member of the isomorphous series must be known.
E. Mitscherlich deduced the number 79 for the atomic weight of selenium by this
method, and he also gave selenious and selenic acids formulae corresponding with
sulphurous and sulphuric acids respectively, on account of the isomorphism of the
sulphates and the selenates. The analyses of potassium sulphate and of potassium
selenate gave :
Potassium.
Oxygen.
Sulphur.
Selenium.
Total.
Potassium sulphate .
. 44-83
36-78
18-39
. — .
-100-00
Potassium selenate .
. 44-83
36-78
. —
45-40
-127-01
Assuming that the molecule of potassium sulphate contains one atom of sulphur ;
that the molecule of potassium selenate contains the same number of atoms ; and
that the atomic weight of sulphur is 32, we have :
Atomic weight S : Atomic weight Se=18'39 : 45*40,
Hence, 32 : atomic weight Se=18'39 : 35*34 ; consequently, the atomic weight of
selenium is 79*00. About 1836 the atomic weight of copper was supposed to be
63*4, and of silver, 216*6. The analysis of the native sulphides of these elements
were accordingly represented by the formulse CU2S and AgS. But J. B. A. Dumas
(1837) pointed out that the two minerals are isomorphous, and various mixed sul-
phides of the two elements are known by the general term, Fahlerz. Hence the
constitution of the two sulphides is probably the same ; assuming the formula of
the one to be CugS, that of the other will probably be Ag2S, and the atomic
weight of silver 108*3, not 216*6. This result agrees with evidence deduced from
other independent sources. More exact determinations of the atomic weight of
silver make this element 107*9 ; but this does not affect the principle of the argument.
Example. — Analyses of alumina show that Al : 0=^18-1 : 16 ; the equivalent of
aluminium in 9-03 ; hence the formula of alumina might be :
AlO AI2O3 AIO2 AIO3 . . .
Ratio . . . 0 : Al . . . 16 : 18*1 48 : 27-1 32 : 36-2 48 : 542, . . .
that is, the atomic weight of aluminium might be 18*1, 27-1, 36*2, 54-2 . . . There is
nothing in the composition of the oxide to show which of these numbers should be selected.
It is known, however, that ferric oxide— FcaOg- — forms a series of iron alums isomorphous
with the aluminium alums ; hence, it is inferred that the constitution of aluminium oxide
is AI2O3 — -like that of ferric oxide — -and that the atomic weight of aluminium is 27-1.
§ 12. The Formulae o£ Minerals, and of Isomorphous Mixed Salts
Whether and when formula* can be employed for minerals must be learned from faith-
ful analyses.- — T. Bergmann (1779).
We can scarcely doubt that there is a fixed proportion of elements in each mineral
CRYSTALS AND CRYSTALLIZATION 669
substance, which constitutes its true nature, so that what exceeds a given limit should
be to that degree regarded as accidental and foreign. — R. J. Hauy (1801).
A great many minerals can be synthesized in the laboratory. The chemist can
then use fairly pure materials and obtain fairly pure products. On the contrary,
nature, in her great laboratory, has rarely dealt with pure materials, and accordingly,
her products — the minerals — are usually contaminated with much impurity. The
determination of the formulae of minerals is exceptionally difficult mainly because
(1) the molecular weights can seldom be determined, and the formulae are therefore
nearly always empirical ; (2) the material available for analysis is more or less impure ;
(3) members of certain isomorphous groups of elements — iron, aluminium, chromium,
etc. ; calcium, magnesium, iron, manganese, etc. ; sodium, potassium, lithium,,
etc.— can replace one another in every conceivable proportion. Usually the crystal-
line form as well as the analysis, is necessary for establishing the individuality of
any mineral. Thus, H. A. Miers (1902) has said : It is necessary to employ at least
two properties, namely, the chemical composition and the crystalline form ; these
two when completely known are necessary and sufficient for the definition and deter-
mination of any mineral. Colour, structure, state of aggregation, and minor details
of chemical composition are used to distinguish subordinate varieties of the main
types.
The ultimate composition of any native calcium carbonates is exceedingly com-
plex. The same remark is more or less true for most native minerals ; at least
chemical formulae which rigorously followed the analyses would be very complex.
The formulae for minerals are commonly represented as if pure minerals occurred
in nature. Ideally pure minerals are seldom found native, and accordingly the stan-
dard formulae represent idealized or imaginary minerals to which real minerals
approximate more or less closely. The secondary constituents present in but small
quantities are usually ignored and the main constituents are alone included in the
formulae. For example, the analyses of a sample of limestone from Buxton furnished :
CaO MgO KaO NajO COa FcaO. and ALOa SiO,
54-76 0-31 0-25 024 43'78 0-26 0-88
Neglecting constituents less than one per cent., the remaining CaO and CO2 are in
the proportions needed for CaO.C02, or CaCOs. The sample here selected was
fairly pure and clean ; it contained 98*5 per cent, of calcium carbonate. In the
case of less pure minerals there is sometimes a doubt as to what is the best repre-
sentative formula, and the identification of the mineral is then based on its external
or physical properties rather than on its ultimate composition, for it has to be assumed
that the deviations are due to admixed impurities in order to avoid conflict with
the constant composition law. The difficulty is so real that many mineralogists
define a mineral species as a natural inorganic substance whose chemical and
physical properties are constant only within certain limits.
Examples.— (1) Clean crystals of cerussite from Tsumeb (S.W. Africa) furnished
H. Dubigk (1913) on analysis : PbO, 83-27 ; CO2, 16-64 ; insoluble matter, 0*24 per cent.
Show that the best representative formula is PbCOj.
(2) J. T. Bell (1892) analysed a sample of cuproplumhite, and found: copper, 61-32
percent.; lead, 18-97; sulphur, 17-77; and silica (SiO 2), 1*58 per cent. Show that this
analysis agrees with the formula SCuaS.PbS, assuming that the copper occurs in the
mineral in the cuprous condition.
(3) A sample of hodgkinsonite from New Jersey (U.S.A.) furnished C. Palache and
W. T. Schaller (1913) with: SiOj, 19-86 ; MnO, 20-68 ; ZnO, 52-93 ; CaO, 0-93 ; MgO, 0-04 ;
H2O, 5-77 per cent. Show that MnO. 2ZnO.SiO2.HaO best represents the analytical data.
(4) A marl from South Lincolnshire furnished on analysis : CaO, 42-6 ; MgO, 7-9 ;
CO2, 41-6 ; K2O, 0-3 ; AljOg, 08 ; FejOg, 0-5 ; SiOj, 5-1 ; H2O, 0-8 per cent. Here the
water, potash, alumina, ferric oxide appear to be of secondary importance. Divide the
weights of each of the remaining constituents by the respective molecular weights, and
reduce to the nearest whole number. There remains: 9Ca0.2Mg0.11C02,SiO, or
9CaCO3.2MgCO3.SiO2. The substance is thus considered to be a mixture of calcium and
magnesium carbonates with silica, and other impurities.
670 INORGANIC AND THEORETICAL CHEMISTRY
Minerals with isomorphous groups of elements. — In minerals, the members of
an isomorphous series of bases appear to be replaceable one with another to an unde-
termined extent, while the type of mineral or the crystalline form remains the same.
For example, in lime garnet, 3CaO.Al203.3Si02, a certain proportion of the lime,
CaO, may be replaced by ferrous oxide, FeO, and by magnesia, MgO, while a certain
proportion of the alumina, AI2O3, by ferric oxide, Fe203. If the composition
of the mineral is to be represented by a simple forrnula of the garnet type, the sum
of the lime, magnesia, and ferrous oxide and not their individual proportions must be
taken to represent the bases, and the sum of the alumina and ferric oxide must be
taken as one sesquioxide. An analyses will then be represented by the formula,
(Ca, Mg, Fe)0.(Al, Fe)203.3Si02 ; or more generally R^CRg^'Og-SSiOg.
Example.- — The analysis of a sample of aideritc furnished :
COa FeO MnO MgO CaO
38-5 55-6 2-8 IS I'O Total, 99-7.
The basic elements belong to the same isomorphous group. To find if the mineral is of
the type RO.CO2, where R may represent Fe, Mn, Mg, or Ca, it is usual to first calculate
the percentage amount of oxygen in the acids and bases. This gives
Basic oxygen in
Acidic oxygen in "^ :: — -
CO 2 FeO MnO MgO CaO
280 12-3 0-6 0-7 03 Total, 13-9.
There is thus 139 per cent, of oxygen in the basic radicles and 28'0 in the acidic radicles.
This is very nearly in the ratio 1*2 required by the general formula RCCOg. The formula
for this type of mineral may therefore be written, (Fe, Mn, Mg, CajCOj ; or (Fe, Mn, Mg,
Ca)0.C02; orRO.COa; or RCO3.
Instead of proceeding in this way, the amounts of the isomorphous bases CaO, MgO,
MnO, which can be replaced by the equivalent amounts of the isomorphous FeO, can be
calculated. Thus, what amount x of FeO is equivalent to 2-8 of MnO ? The molecular
weight of FeO is 72, and of MnO 71, consequently the proportion 71 : 72 = 2*8 : x gives
nearly a; = 2'8. By treating the MgO and CaO in a similar way, 3'2 and 1*3 are obtained
respectively. The basic radicles are thus equivalent to 55-6 + 2-8 + 3-2 + l-3 = 62
of FeO. The composition of the idealized siderite is thus FeO, 63 per cent., CO2, 38-5
per cent. Converting these numbers into molecular ratios in the usual manner, the formula
of the idealized siderite becomes FeCOa.
Consequently, while mixed crystals or solid solutions are not accepted as chemical
units, individuals, or compounds, in mineralogy, those mixed crj^stals which give
rise to known minerals are regarded as mineralogical individuals ; and mineralogy
merges into chemistry when all the possible products of isomorphous crystallization
are taken into account whether they occur in nature or are known solely as laboratory
products. The so-called earthy and non-crystalline minerals are regarded as bearing
the same relation to possible crystalline minerals as an amorphous precipitate of,
say, barium sulphate bears to the mineral barytes.
§ 13. Index o£ Refraction and Dispersion
The physical character of any chemical compound, and its composition are dependent
each on the other, and present two aspects of the same problem, which can never be solved
but by a consideration of both. — T. S. Hunt (1891).
When a ray of light travels from one medium to another of different density,
it is refracted, bent, or deflected towards or away from the vertical according as the
density of the second medium is greater or less than the first. Between 1620 and
1625, \V. Snell discovered the law — the law of sines — which determines the angle of
deflection. If a ray of light, 10, Fig. 95, enters the denser medium at 0, not normal
to the surface, it travels through the new medium along the path OR, such that if
NON' is perpendicular or normal to the surface, and i the angle which the incident
ray makes with the normal, and r the angle which the deflected or refracted ray also
makes with the normal, the ratio of the velocities of the light in the two media is
CRYSTALS AND CRYSTALLIZATION
671
proportional to the ratio of the sines of the angles of incidence * and refraction r,
and the ratio of the sine of the angle of incidence to the sine of the angle of reflection
has alifays the same numerical value ; that is,
Velocity in rarer medium, Fj sin i
Velocity in denser medium, V^, sin r
Constant
(1)
Fig. 95.
The constant is usually symbolized by /x or n. This ratio, sin i/sin r, is called the
index of refraction. The index of refraction can also be regarded as a number
which expresses the ratio of the velocity of light in vacuo to
its velocity in the medium, and the index of refraction less
unity, jLt— 1, expresses the fractional shortening of the velocity
which occurs when light passes through a transparent sub-
stance. For instance, the refractive index of air is 1-000292 at
0° and 760 mm. This means that while passing through air
of standard temperature and pressure, light is retarded nearly
three-ten-thousandths of its velocity in vacuo. The magnitude
(/Lt— 1) X 10^ is sometimes called the refractivity of the substance.
The determination of the index of refraction involves
fundamentally the measurement of these two angles, and the
intruments used for this purpose are called refractometers, etc.
A ray of white light is spread out or dispersed into a number of coloured rays
when it passes from one medium to another ; and, in consequence, a definite
ray of light, corresponding with a definite part of the spectrum, is used for the
measurement of indices of refraction, and two such points are used for the measure-
ment of the dispersion. The index of refraction varies with the wave-length of the
particular ray so that the index of refraction with rays from
different parts of the spectrum is different. This is illustrated
diagrammatically in Fig. 96, which shows how the rays at the
violet F-end of the spectrum are refracted more than those at
the red R-end. The earlier observations of the indices of
refraction were not very precise, for they were referred to a
part of the spectrum having a particular colour. Each colour
occupies a certain range in the spectrum whose boundaries are
not well defined, since two independent observations with the
same coloured ray may really refer to different parts of the
spectrum. The more conspicuous lines A, D, F, and H of
the solar spectrum, or the a-, j3-, and y-lines of the hydrogen spectrum are
in use, and they are recorded respectively as /t^, jjLj^, /u,^,, (jl^, jtXa, H'p, and jj^.
The wave-lengths of the chief Fraunhofer lines in Angstrom units, are
Fig. 96.
A
7677
K
B
6867
C
6563
Ha
Red
D
5893
Na
Yellow
5270
Green
P
4861
Ha
G
4841
Hy
h H
4103 3969
H6
Blue
The index of refraction for sodium light is a characteristic constant, e.g.
Cryolite.
1-34
Potash alum.
1-46
HaUte.
1-54
Baryt€s.
1-64
Xenotime.
1-72
Strengite.
1-81
Diamond.
2-42
The index of refraction of crystalline substances may vary with the direction in which
the ray of light is propagated ; it may also vary with temperature. J. L. C. Schroeder
van der Kolk, in his Tabellen zur 7nikroskopischen Bestimmung der Mineralien nach
ihrem Brechungsindex (Wiesbaden, 1906), has arranged a long list of minerals in the
order of the refractive indices of their crystals with the idea of facilitating
the rapid identification of small fragments of the mineral from measurements
of their refractive index. In some cases, the method needs supplementing by other
tests because a very limited number of minerals have indices of refraction so nearly
672 INORGANIC AND THEORETICAL CHEMISTRY
alike, and within the limits of experimental error, that they cannot be distinguished
with certainty.
The relation between the index of refraction and density.— As a rule, substances
with the greatest density have the greatest index of refraction, but there are a few
exceptions — e.g. methyl iodide has a density 2 258 and an index of refraction 1524: ;
while quinoline has a density 1 '095, and an index of refraction 1 '609. Isaac Newton i
argued from the corpuscular theory of light :
If light be swifter in bodies than in vacuo in the proportion of the sines which measure
the refraction of the bodies, th© forces of the bodies to refract light are veiy nearly propor-
tional to the densities of the same bodies excepting that unctuous and sulphurous bodies
refract more than others of the same density.
Newton then demonstrated that in spite of differences in density, D, the ratio
(/Lt2— 1)/Z> is approximately constant ; P. S. de Laplace (1805) put Newton's proof
on a sounder theoretical basis.
If the molecular force, like gravitation, is proportional to the mass, the force exerted
by a body on a corpuscle of light near its sxu-face will be proportional to the density of the
substance when, unlike gravitation, the molecular force is sensible at insensible distances,
and insensible at sensible distances. P. S. de Laplace then showed how the change in the
velocity of the light corpuscles occurs only at the superficial transitional layer between
the aether and the particles of matter ; and he also showed how this change can be expressed
as a definite integral which reduces to jit' — 1, and which is proportional to the density
D, such that (ft'' — 1 )/-D is a constant.
Serious experimental work on the relation between the index of refraction and the
density of a substance was undertaken by J. P. Biot and E. J. D. Arago in 1806
and by P. L. Dulong in 1826, in order to test Newton's formula (jli^— l)/D=constant.
As a result the formula was found to be an accurate description of the relation
between the index of refraction and density of gases. The agreement between
theory and experiment did not prove the truth of the corpuscular theory of light,
for that hypothesis was soon afterwards abandoned in favour of the undulatory
theory ; but of course the coincidence still remained as an empirical fact.
In 1858, J. H. Gladstone and T. P. Dale 2 tried how Newton's formula held for
substances other than gases, and found the empirical rule
= Constant
to be much more nearly in accord with their observations than the older formula.
This expression had been found by D. Beer, in 1853, to hold good for gases. Mean-
while, in 1880, L. Lorenz of Copenhagen and H. A. Lorentz of Leyden independently
deduced the relation :
g=i . 1 = Constant ; or ^ . ^^ = Constant
H. A. Lorentz developed the expression from the electromagnetic theory of light,
and also from the electron theory. L. Lorenz worked from the undulatory theory
of light.
L. Lorenz assumed that the body is isotropic and consists of spherical molecules
between which light is propagated with the same velocity as in free space. He further
assimaed that in the mixed discontinuous medium- — ^aether and molecules- — the light may
be regarded as if it were propagated with a definite mean wave-length and mean velocity,
and with a periodically varying amplitude of vibration. As a matter of fact the wave-
length of the light in a molecule may be much smaller than it is in the aether. Hence
L. Lorenz really replaced the actual discontinuous medium by a hypothetical mean medium,
a method which is not justifiable a priori.
All three expressions hold fairly well for gases at ordinary pressures because when
jx is nearly unity, as is the case with gases, (/Lt-f l)/(/>t^ + 2) is nearly f ; and for gases,
CRYSTALS AND CRYSTALLIZATION 673
therefore, H. A. Lorentz and L. Lorenz's expression is nearly equivalent to J. H.
Gladstone and T. P. Dale's formula multiplied by § ; and to twice I. Newton's
formula. As a result of numerous comparative experiments on liquids and solids,
it has been found that neither J. H. Gladstone and T. P. Dale's nor H. A. Lorentz
and L. Lorenz's formula is entirely in accord with all the facts. Each formula has
its own special advantages, and each breaks down under quite different experimental
conditions. Many other formulaB have been proposed — e.g. by E. Ketteler (1888) 3
and W. F. Edwards (1894) — but after much theoretical and mathematical work, no
formula yet proposed has been proved entirely in accord with observations. This
probably means that the simplifying assumptions, used in deducing the formulae,
want revising.
Nomenclature. — The ratio (/x — l)/i> or (/x2_i)/(^2-f2)i) for an element or compound
is called the specific refractory power or the specific refraction of the substance. If the
index of a substance is I'SOl, and the specific gravity 0-880, the specific refraction by
H. A. Lorentz and L. Lorenz's formula is (I'SOl^ — l)/(l-50l2 + 2)0-880=:0-337 ; and by
Gladstone and Dale's formula (1-501 — l)-f-0-880=-0-57. The product of the specific
refraction and the molecular weight of a compound is called the molecular refraction, and
the product of the atomic weight and the specific refraction of an element is called the
atomic refraction of the element. In measuring the index of refraction, light of a definite
wave-length is employed, and the difference between the indices of refraction of a substance
for light of two definite wave-lengths — say the red (?) ray and the blue (?) ray of the hydrogen
flame ; or the C and D lines of the solar spectrum- — is called the specific dispersion of
the substance, and the difference between the molecular refractions of a substance for
light of two definite wave-lengths is called the molecular dispersion ; and similarly, for the
atomic dispersion, and for the specific dispersion. The dispersive power was represented by
J. H. Gladstone as the ratio of the dispersion and the density— thus, the specific dispersive
power when multiplied by the molecular weight, furnished the molecular dispersive power.
Thus, the specific dispersion for the H and ^-rays of the spectrima is /a^— /-t^, and the
specific dispersive power is (/li^— ^^)/Z>. The dispersive power of a substance has also
been defined as the ratio of the specific dispersion of the index of refraction of the mean
ray less unity. Thus, this dispersive power is (/x^— jLt^)/(/Xjp — 1). The term //.— 1 in
J. H. Gladstone and T. P. Dale's formula is sometimes called the refractive energy of
the substance, and (/x — 1)/D, the specific refractive energy. The term /x— I is now usually
called the refractivity of the substance.
The refractivities or refractive indices of gases can be corrected for temperature
and pressure either by reference to the equation pv=RT, or to one of the many
corrected forms ^^ — J. D. van der Waals', D. Berthelot's, etc. H. A. Lorentz and
L. Lorenz's formula is generally preferred because it has a sounder theoretical
foundation ; and it is more generally adaptable to the experimental material than
J. H. Gladstone and T. P. Dale's — e.g. it is valid for a wider interval of temperature,
and it is less influenced by the state of aggregation of the substance under
examination.
G. Quincke (1883) tested the two rival formulae by varying the density by hydro-
static pressure, and calculated the values for the density and accordingly also the
compressibility, from the observed pressure and the index of refraction. He
found H. A. Lorentz and L. Lorenz's formula gave values too small ; I. Newton's
too large ; and with J. H. Gladstone and T. P. Dale's formula the values were some-
times too large and sometimes too small. The mean percentage errors were respec-
tively —14, +17, and ± 16. H. Landolt and R. Weegmann ^ found the results
indicated in Table IV showing that J. H. Gladstone and T. P. Dale's formula fails
signally in bridging the gap in density between liquid and vapour. The general
conclusions of G. Quincke's and H. Landolt's comparison of the two rival formulae
are :
(1) When the same specimen of a substance has been examined in the
liquid and gaseous state, H. A. Lorentz and L. Lorenz's formula has proved
superior in the marked degree.
(2) The efiect of increasing the density by hydrostatic pressure fits better
with J. H. Gladstone and T. P. Dale's formula than with H. A. Lorentz and
L. Lorenz's.
VOL. T. 2 X
674
INORGANIC AND THEORETICAL CHEMISTRY
(3) Both formula usually apply well to the change from the liquid to
the solid state, but the results are more in favour of J. H. Gladstone and
T. P. Dale's formula.
(4) J. H. Gladstone and T. P. Dale's formula gives more accurate results
when applied to calculate the index of refraction of a mixture from those of
its constituents, for it varied on the average 0*05 per cent. — the deviation in
the worst case was 0*16 per cent. ; with H. A. Lorentz and L. Lorenz's formula
the average deviation was 016 per cent. , in the worst case 06 per cent. Hence,
J. H. Gladstone and T. P. Dale's formula gives the better results with chemical
optical analysis when it is used, say, to determine the amount of a particular
substance dissolved in a solvent from the observed index of refraction and
density.
Table IV.— Comparison of Lorentz and Lorenz's and Gladstone and Dale's
Formula for Specific Refraction.
1 /x2-i
li-l
D /M2+2
D
Compound.
Liquid.
Liquid.
Vapour,
100°
Vapour,
100°
10»
20°
10°
20°
Water .
0-2062
0-2061
0-2068
0-2068
0-3336
0-3101
Ethyl alcohol
0-2804
0-2807
0-2825
0-4582
0-4581
0-4237
Ether .
0-3026
0-3029
0-3068
0-4935
0-4930
0-4599
Ethyl acetate
0-2547
0-2549
0-2683
0-4174
0-4172
0-4024
Ethyl iodide .
0-1557
0-1558
0-1571
0-2663
0-2658
0-2356
Chloroform .
0-1790
0-1791
0-1796
0-3000
0-2996
0-2694
Carbon disulphide .
0'2805
0-2809
0-2898
0-4977
0-4970
0-4348
(5) The refraction equivalents^ — vide infra — obtained by J. H. Gladstone
andT. P. Dale's formula for carbon dioxide, oxygen, hydrogen, nitrogen, and the
halogens, hold good for H. A. Lorentz and L. Lorenz's formula ; but according
to J. W. Briihl, the latter formula is preferable since it gives a smaller percentage
error in the calculation of the refraction equivalent of a molecule from those
of its atoms.
These facts, said W. Sutherland (1889), ought to furnish logicians with instructive
examples in the theory of evidence. H. Dufet (1883) 6 and W. Sutherland (1889)
have further shown that a theoretical foundation can be given to the empirical rule
of J. H. Gladstone and T. P. Dale.
If Fo be the velocity of light in vacuo, i.e. in free aether ; F,, the velocity in passing
through the atoms themselves ; and F, the velocity in an atom-strewn medium, then in
a unit length of path' — since velocity x time = distance traversed— the time occupied
in passing through a given medium will be the reciprocal of the respective velocities ; and the
loss of time in the passage of light through the atom-strewn medium, owing to the retarda-
tion produced by the atoms, will be proportional to the retardation in passing through
an atom, and
?=*KF.-fjS
where DjM represents the number of atoms in unit volume ; I the mean length of the
path through an atom ; a the mean sectional area of an atom ; and k is the constant
of proportion. When multiplied through by Vq, and substituting /x for Fo/F from
(1), this relation reduces to {fj,— l)M/D=kl8{N — l), where fj, denotes the refractive
index of the medium, and iV that of the substance of the atom itself ; M/D, the atomic
domain ; and the product is, the average volume of the atom itself. W. Sutherland
assumed that the bracketed term on the right is constant, so that the expression reduced
to J. H. Gladstone and T. P. Dale's rule, (/a— !)/£) = constant. The delay produced by
CRYSTALS AND CRYSTALLIZATION
675
matter is due to the breaking up of the front of the light-wave by the interspersed
atoms, and the subsequent loss of time in travelling from atom to atom before the front
of the wave recovers its plane form. Assuming that this retardation is proportional to
the length of the path, and to a function of the density, say, a'D + bD^-\- . . ., then
for unit path, {iJL—l)M/D=kl8{N — l)-{-m{a+bD-{- . . .) ; and the specific refraction is
{fi — l)/D=[kls{N — 1)/M -l-a']-\-bD, where the bracketed term is constant, say a.
W. Sutherland's formula for specific refraction resembles J. H. Gladstone and
T. P. Dale's specific refraction, but a term hD, proportional to the density, is added
to the constant. This furnishes.
D
= a+6Z)
where a and b are constants to be evaluated from observations of ft and D at two
different temperatures. For gases and vapours, the term bD may be neglected pn
account of the smallness of the term D. W. Sutherland also claims that the revised
formula is the best yet advanced, and that it is capable of representing the relation
between the index of refraction under all circumstances within the limits of experi-
mental error, where the uncorrected formula of J. H. Gladstone and T. P. Dale
fails. Some examples are indicated in Table V.
Table V. — W. Sutherland's Formula for Specific Refraction.
Compound.
Constant
b
Liquid.
Vapour (100°).
10° Obs.
20° Obs.
Observed.
Calculated.
Ether .
Ethyl acetate
Ethyl iodide .
Chloroform .
Carbon disulphide .
0-044
0-017
0-022
0023
0-050
0-4935
0-4174
0-2663
0-3000
0-4977
0-4930
0-4172
0-2658
0-2996
0-4970
0-4599
0-4024
0-2356
0-2694
0-4348
0-461
0-407
0-223
0-265
0-435
The agreement between the observed and calculated results in all cases excepting
ethyl iodide is good.
E. T. Wherry ' found that with the refractive indices for the ordinary wray and the
extraordinary e-ray of tetragonal crystals are related with the axial ratio so that
a>8+2 6^-2 a
where the axial ratios are based on the atoms present, and not on the standard axial ratio
obtained by taking the most prominent pyramidal form to be (111). There are some
disturbing factors with complex compounds, but with the simpler compoimds the rule
gave good results with a few organic compounds, and minerals of the zircon group. With
cassiterite, the two ratios are 0*945 and 0-951 ; with rutile, 0-926 and 0-911 ; this is taken
to mean that in these minerals the space lattice must have the same number of layers of
atoms in the horizontal as in the vertical direction. In zircon and xenotime, the refraction
ratios are equal to 3 : 2 times the standard ratio c : a, meaning that in the unit cell of these
minerals there are three horizontal layers of atoms for every two vertical layers.
The effect of pressure on the refractive index. — The effect of variations of
pressure on the index of refraction of gases has been investigated between 0*05 and
200 atm. W. Kaiser ® found that for pressures between 20 and 760 mm., with
sulphur and carbon dioxides, the variation of the index of refraction with pressure,
dfx/dj), increases faster than the variation of the density of the gas with changes
of pressure ; with pressures higher than atmospheric, E. Mascart ^ found the
relation [x=l+ap-}-bp^ described his results for air, nitrogen, oxygen, carbon
monoxide, carbon dioxide, nitric oxide, nitrous oxide, and cyanogen very well.
67G INORGANIC AND THEORETICAL CHEMISTRY
The effect of pressure on the refractive index of a few liquids (Na light) is as
follows :
Water (20°) Alcohol (17-5°) Benzene (20°) Ether (8°) Carbon disulphide (20°)
d/i/rfp . 0-00001514 000004174 0-00005060 0-00006161 0-00006583
The effect o! temperature on the refractive index. — The exact relation between
the temperature and the refractive index has not been established. J. P. Biot^®
represented his measurements of the refractive index of gases between 0° and 25°,
by fi=fiQ—aJd, where jjlq represents the refractive index at 0°, and a is a constant
independent of temperature. V. von Lang added another term to Biot's formula for
his results between 0° and 100° ; thus, for air, he used /x=/xo— OO69O50+O-O72^2,
E. Mascart used the more complicated expression: {fjL— 1) {I -\-ad)=coiist&nt.
If the relation (/a— 1)Z)= constant be valid, then, for ideal gases, at pressure p
and j)q with the corresponding temperatures 6° and 6q°,
^ Po
and hence /x— 1 varies inversely as (l+ct^). He assumed that the temperature
coefficient of refraction a, and the ordinary coefficient of thermal expansion a to
be the same. The observed difierences showed that a was about 12 or 15 per cent,
greater than a' ; V. von Lang found the opposite, for a' was greater than a ; and
J. R. Benoit obtained a=a'. Hence G. W. Walker made some careful measurements
of the two constants, and found a to be less than a' for air and hydrogen by respec-
tively 0*047 and 0'03l6 ; and a greater than a' for carbon and sulphur dioxides
by respectively 0*049 and 0-0326. The influence of temperature on the refractive
index of a number of minerals and liquids has been represented by formula of
the type fjL=a-\-hd-\-cd^-\- . . . Since changes of temperature are always
accompanied by changes of density, it is generally assumed that the velocity of
the propagation of light in a body is not affected by variations of temperature except
in so far as the density of the substance is simultaneously altered ; and the influence
of density, D, is given by J. H. Gladstone and T. P. Dale's or L. Lorenz and H. A.
Lorentz's formula. H. D. Ayres i^ found both Lorentz and Lorenz's and Gladstone
and Dale's formulse to apply equally well at — 189*2° between 10*1 and 149*5 cm.
pressure of mercury. He found that the refractivity /a— 1 of hydrogen, oxygen,
nitrogen, and carbon dioxide varied lineally with the density at temperatures from
0° to —189*0°. K. Scheel compared the calculated values of the density D at —190°
with the values D' observed by M. W. Travers and G. Senter, and by A. Bestelmayer
and S. Valentiner. Assuming that the density of the gas is unity at 0° and 760 mm.,
the density D at 6° were calculated, by K. Scheel, from (/x— l)/Z)=constant,
and the observed densities of hydrogen and nitrogen between 0° and —190° were
found to be about 0*4 per cent, greater than the calculated values.
The refractive index at the critical temperature.— P. A. Guye 12 has shown that
J. D. van der Waals' constant h is related with the molecular refraction /x, by the
expression ;
where M denotes the molecular weight, D the density, and k is a. constant which,
according to I. Traube, is equal to 4*03. With the 35 inorganic compounds examined
by P. A. Guye,
'% fi2+2 D
where He denotes the absolute critical temperature and pc the critical pressure, if
Vc denotes the critical volume. J. D. van der Waals showed that iic=^b, and that
PcVc=2l7QTc. This means that the refraction constants of a substance are
CRYSTALS AND CRYSTALLIZATION
677
independent of the pressure, temperature, and state of aggregation, and for all
substances at the critical temperature Vc=1126. V. Smith has compared this
deduction for the gases and liquids indicated in Table VI.
Table VI. — The Critical Values of fj. for Some Gases and Liquids.
e
H'n^"
T
Critical value
of/t
Deviation,
per cent.
Gases:
Oxygen
0°
1-000271
-118-8
1-126
0
Ethylene
0°
1-000723
13-0
1-124
-0-2
Carbon dioxide
0°
1-000449
31-35
1-109
-1-6
Sulphur dioxide
O*'
1-000686
1560
1-128
+0-2
Nitric oxide .
0°
1-000576
35-4
1-110
-1-4
Liquids :
Ammonia
16-5°
1-325
131-0
1-120
-0-5
Hydrogen chloride .
10-5°
1-254
52-3
1-109
+0-4
Hydrogen bromide .
10-0°
1-325
91-3
1126
-1-5
Chlorine
140°
1-367
148-0
1-131
0
Carbon dioxide
15-5°
1-192
31-35
1-101
-2-2
The effect oJ dispersion on the refractive index.— The effect of dispersion on
the index of refraction fju for rays of wave-length A can be calculated from A. L.
Cauchy's formula,i3 fjL=a-\-b X—'^-\-c A— *+ • • -, where a, 6, c, . . . are constants
to be computed from measurements of /x and A. The constant a is sometimes
called A. L. Cauchy's coefficient of refraction ; h, c, . , . are coefficients of
dispersion. Several other formula have been proposed — e.g. by F. Kedtenbacher,
and by C. Briot — and they all give quite good results for the visible part of the spec-
trum, but fail as the invisible red is approached, S. P. Langley found C. Briot's
formula gave the best results with the invisible red rays, but even these were not
satisfactory. A. L. Cauchy's formula is usually preferred because of its simplicity.
It agrees very well with observations for substances of low dispersive power, but
not so well with substances of high dispersive power. The relation between the
refractivity and the wave-length is then represented by an expression of Cauchy's
type:
/^-1=<1+A^)
where a and h are constants. For mercury, a=0'001755, and 6=22 '65x10— n.
It is found that in four cases — helium, argon, krypton, and xenon — where the
measurements are available, if the refractivity /x — 1 for infinite wave-length A be
plotted against the value of h in this formula, the result is a straight line.^*
It has been shown, by F. L. Perrot, A. E. H. Tutton, etc.,i5 that the index of refrac-
tion and the dispersion of a series of isomorphous crystals usually increase when one
element is replaced by another with a greater atomic weight. For example, potas-
sium, rubidium, and caesium sulphates have respectivelv the values Z)=l"4947,
1-5113, and 1-5644; and for anhydrite, CaS04, Z)=l -57518 ; celestine, SrS04,
1-62367; barytes, BaS04, 1*63717 ; and anglesite, PbS04, 1-88226.
The relation between the refractive index and chemical composition. — In 1826,
P. L. Dulong 1^ concluded from his experiments that the specific refraction of a mixture
of gases is the mean of the specific refractions of the constituents calculated for the
partial pressures of the gases in the mixture ; while the specific refraction of a com-
pound is not a mean of those of the component gases, for it is sometimes greater and
sometimes less. J. H. Gladstone and T. P. Dale (1863) investigated the effect of
chemical constitution on the refractive energies. One of the most important facts
developed by this study from a chemical point of view is that the refractive
678 INORGANIC AND THEORETICAL CHEMISTRY
equivalent of an atom is not a constant, but depends upon the way the atom
is linked with other atoms. J. H. Gladstone and T. P. Dale said :
We sought to determine the amount of change in the optical properties which results
from a replacement of one element by another, the type remaining the same, ... in order
to attain a knowledge of the action of the individual elements on the rays of light transmitted
by them. . . . The general conclusion is that every liquid has its own specific refractive
energy composed of the specific refractive energies of its component elements, modified by the
memner of combination, and which is unaffected by change of temperature, and accom-
panies it when mixed with other liquids.
The subject was followed up by H. Landolt (1864), J. W. Briihl (1886), etc.,
and as a result, it was found that the molecular refraction or dispersion of com-
pounds is :
(1) An additive property in that it depends on the number and kind of atoms
in the molecule — e.g. the atomic refractions and dispersions of hydrogen and
chlorine are virtually the same whether they are free, or combined.
(2) It is also a constitutive property in that it depends on the mode of
combination of the difEerent elements— e.^f. the atomic refraction and dispersion
of carbon is very difEerent according to the way it is combined— single, double,
or triple-bonded carbon atoms have difEerent values. While the atomic refrac-
tion of single-bonded carbon and carbonyl, CO, carbons are nearly the same,
the atomic dispersion of the one is nearly double that of the other. Similar
remarks apply to oxygen, and more particularly to nitrogen.
The refractive indices of the elements gaseous at ordinary temperatures have been
measured directly ; the opaque metals do not lend themselves to this treatment
although the method has been used in case of a few metals which can be beaten into
thin enough sheets to permit the passage of light. A. Kundt ^^ determined values
for half-a-dozen metals in this way ; the values for gold, silver, and platinum so
determined are respectively 0'58, 0*27, and 1*64. P. Drude also developed a method
for measuring the refractive indices of the metals which are based on the angle of
reflexion, and accordingly obtained values for the refractive indices of over a dozen
metals. The results of A. Kundt and of P. Drude did not agree very well. The
molecular refractions of a number of compounds has been measured directly, and
values for the constituent elements have been computed on the assumption that
the atomic refraction is an additive quality. The indices for a number of inorganic
salts have also been estimated from the indices of refraction obtained for their
aqueous solutions on the assumption that J. H. Gladstone's additive formula
Molecular refraction = (18w-}-M)J?i — l^nR^
holds good. Here M denotes the molecular weight of the compound ; R^ and J?2
the refraction constants of water and of the solution respectively ; and n denotes
the number of gram-molecules of water per gram-molecule of salt. After comparing
the observational data of a large number of compounds, J. W. Briihl drew up tables
of the atomic refractions and dispersions of a number of elements based on constants
calculated from observations on the refractive index reduced by Lorentz and
Lorenz's formula. Different constants are obtained when J. H. Gladstone and
T. P. Dale's formula ^^ is used. Table VII contains values of the atomic refrac-
tion, w{iJ?—\)l{iL^-\-2)D, compiled by W. A. Roth and F. Eisenhohr in their
Refractometrisches Hilfshuch (Leipzig, 1911).
The additive or mixture law. — The molecular refractions of a number of organic
compounds were found by H. Landolt to be the sum of the atomic refractions of their
constituent atoms when due allowance is made for the modifications in the atomic
refractions of elements united in special ways. If a substance of molecular weight
M contains Ui atoms each of atomic refraction R^ ; n^ atoms of atomic refraction
R.^\ . . . then the molecular refraction is
Molecular refraction=Wii2i-|-?22^2"i~ • • • =^^R
CRYSTALS. AND CRYSTALLIZATION
679
The results of this method of investigation have established the proposition
that the atoms of the elements have the power of retarding light ; and that in some
cases this power is not materially changed when the atoms pass from one compound
Table VII.' — Atomic Refractions and Dispersions- — w{^^ — l)l{ii*-\-2).
Carbon
Hydrogen .
Carbonyl oxygen
Ether oxygen
Hydroxyl oxygen
Chlorine
Bromine
Iodine
Ethylene bond .
Acetylene bond . ,
Nitrogen in primary aliphatic amines
Nitrogen in secondary aliphatic amines
Nitrogen in tertiary aliphatic amines
Nitrite nitrogen ....
Nitrogen in amides (C-N-C)
Atomic refraction.
2-413
1-092
2-189
1-639
1-522
5-933
8-803
13-757
1-686
2-328
2-309
2-475
2-807
3 064
3-740
2-418
1-100
2-211
1-643
1-625
6-967
8-865
13-900
1-733
2-398
2-322
2-499
2-840
3-070
3-776
B/3 Hy
2-438
1-115
2-247
1-649
1-531
6-043
8-999
14-224
1-824
2-506
2-368
2-561
2-940
3-108
3-847
2-466
1-122
2-267
1-662
1-641
6-101
9-152
14-621
1-893
2-638
2-397
2-603
3-000
3129
3-962
Atomic
dispersions.
Hp — Ha Hy — H,
0-026
0-023
0-067
0-012
0-006
0-107
0-211
0-482
0-138
0-139
0-059
0-086
0-133
0-055
0-139
0-066
0-029
0-078
0-019
0015
0-168
0-340
0-775
0-200
0-171
0-086
0-119
0-186
0-065
0-220
to another ; >but closer investigation shows that the specific refractive energies of
the atoms are greatly modified by the nature of the combination. The relation
between the different atoms is an important factor. Thus, J. W. Briihl showed that
the refractive effect of oxygen is greater when the oxygen is united to carbon than
when it is united to two other elements ; and E. Conrady i^ further showed that still
a different value for oxygen is obtained with oxygen in the ethers. J. H. Gladstone
showed that hydrogen in the weak acids has but 40 per cent, of its value in the strongly
ionized acids ; M. le Blanc found two distinct values for chlorine ; E. Wiedemann
found two for sulphur ; and seventeen values have been obtained for nitrogen.
In some cases, the additive rule is applicable more particularly when the compound
contains but a few elements combined in the same way ; but in general, the additive
mixture law breaks down completely for chemical compounds. It is, however, valid
for mere mixtures which exert no chemical action on one another. The observed
and calculated values for air, for example, coincide — within the limits of experimental
error. The principle has been applied to the technical analysis of gases, and of
many solutions. The molecular refraction (and likewise also the molecular dis-
persion) of compounds calculated in this way may then furnish concordant results ;
and the results have been used as circumstantial evidence in favour of particular
hypotheses about the constitution 20 of the compound under investigation.
Examples.- — (1) Assuming that in benzene, CgHe, there are two ethylene linkages,
compare the calculated and observed molecular refractions, given the index of refraction is
1-50144 and the density 0-880, and the molecular weight 78. From Table VII (6 X 2-365)
+ (6 X 1-103) + (3x1-836) =26-3; etc.
(2) Compare the atomic refraction and dispersion of hydroxylamino on the assumption
that the formula is HgN — OH.with the observed values 7-23 and 0*19 respectively.
N in ammonia
O in OH group
3 H atoms
Atom refraction.
. 2-497
. 1-610
. 3-150
Dispersion.
0072
0-019
0-120
HaN-^OH . 7-257
Found 7-23
The approximate agreement of the observed and calculated values of the atomic refraction
0-211
0-19
680
INORGANIC AND THEORETICAL CHEMISTRY
and dispersion is taken to favour the hypothesis that the constitution of hydroxy lamine
is H2=N — OH. Hydrogen peroxide furnishes another example.
The errors involved in the computation of the refractive equivalents of some of
the elements are sometimes as great as 12 per cent. ; and when it is remembered that
in comparing the observed and calculated refractivities of a compound, the differ-
ences seldom exceed 10 per cent., it will be obvious that the errors in the data are as
large as the magnitude under observations. In such cases, the process of investi-
gation is of comparatively little value. Many attempts have also been made to
establish a relation between the dispersive power of a substance and chemical
composition,2i but the results are not so good as with refractivities, probably because
dispersion is more readily influenced by composition than refractivity.
The refractivity of compounds is an additive property for liquids and solids
when the refractive constants are determined from the compounds themselves.
In gases, this is not true. Out of sixteen gaseous compounds of which the refrac-
tivities of the components have been measured in a free state, C. Cuthbertson 22 did
not find one to agree with the additive rule — with the nitrogen compounds the
deviations varied between 5 and 14 per cent. ; with sulphur dioxide the deviation
is 18 per cent. ; with selenium hexafluoride, 30 per cent. ; and with tellurium
hexafluoride, 45 per cent. C. Cuthbertson says the failure of the additive rule is
not far to seek :
The true refractive constant of an element is, evidently, the retardation caused by the
free gaseous atom. The forces which compel atoms of different elements to combine,
and to assume the liquid or solid states, whatever they may be, are evidently very powerful,
and it is unreasonable to expect that causes which can modify other attributes of matter
in ways and to an extent which we are unable to predict should have no effect, or always
the same effect, on its power to retard light. It is therefore to the study of the gaseous
refractive indices that it is necessary to turn in the hope of obtaining really accurate infor-
mation with regard to the optical properties of matter.
Table VIII. — Refractive Indices of Some Gaseous Elements (C. Cuthbertson).
Element.
.-i=.(i+|.)
Refractivity,
(iLi--l)xlO«
Empirical
vat\r\
a
h
Ah=00
A = 5893
raiio.
Hydrogen
0-031358 6-67 x IQ-ii
135-8
138-4
Heliima .
0-046956 2-2 X 10-11
69-56
70-0
1
Neon
00001374
. —
137-4
2
Argon
0-035584 5-6 x lO-^^
558-4
567-4
8
Krypton
0-038378 6-97x10-11
837-8
854-6
12
Xenon .
O-O3I364 610-14x10-11
1364-6
1404
20
Fluorine .
0-000195
.
195
2
Chlorine .
0-000768
i —
768
8
Bromine .
0-001125
—
1125
12
Iodine .
0-00192 (violet)
0-00205 (red)
—
— "
20
Oxygen .
O-O32663 5-07 x 10-11
266-3
270-2
2
Sulphur .
0-001045 721-2x10-11
1046
1111
8
Selenium
0-001565
. —
1565
12
Tellurium
0-002495
• —
2495
20
Nitrogen .
O-O329O6I 7-7 X 10-11
290-6
297-1
2
Phosphorus
0-001162 15-3x10-11
1162
1212
8
Arsenic .
0001552
• —
1552
12
Zinc
0-002050
, .
2050
Cadmium
0-002675
. —
2675
Mercury .
0001755 22-65x10-11
1755
1866
CRYSTALS AND CRYSTALLIZATION 681
The refractive indices of gases. — J. P. Biot and F. J. D. Arago23 measured the
refractivities of hydrogen, oxygen, and nitrogen ; to these P. L. Dulong added chlo-
rine ; F. P. le Roux, iodine, sulphur, phosphorus, arsenic, and mercury ; E. Mascart,
bromine ; W. Ramsay and co-workers, the five inert gases — helium, neon, argon,
krypton, and xenon ; and C. Cuthbertson and co-workers added selenium, tellurium,
zinc, cadmium, and fluorine. The earlier less accurate determinations have also
been revised by a number of different workers. The results are indicated in Table
VIII. The refractivities of the five mono-atomic gases are almost exactly in the
ratios 1 : 2 : 8 : 12 : 20 — the last three being as 2 : 3 : 5. A similar ratio obtains
with the halogens ; and with the other families indicated in the table. There is a
4'7 per cent, error with selenium ; a 8*7 per cent, error with tellurium ; arsenic
has a deviation of about 17 per cent. No simple ratio was observed with the zinc
family, for mercury with the highest atomic weight has the lowest refractivity.
However, it must be added that some chemists doubt if mercury is rightly placed
with the zinc family. The anomalous dispersion of iodine in the red will be noticed.
The estimated refraction equivalents of the alkali metals by J. H. Gladstone also
fall in line — the values for potassium rubidium and caesium being respectively
7'85, 121, and 192. Rearranging the data,
Table IX. — Refractivities of the Elements.
Li Be . . .
Na Mg . . .
K Ca . . .
Ge As Se Br Kr Rb Sr . . .
— 129x12 130x12 93-8x12 70x12 —
Sn Sb Te I X Cs Ba . . .
— — 125x20 96x20 70x20 — >-
An examination of the horizontal rows in Table IX shows that refractivity must
be closely connected with valency ; the elements with the higher atomic weights
have the lowest refractivities although the relation between the two is not known.
J. H. Gladstone thought that the product of the specific refractivity and the square
root of the chemical equivalent is approximately constant— 1 '3 for the univalent
elements, and 1*01 for the bi-, ter-, quadri-, and quinquevalent elements. Improved
observational data do not support Gladstone's rule. An increase of valency is accom-
panied by an increase of refractivity. Only part of the refractivity is concerned with
valency, or the non-valent elements would not retard light at all.
According to the electronic hypothesis of matter, the electrons bound within
the atoms by quasi-elastic forces, are supposed to be stimulated into oscillatory
motions by incident waves of light ; the mode of motion is influenced by the orien-
tation of the atoms in the molecules. The same electrons probably also play an
important role in the union of the atoms to form the molecules. It is now assumed
that the retardation of light in passing through a material medium is caused by the
expenditure of energy in starting and maintaining the motion of the electrons which
form part of the atom. These electrons are supposed to have a natural period of
vibration of their own, and the loss of energy and velocity is greatest with rays whose
wave-length approaches most nearly to the natural period of vibrations of the
constituent electrons— the long red waves, for example, are retarded 1 to 2 per cent,
less than the shorter violet waves.
Valency and the refractive index.— According to H. A. Lorentz's theory, the
refractivity is directly proportional to the number N of electrons in unit volume of
the medium, and inversely as the difference in the frequency of the free vibrations
He
70
N
0
F
Ne
149x2
135x2
96x2
70x2
P
S
CI
Ar
149x8
138x8
96x8
70x8
As
Se
Br
Kr
129x12
130x12
93-8x12
70x12
Sb
Te
I
X
INORGANIC AND THEORETICAL CHEMISTRY
Wq of the atoms which are instrumental in effecting refraction, and n that of the inci-
dent light; wA=F, so that
^-1 = ^2
Calculations show that the numerator of this expression is proportional to N, the
number of electrons in unit volume such that C=i7r~^e^Nlm, where tt denotes
the well-known constant, m is the mass, and e the electric charge. Then if v denotes
the number of valency bonds associated with an element, P. Drude's theory of disper-
sion 24 leads to the conclusion that the sum of the valencies contained in the molecules
of a compound is proportional to the accepted chemical valencies ; or that C is pro-
portional to the positive valency v of the atom, or that C/u is a constant which has
the same value for all gases. Thus,
Table X. — The Rei.ation between Valency and the Refractive Index.
CX 10-27
V
C
V
Hydrogen .....
Oxygen
Phosphorus ....
Nitrogen .....
Sulphur
1-692
3-397
7-610
5-034
4-808
1
2
3
3
1-692
1-699
1-691
1-678
1-603
The results are not so good with the last two elements ; but the agreement in all
cases is fairly close, giving grounds for the assumption that the number of electrons
concerned in dispersion is proportional to the received valency of the element.
According to this theory, L. Natanson inferred that the product of v with the
term A should be independent of the nature of the gas, and be the same for all gases
referred to the normal state. Here :
A=
3(/Zi-/.2)Ai2A22
2(/^i~l)(i^-
■l)(Ai2.
-A22)
where fii and /Lt2 ^^^ *^® refractive indices corresponding with two different rays
of wave-length A^ and A2 respectively. The mean values of the product vA for
hydrogen, oxygen, nitrogen, carbon monoxide, sulphur dioxide, hydrogen disulphide,
and carbon monoxide are approximately constant, but deviations occur with some
of the hydrocarbons. For example :
H2
Og
Na
CO2
SO2
HaS
CO
V .
2
4
6
8
8
6
4
A .
7-99
3-87
2-73
216
2-23
2-80
3-68
vA
. 15-98
15-48
16-38
17-28
17-84
16-80
14-72
whereas the values of vA for methane (t^— 8), ethane (v=14), ethylene (v=12),
and acetylene (v=10) are respectively 2795, 2550, 32-72, and 34-45.
The study of the refractive indices of gases is therefore promising to throw light
on the intimate structure of atoms and molecules. The sterility of the enormous
amount of work which has been done on the refractivities of liquids and solids, says
C. Cuthbertson, proves that in these states of aggregation the causes which obscure
the simplicity of the results are too powerful, and that it is vain to hope for much
addition to our knowledge in this direction. The little work which has been done on
the refractivities of gases has furnished a few simple relations, which by th-eir very
simplicity seem to be the outward and visible truth which ought to be pursued by the
accumulation of more data.
The relation between the index oi refraction and the magnetic rotatory
power. — J. H. Gladstone and W. H. Perkin 25 have shown that there is some
CRYSTALS AND CRYSTALLIZATION
683
connection between the rotation of a polarized ray under the influence of magnetiza-
tion, and the retardation of the rays in passing through a material substance as
represented by the index of refraction and dispersion. The three properties are
additive in an analogous manner, and a change in the active valency of an element
is attended by a parallel change in all three properties. With the halogen acids all
three properties exhibit parallel deviations from the normal when measured in
solutions of iso-amyl ether, and in water :
Table XI.' — ^Magnetic Rotatory Power of Hydrochloric Acid.
Hydrochloric acid.
Molecular magnetic
rotation. ,
.,^-1
Molecular
dispersion.
Free
Aqueous solution
Solution in isoamyl ether .
2-187
4-412
2-238
11-20
14-45
71-36
0-54
1-12
0-51
The relation between the index o! refraction and the dielectric constant.—
The numbers in Table XI point to a relationship between electromagnetism and the
velocity of light. According to J. C. Maxwell's electromagnetic theory of light,26
if K and K^ respectively denote the dielectric constants of two transparent media,
and IX the limiting value towards which the index of refraction approaches when the
wave-length of the rays become indefinitely large, then fji^=KIKi. If one of the
two media be air, for which iLi=unity, then fJi^=K, meaning that the specific
inductive capacity, or the dielectric constant of any medium relative to air unity,
is equal to the square of the index of refraction of that medium when fi^ is measured
for the slowest vibrations of light, and K for the most rapid electrical oscillations ;
the agreement between the observed values of K and /x^ becomes wery close, as is
illustrated in Table XII.
Table XII. — Observed Values of K=fi,^.
Substance.
K
/^^
Air
1-000690
1-000688
Hydrogen
1-00264
1-000276
Phosphorus .
4-20
3-60
Selenium
5-96
6-60
Liquid chlorine
1-87
1-88
Bromine
2-57
3-10
Iodine .
4-00
4-00
Carbon disulphide
2-67
2-67
The value of K for water is 80 and jjl varies between 1 '33 and 1*34 ; and for alcohol,
K=26, and fi varies between 1-36 and 1"37. Hence, as F. Heerwagen showed,
there appears no kind of relation between fju and K for water. H. A. Lorentz
and L. Lorenz's relation ()l(,2— l)/(^2_^2)D=constant becomes {K—l)l{K-\-2)D
=constant, when K=ix^, and L. Boltzmann found the results to be satisfactory.
For gases, (Z—l)/Z)= constant, and at a constant temperature, therefore, K—l
is proportional to the pressure. The results were confirmed by A. Palaz, P. Fuchs,
A. Rosa, P. Lededew, and K. Badeker. F. Linde found that the formula is not
applicable to liquid and gaseous carbon dioxide, sulphur dioxide, nitrous oxide, and
chlorine. F. Ratz could not use the formula satisfactorily for his experiments on the
influence of temperature and pressure on the dielectric constant, while 0. Hasenhorl
obtained satisfactory results. A. Batschinsky found the formula fails when K is
greater than [jfi. R. Millikan, F. Beaulard, and V. Boccara and M. Pandolfi obtained
good results with many mixtures. S. Pagliani found that better results are obtained
684 INORGANIC AND THEORETICAL CHEMISTRY
with the formulae (A"— l)iV/A'F=constant ; and (Z— l)/A'\/-^M^=constant,
where N denotes the number of atoms in a molecule ; M, the molecular weight ;
and F, the molecular volume.
The experiments of A. P. Cole, L. Arons and H. Rubens, and of A. Ellinger 27
show that for rays of very great wave-length — say 60 to 600 cm. in air — the index of
refraction of water is nearly /x=9. Similarly, for alcohol, the index of refraction
for waves 209 cm. long in air was 5*24, and the square of this number agrees well
with the dielectric constant of alcohol for low-frequency oscillations. This shows
that J. C. Maxwell's formula, ijfi=K, is valid for these substances with radiations
of great wave-length.
According to J. A. Fleming,^ substances of simple symmetrical constitution —
e.g. liquid elemental gases, saturated hydrocarbons, paraffins — follow Maxwell's
rule yu^—K for light waves ; and they have dielectric constants between 2 and 3,
and values of /x lying between 14 and 17, and these values are not much affected by
changing the frequency of the incident waves from zero to billions per second. The
molecules of which bodies are composed can double the dielectric constant of the
intermolecular spaces without changing the qualitative characteristics of the aether —
according to Th wing's rule : the dielectric constant of these bodies is nearly 2 '6
times their density. Substances made up of molecules with groups of radicles —
hydroxyl, nitryl, etc. — do not generally follow Maxwell's rule, and they have dielec-
tric constants which are much more sensitive to changes in frequency — as a rule,
increasing the frequency decreases the dielectric constant.
Again, electromagnetic waves travel much more slowly through dielectrics than
through empty space ; with water, the velocities in space and in the liquid are as
9 : 1 for all electric waves yet produced, while for visible light waves the ratio is nearly
1*3 : 1 ; for alcohol, the ratio varies from 5 : 1 to 2 '5 : 1 in passing from the longest
to the shortest waves yet produced electrically, while for visible light rays the ratio
is 1*3 : 1. Low temperatures annul the difference in the velocity ratios for long
and short waves. For substances like paraffin, hydrocarbons, liquid oxygen, and
bodies of simple chemical constitutions there is no marked difference between the
velocities of the waves of different wave-length.
J. A. Fleming assumes that there is a slight displacement of the electric charges
(electrons) on the atoms in opposite directions when the molecules of inert substances
— like the paraffins and saturated hydrocarbons with symmetrical atoms — are
subjected to an electric force. This displacement is the same whether the stress
has a frequency of some billions per second, as in the case of a ray of light, or 100
per second, as in the case of electrical oscillations ; and Maxwell's rule is fulfilled.
On the other hand, with unsymmetrical molecules^ — like water H*— OH' — the electric
charges are so displaced that the molecules have an electrical moment, and under
the influence of an electric force, they are oriented in space like small magnets in a
magnetic field. This displacement is over and above the strain due to the charges
of the molecule in that it bestows an abnormal value on the dielectric constant.
These abnormal values become normal on lowering the temperature because the
molecules aggregate into more complex groups which no longer possess an electric
moment, and are no longer liable to orientation in the electromagnetic field.
Waves of high frequency produce the same result as very low temperatures,
because the inertia of the molecules under rapid alternations of electric force
(billions per second) prevents the orientation of the molecules.
J. Stefan (1872) 28 drew attention to a simple relation between the index of
refraction, ii, of a gas and the mean free path, /, of the molecules such that (/x— l)/=a
constant. R. Clausius (1879) developed the idea, on the lines of J. C. Maxwell's
theory, and showed that the expression (A— l)/(A-f-2) represents the ratio of the
real volume of the molecules of a substance to the volume they actually occupy en
masse. The idea was extended by 0. F. Mosotti and F. Exner (1885), who developed
the idea with respect to the equivalent expression (/x^— l)/(/x24-2) such that if the
molecules have a spherical form, and do not touch, the relation between the specific
CRYSTALS AND CRYSTALLIZATION
685
inductive capacity or the dielectric constant, K, and the space v actually occupied
by the molecules v=(Z— l)/(K+2), and since, for vibrations of great wave-length,
fJ'^=K, the spaces occupied by the molecules will be v=(/x2— 1)/(^2_^2). If the
true molecular volume be represented by v, and if 2JM represents the sum of
the atomic volumes, when Ai, A^, . . . represent the atomic volumes and Wi, 712,
. . . the number of atoms of each kind respectively present in the molecule, the
fraction of unit space actually occupied by the molecule will be EnAjV, and
I. Traube assumed this to be proportional to F, that is, to (jLt2_i)/(^2_|.2), po that :
/>t2_j-2 i"'2+2 D
where ^ is a constant, and V=MID, where M denotes the molecular weight, and D
the density. I. Traube obtained a value 3"4:6 for the constant k for a long series
of saturated organic compounds. The constant k decreases with an increase in the
number of double linkages in saturated organic compounds.
H. Da vies showed that if the absolute volume of unit mass of liquid be v, the
volume at absolute zero be Vq— the h of J. D. van der Waals' equation — and if the
volume of the liquid at the temperature T° K. be F ; then, if VQ=kV :
2akTc
K-\
K^2
1
2akTc
where a denotes the coefficient of cubical expansion. The values of the constant
k calculated from this expression for the liquids tried deviate very little from 2*5.
Hence, the ratio of the volume at the absolute zero to the real volume occupied by
a number of molecules is 2*5. H. Davies also showed that :
K-l M Mv
• -^ = ^TT^ ; and
K4-2 ' D 10
K-l M
K+2' D
-2-3^
Vc
observation gave 2'4:STclpc, when ]).■ is measured in atmospheres. P. A. Guye
found empirically that the molecular refractive power is 1 'STc/Pc
The refractive index of crystals of isomeric compounds. — ^The specific
refraction of isomeric substances has been suggested as a possible method of
distinguishing between metamerism and polymerism. With metamers the difference
in the specific refraction is small, and said to be less than about 0*5 per cent. The
following examples have been cited in illustration :
Table XIII. — Metamers.
TiOa
Specific gravity
^
Difference.
Percentage
deviation.
Anatase
Brookite
Rutile
3-840
4-065
4-239
2-5011
2-5872
2-6642
0-3909
0-3905
0-3926
0-0004
0-0017
0-1
0-4
With polymers, the difference in the specific refraction is large, say, over one per
cent. For example :
Table XIV.~
-Polymers.
CaCOg
Specific gravity,
f*
D
Difference.
Percentage
deviation.
Calcspar
Aragonite
2-713
2-960
1-5958
1-6277
0-2196
0-2128
0-0068
3-1
686
INORGANIC AND THEORETICAL CHEMISTRY
With the three aluminium silicates there appears both metamerism and poly-
merism.
Table XV.
— Metamebs
AND Polymers.
AlSlOj
Specific gravity,
/*
%^
Difference.
Percentage
deviation.
Andcdusite .
Cyanite
Sillimaxdte .
3-180
3-603
3-236
1-6367
1-7182
1-6613
0-1999
0-1883
0-2044
0-0003
0-0045
0-3
2-3
It will be observed, however, that the molecular structure of none of these
examples has been established, and the argument therefore proceeds in a vicious
circle.
Refebbnces.
1 I. Newton, Optiks, London, 1704 ; P. S. Laplace, Micanique celeste, Paris, 4. 237, 1805 ;
J. P. Biot and P. J. D. Arago, Mem. Acad., 7» 301, 1806 ; P. L. Dulong, Ann. Chim. Phys., (2),
34. 154, 1826.
2 J. H. Gladstone and T. P. Dale, Phil. Trans., 148. 887, 1858 ; ib., 153. 317, 1863 ; W. Suther
land, Phil. Mag., (5), 27. 141, 1889 ; D. Beer, Einleitung in die hohere Optik, Braunschweig, 1853 ;
H. A. Lorentz, Wied. Ann., 9. 641, 1880; Theory of Electrons, Leipzig, 1909 ; L. Lorenz, Wied,
Ann., 11. 70, 1880.
» E. Ketteler, Pogg. Ann., 124. 390, 1865 ; E. Mascart, Compt. Bend., 78. 617, 679, 1874 ;
86. 321, 1876 ; V. von Lang, Pogg. Ann., 153. 448, 1874 ; W. F. Edwards, Amer. Chem. Journ.,
16. 625, 1894 ; 17. 243, 1895 ; W. Johst, Wied. Ann., 20. 47, 1883 ; M. Zerchini, Gazz. Chim.
Italy 25. 269, 1885 ; W. Hibbert, Phil. Mag., (5), 40. 268, 1895 ; J. F. Eijkmann, Bee. Trav.
Chim. Pays-Bas, 14. 185, 1895 ; 15. 52, 1896.
* C. Cuthbertson and E. P. Metcalfe, Proc. Boy. Soc, 80. A, 406, 1908 ; L. Stuckert, Zeii.
Ekktrochem., 16. 37, 1910 ; W. J. Jones and J. R. Partington, Phil. Mag., (6), 29. 28, 1915.
^ R. Weegmann, Zeit. phys. Chem., 2. 218, 257, 1888 ; H. Landolt, Ber., 15. 64, 1882 ;
L. Bleekrode, Proc. Boy. Soc, 37. 339, 1884 ; K. Prytz, Wied. Ann., 11. 104, 1880 ; G. Quincke,
Wied. Ann., 19. 401, 1883 ;" Phil. Mag., (5), 17. 65, 1884.
« H. Dufet, BuU. Soc. Min., 6. 261, 1885; W. Sutherland, Phil. Mag., (5), 27. 141, 1889.
' E. T. Wherry, Amer. Min., 3. 134, 1918 ; Journ. Washington Acad., 8. 277, 319, 1918.
« W. Kaiser, Ann. Physik, (4), 13. 210, 1904.
» E. Mascart, Compt. Bend., 78. 617, 679, 1874 ; 84. 321, 1182, 1878 ; J. Chappius and
C. Riviere, ^nri. Chim. Phys., (6), 14. 1, 1889 ; F. Perreau, ib., (7), 7. 298, 1896 ; H. G. Gale, Phys.
Bev., 14. 1, 1902 ; P. Carnazzi, Nuovo Cimento, (4), 6. 385, 1897 ; L. Magri, Phys. Zeit., 6. 629,
1905.
i» J. P. Biot and F. J. D. Arago, M6m. Acad., 7. 301, 1806 ; E. Mascart, Compt. Bend.,
78. Q17, 679, 1874; 84. 321, 1182, 1878; V. von Lang, Sitzber. Akad. WienA^. 451, 1874;
J. R. Benoit, Journ. Phys., (2), 8. 451, 1889 ; G. W. Walker, Proc. Boy. Soc, 201. 435, 1903.
11 H. D. Ayres, Phys. Bev., (2), 2. 161, 1913 ; K. Scheel, Verh. deut. phys. Ges., 9. 24, 1907 ;
M. W. Travers and G. Senter, B. A. Bep., 646, loOl ; A. Bestelmeyer and S. Valentiner, Ann.
Physik, (4), 15. 61, 1904.
" P. A. Guye, Ann. Chim. Phys., (6), 21. 206, 1890 ; Archiv. Sciences Geneve, 23. 197, 204,
1890 ; Journ. Phys., (2), 9. 312, 1890 ; I. Traube, Ann. Physik, (4), 5. 552, 1901 ; V. Smith,
Proc Boy. Soc, 87. 366, 1912.
1' A. L. Cauchy, Mimoire sur la dispersion de la lumihe, Paris, 1836 ; C. Briot, Essais sur
la thiorie mathematique de la lumihre, Paris, 1864; F. Redtenbacher, Das Dyrmmidensysiem,
Mannheim, 1857 ; A. Wiilhier, Wied. Ann., 17. 582, 1882 ; E. Ketteler, ib., 7. 658, 1879 ; 12.
363, 1881 ; 30. 300, 1887 ; H. von Hehnholtz, Pogg. Ann., 154. 512, 1874 ; E. B. Christoffel,
ib., 117. 27, 1862 ; W. Sellmayer, ib., 143. 272, 1871 ; 145. 399, 1872 ; 147. 386, 1872 ; P. Drude,
Ann. Physik, (4), 14. 677, 1904.
" A. WiiUner, Liebig's Ann., 133. 1, 1868 ; J. W. Briihl, ib., 235, 233, 1886 ; S. P. Langley,
Phil. Mag., (5), 17. 194, 1884 ; R. Nasini and P. Bemheimer, Atti Accad. Lined, (3), 18. 608,
1884 ; (3), 19. 195, 1884 ; C. Cuthbertson, Phil. Mag., (6), 24. 69, 1912.
" F. L. Perrot, Archiv. Sciences, Genhve, (2), 21. 123, 1889 ; (2), 25. 54, 1891 ; (2), 29. 128,
1893 ; A. E. H. Tutton, Journ. Chem. Soc, 69. 344, 1896 ; W. Orthoff, Zeit. phys. Chem., 19.
201, 1896 ; F. L. Bishop, Amer. Chem. Journ., 35. 84, 1906.
!• P. L. Dulong, Ann. Chim. Phys., (2), 31. 154, 1826 ; J. H. Gladstone and T. P. Dale, Phil.
Trans., 153. 317, 1863; H. Landolt, Pogg. Ann., 117. 122, 545, 1864; 123. 595, 1864;
Liebig's Ann. Suppl, 4. 1, 1866 ; Liebig's Ann., 213. 75, 1882 ; J. W. Bruhl, ib., 235. 1, 1886 ;
CRYSTALS AND CRYSTALLIZATION 687
Zeit. phys. Chem.y 1. 307, 1887 ; 7. 1, 140, 429, 521, 1891 ; 12. 681, 1893 ; 16. 193, 226, 497, 512,
1895 ; 21. 385, 1896 ; 22. 373, 1897 ; 23. 564, 1897 ; 25. 577, 1898 ; 26. 47, 1898.
1' A. Kundt, Wied. Ann., 34. 469, 1888 ; 36. 824, 1889 ; P. Drude, ib., 34. 523, 1888 ; 36.
548, 1889 ; 39. 537, 1890 ; 42. 189, 1891 ; 64. 159, 1898.
18 J. H. Gladstone, Phil. Trans., 159. 13, 1869 ; Proc. Boy. Soc, 18. 49, 1870.
19 E. Coiirad3% Zeit. phys. Chem., 3. 210, 1889 ; M. le Blanc, ib., 4. 553, 1889 ; M. le Blanc and
P. Rohland, ib., i9. 261, 1896 ; E. Wiedemann, Pogg. Ann., 150. 380, 1876 ; R. Nasini, Ber.,
15. 28, 1882 ; Gazz. Chim. Ital, 13. 296, 1883 ; J. H. Gladstone and W. Hibbert, Journ. Chem.
Soc., 67. 831, 1895 ; 71. 822, 1897.
*° S. Smiles, The Belation between Chemical Constitviion and Some Physical Properties, London,
306, 1910.
21 A. Schrauf, Pogg. Ann., 116. 193, 1862 ; 119. 461, 1863 ; J. H. Gladstone, Proc. Roy. Soc.,
42. 401, 1887 ; Journ. Chem. Soc, 50. 609, 1886 ; J. W. BriiM, Zeit. phys. Chem., 7. 140, 1891.
2* C. Cuthheitson, Science Prog., 3. 273, 1908 ; S. Loria, Die Lichtbrechung in Gasen als physi-
kalisches und chemisches Problem, Braunschweig, 1914.
23 J, P. Biot and F. J. D. Arago, Mem. Acad., 7. 301, 1806 ; P. L. Dulong, Ann. Chim. Phys.,
(2), 31. 154, 1826 ; F. P. le Roux, ib., (3), 61. 385, 1861 ; Compf. Bend., 51. 800, 1860 ; E. Mascart,
ib., 78. 617, 679, 1874 j 86. 321, 1182, 1878; K. Scheel, Verh. deut. phys. Ges., 9. 24, 1907;
K. Scheel and R. Schmidt, ib., 10. 287, 1908; W. Ramsay and M. W. Travers, Proc. Roy. Soc,
62. 225, 1898; W. Burton, ib., 80. 390, 1908; A. Hurion, Ann. V^cole Norm., 6. 380, 1877 ;
0. M. Cuthbertson, Proc Boy. Soc, 81. 440, 1908 ; 83. 149, 1909 ; 84. 13, 1910 ; C. Cuthbertson
and E. P. Metcalfe, ib., 80. 406, 1908 ; Phil. Trans., 207. A, 135, 1906 ; C. Cuthbertson and
E. B. R. Prideaux, ib., 205. A, 319, 1906.
2* P. Drude, Ann. Physik, (4), 14. 677, 1904 ; L. Natanson, Bull. Acad. Cracovie, 939, 1909.
26 J. H. Gladstone and W. H. Perkin, Journ. Chem. Soc, 55. 750, 1889.
2« J. C. Maxwell, Phil. Trans., 155. 459, 1865 ; R. Clausius, Die mechanische Wdrmetheorie,
Braunschweig, 2. 94, 1879.
27 L. Arons and H. Rubens, Wied. Ann., 41. 580, 1891 ; 44. 206, 1891 ; E. Cohn and L. Arons,
ib., 28. 454, 1886; 33. 13, 31, 1888; S. I. Tereschine, ih., 36. 792, 1889; F. Paschen, ib., 54.
668, 1896; A. EUinger, ib., 46. 514, 1892; H. Rubens and E. F. Nichols, ib., 60. 455, 1897;
A. D. Cole, ib., 57. 290, 1896 ; F. Heerwagen, ib., 49. 276, 1893 ; P. Lebedeff, ib., 44. 304, 1883 ;
F. Linde, ib., 56. 646, 1895 ; R. A. Millikan, ib., 61. 377, 1897 ; R. Lang, ib., 56. 534, 1895 ;
L. Boltzmann, Sitzber. Akad. Wien, 69. 812, 1874 ; P. Fuchs, ib., 98. 1240, 1889 ; O. Hasenhorl,
ib., 105. 460, 1896 ; F. Hlawatz, ib., 110. 454, 1901 ; A. Palaz, Journ. Phys., (2), 5. 370, 1885 ;
A. Rosa, Phil. Mag., (5), 31. 188, 1891 ; K. Badeker, Zeit. phys. Chem., 36. 305, 1901 ; F. Ratz,
ib., 19. 86, 1891 ; V. Boccara and M. Pandolfi, Nuovo Cimento, (4), 9. 254, 1899 ; F. Beaulard,
Compt. Rend., 119. 268, 1894 ; 129. 149, 1899 ; S. Pagliani, Atti Accad. Lincei, 2. 48, 1893 ;
W. Schmidt, Ann. Physik, (4), 11. 121, 1903 ; J. A. Fleming, Journ. Soc Arts, 49. 69, 86, 97, 113,
1900.
28 J. Stefan, Sitzber. Akad. Wien, 65. 341, 1872; O. F. Mosotti and F. Exner, ib., 91. 850,
1885 ; R. Clausius, Die mechanische Wdrmetheorie, Braunschweig, 2. 64, 1879 ; I. Traube, Ber.,
29. 2730, 1896 ; J. C. Maxwell, Phil. Trans., 155. 459, 1865 ; P. A. Guye, Ann. Chim. Phys.,
(6), 21. 222, 1890; H. Davies, Phil. Mag., (6), 24. 415, 1912; (6), 23. 657, 1912; W. C. McC.
Lewis, ib., (6), 28. 104. 1914.
CHAPTER XII
THERMODYNAMICS AND THERMOCHEMISTRY
§ Ic Matter and Energy
All physical science starts from certain postulates. One of them is the objective exist-
ence of a real world.— T. H. Huxley (1887).
Substance is like a river in continual flow ; the energies undergo constant changes and
do work in infinite variety. There is hardly anything that stands still or remains still.' —
Marcus Aurelius.
Side by side with ponderable matter, capable of being weighed, early science accepted
imponderable matter such as electricity, fire, heat, etc. ; and, up to the beginning of
the nineteenth century, it was generally assumed that heat is a substance which was
variously styled caloric, igneous fluid, phlogiston, etc. Indeed, A. L. Lavoisier
was influenced by the time-honoured tradition, and in his Traite elementaire de
chimie (Paris, 1793), included lumiere and calorique in his Tableau des substances
simples to be regarded as elements, although he knew quite well that they had no
perceptible weight. Even J. Fourier, in his classical Theorie de la chaleur (Paris,
1816), regarded heat as a material substance ; and J. J. Berzelius, in his Lehrbuch der
Chemie (Dresden, 1825), classed Licht- und W drmestoff a,mong the einfache unwdgbare
Stoffe.
As a result of Newton's theory of gravitation, it was soon recognized that matter
seems to have a property, called mass, which shows itself as weight under the influence
of gravity. Matter also occupies space so that it is always extended in some shape
or form ; and further, matter is invariably associated with energy. Consequently,
matter, as perceived by the senses, possesses certain attributes — weight and form —
which appear to be permanent and essential qualities abiding in all known kinds of
matter ; whereas other properties — e.g. colour, odour, etc. — appear to be secondary
and accidental attributes which are peculiar to specific forms of matter. This
distinction between the primary and secondary qualities of matter was recognized
by Democritus, c. 350 B.C. ; it appeared among the tenets of Albertus Magnus, who
in his De generations, elementorum said :
That matter and power are the principles of each body is clear, for having taken away
all the accidental forms, we arrive at length at a substantial form which, being abstracted
per irUellectum, there remains a something very occult which is prima materia.
The idea was further emphasized by J. Locke in 1689. Matter may also be found
under different conditions of temperature, electrification, motion, etc. ; and daily
experience teaches us that changes are continually taking place in the conditions
of bodies around us. Change of position, change of motion, of temperature, volume,
and chemical combination are but few of the myriad changes associated with bodies
in general.
The forms of energy. — There are many different forms of energy— electrical,
chemical, mechanical, thermal, and actinic — and by suitable means these can be
mutually converted one into the other ; e.g. the galvanic battery converts chemical
into electrical energy, and the dynamo converts mechanical into electrical energy.
Much of the motive power used in the industrial arts is derived from the chemical
action between coal and oxygen in the furnace of a steam engine. Heat and light
THERMODYNAMICS AND THERMOCHEMISTRY 689
are also well-known concomitants of chemical action. Hence, it is inferred that heat,
electricity, mechanical motion, light, and chemical action are all different forms of
one distinct entity — energy. The different forms of energy are supposed to be the
external aspects of one single basic form of energy, which, in the words of S. A.
Reeve (1909), we may never hope to comprehend. Examples of the mutual identity
of the different forms of energy multiply daily in familiar experience. The idea was
dimly foreshadowed in the seventeenth century by F. Bacon in his essay De forma
calidi (1627), which was offered as a model of method for investigating nature. Here
Bacon argued that the facts could be satisfactorily explained only by assuming that
heat is a kind of motion among the particles of a body ; heat and mechanical motion
are mutually convertible. The same idea was accepted by R. Hooke (1667), Isaac
Newton (1675), R. Descartes (1677), R. Boyle (1680), J. Locke (1689), and others.i
The same idea was emphasized by Count Rumford 2 who, in An inquiry concerning
the iveight of heat (1798), proved that heat could not be a material substance because
unlimited quantities can be developed by friction, and concluded that heat must
be motion. H. Davy, likewise, in 1812 showed that ice can be liquefied by friction,
and the resulting liquid contains a far greater amount of heat than the ice, and he
concluded :
The immediate cause of the phenomena of heat is motion, and the laws of its communica-
tion are precisely the same as the laws of the communication of motion.
The later work of J. P. Joule (1846-9) and others has fixed the generalization
that any one form o£ energy can be transformed, wholly or partially, directly or
by intermediate steps, into any other form. This is the so-called law of transfor-
mation of energy. All types of machinery are devices for transforming energy
from one form into another ; and all phenomena in the material world can be repre-
sented as transformation of energy. Industrial operations usually involve the
expenditure of considerable amounts of energy. For instance, mechanical energy
is expended in crushing and grinding rocks, rolling metals, transporting materials,
etc. ; thermal energy is expended in the steam engine, and used for melting metals,
burning lime and cement, bricks and pottery, etc. ; electrical energy for illumination,
electroplating, refining metals, production of aluminium, etc. ; light energy for
illumination and photography ; chemical energy in the manufacture of chemical
compounds, explosives for blasting and warfare, driving gas engines, etc.
Energy and work. — As a first approximation, every change in the condition of
the various bodies around us is supposed to be due to the action of what T. Young
(1807) 2 called energy. In other words, energy is regarded as an operative physical
agent which has the power of changing the condition of bodies. Whenever a body
is changing its condition, there energy is in action. Energy is the cause, change of
condition the effect. The action of energy may be resisted. Change can take place
only when the restraint is withdrawn or overcome. The action by which energy
produces a tendency to change is called a force. The word tendency here means
that the change will take place the moment the restraining influence is withdrawn.
Force is thus supposed to be an imaginary intermediate link between the physical
cause of a phenomenon and the resulting effect. Force is thus a manifestation of
energy. Whenever resistance is overcome, energy must be expended. Hence,
energy is sometimes defined as " the power to overcome resistance." Work is said
to be performed whenever change takes place in opposition to a force opposing that
change. Work is a manifestation of the transfer or transformation of energy. The
work is done at the expense of the energy, and the amount of work is equivalent to
the quantity of energy transferred. The work performed is equal to the energy
expended, and just as quantity of matter is measured by weight so quantity of
energy is measured by work. Consequently, energy is sometimes defined as the
capacity for doing work ; or, as W. Ostwald (1892) ^ puts it : Wir werden allgemein
Energie als Arbeit, oder alles, ivas aus Arbeit entsteht und sich in Arbeit umwandeln
Idsst, definiren — energy is work and all else that can be produced from or converted
VOL. I. 2 Y
690 INORGANIC AND THEORETICAL CHEMISTRY
into work. Not all energy is capable of doing work. There are two kinds of avail-
ability of energy for work ; energy appears as if it were on two planes, a higher
and a lower. The work value is a measure of the availability of energy on the higher
plane ; the work value of energy on the lower plane is nil. Consequently, the defini-
tion of energy as capacity for work, is valid only when it refers to a particular
form of energy which is in a condition to do work ; it is not a definition of energy.
Two factors are involved during the expenditure of energy in doing work :
(1) The magnitude of the resistance ; and (2) the extent to which the resistance is
overcome. Thus, when a particle moves a certain distance s by the application of
a force F the amount of energy expended is measured by the work done, and is equal
to the product Fs ; if a gas suffers a change in volume dv^ when subjected to a steady
pressure y, the work dW done during the change in volume, or the energy expended
during the operation, is equivalent to the product 'p.dv. This latter problem is so
important to the chemist in studying the energy changes which occur during chemical
reactions in which the volume of the end-prodiicts is different from that of the
initial-products, that it must be considered in more detail.
The work done by a gas when it changes its volume without changing its temperature.—
Imagine a gas occupying a volume v^, confined in a cylinder, Fig. 1, fitted with a piston
to move up and down without friction ; and let the constant pressure p press the piston
downwards. Let the gas expand from a volume v^ to a volume v^ when the work performed
will be equivalent to ^{v^ — Vg). This expression is a convenient approxi-
« mation, for it can be rigorously true only when the change in volume is very
uA small because the pressure of the atmosphere on the expanding gas does not
II remain quite constant, but changes slightly as the volume increases. Hence,
the very small amount of work dW done when the gas suffers an indefinitely
small change of volume dv, iBdW=^p.dv. With an ideal gas, p=RT/v ; and
t when the expression dW—RT{dv/v), obtained by substituting this value of p
'> in pdv, is integrated for a change in volume from Vi to v^, the work of expan-
[ sion W (temperature constant) is W = RT log {v^jv^), or W = RT log {pjp^)-
„ , The second expression represents the work done during the expansion of
a gas from a pressure p^ to a pressure p^ ', and it is obviously derived by
substituting for v^^ and v^ from the relation PiV^=p2V^; or vjvy=pjp^.
Remember also that natural logarithms are supposed to be used ; if ordinary logarithms are
employed, the terms on the right must be multiplied by 2-3026. The same expression also
represents the work required to concentrate a solution with a vapour pressure p^ to one with
a vapour pressure p^*
This shows that the maximum work performed by an ideal gas in increasing its
volume from v^ to v^ is either
W^RT log^; or, W^RT log^^
Otherwise expressed, the maximum work performed by an ideal gas under the given
conditions is (i) dependent only on the initial and final volumes or pressures, but is
independent of their absolute magnitudes ; (ii) proportional to the absolute tempera-
ture ; and (iii) the same magnitude for all gases which obey the ideal gas laws.
The values for the constant R are indicated in Table I. — e.g. if the pressures are
expressed in atmospheres, and volumes in litres, R is 0'0827, and the results are in
litre-atmospheres.
Examples.- — ( 1 ) A unit mass of gas, at 20°, is allowed to expand at a constant temperature
from a pressure of 10 atm. until its pressure is one atm. The work of expansion is 2'3026
X2x (273 + 20) xlogiolOgram-cals., wheni? = 2.
(2) One gram of liquid water (volume 1043 c.c.) at 100° changes to vapour (volume
1660 c.c.) at 100° against atmospheric pressure, show that the work of expansion is equiva-
lent to 168 joules, nearly, given i? = 8*31.
Energy and matter are inseparable. — Our knowledge of the material world
can bo conveniently described in terms of two entities or abstractions : 1 . Energy ;
2. Matter. It is sometimes advantageous to keep these two concepts distinct ;
although energy and matter are separable only in thought, in reality they are indis-
solubly joined together. Energy is not matter, nor matter energy .^ There can be
THERMODYNAMICS AND THERMOCHEMISTRY 691
no matter without energy, nor energy without matter. To summarize, matter
is a term grouping together entities which possess certain properties in common ;
energy likewise is a term grouping together certain phenomena which, like matter,
have many forms ; and sometimes a third term, aether, is used for grouping
together certain relations between matter and energy.
At first sight, common sense and science seem to support the supposition that
there is a real universe existing in all its completeness quite independent of all
relation to the intelligence ; and that observations are made on real things which
are apprehended or perceived as existing fully formed and complete in themselves.
According to the energetic hypothesis of matter, the objective reality of matter is a
derived idea, for the existence of matter as something external to ourselves is assumed
in order to explain certain subjective sensations ; consequently, our knowledge
of the material world is founded upon our perceptions, which are in turn basedf
upon the evidence of our senses. We do not see material objects directly, but
rather experience a sensation presumably due to the formation of a picture of the
object upon the retina, and which is possibly a chemical effect induced by energy
radiated from the object on to the retina. Again, the mechanical energy of vibrating
air may produce sensations in the auditory organs ; and the mechanical energy of
pressure or tension may produce sensations in the organs of touch. In this way,
it can be shown that all our perceptions of the material world are derived from
sensations produced by manifestations of various forms of energy. Take away
the manifestations of energy, and nothing remains, since a body without a quality
is indistinguishable from nothing. In this sense, the objective reality of matter can
be regarded as an hypothesis, postulated to explain our subjective sensations.
W. Ostwald accordingly emphasized the old idea that matter is a redundant hypothesis,
a creature of the imagination designed as a carrier or vehicle of energy ; and that the
only things we really know are manifestations of energy. Energy and only energy
is the thing in itself ; energy is the real substratum of the physical and chemical
world. In his Studien zur Energetik, W. Ostwald (1892) ^ thus describes his conver-
sion to the energy hypothesis :
The more intimately acquainted I became with the properties of energy, the clearer
became the proof that matter is nothing but a complex of different factors of energy which
possess the property of being reciprocally proportioned. The traditional fundamental
properties of matter show themselves as modes of expression or factors of energy.
G. F. Fitzgerald (1896), M. Planck (1896), and L. Boltzmann (1896) 7 and others
have challenged the validity of the energetic view as an elemental hypothesis which
cannot be reduced to simpler terms, for kinetic energy is defined as the product of
half the mass m of a moving body into the square of its velocity F, or \mV^, but
mass is defined in terms of kinetic energy, and the definitions thus proceed
in a circle.
It is best, however, to leave the metaphysical chemist to deal with matter defined
as the unknown cause of known sensations, and answer for himself such questions
as : What is matter in and by itself ? What is the thing matter per se ? Whether
it is better to regard matter as a passive vehicle for energy, or a particular form of
energy having no existence apart from energy ? The working chemist finds it
convenient to assume that all sensible objects occupying space have a material
substratum which accompanies these objects in their motions from place to place ;
and he defines : Matter is that which possesses weight and occupies space. Matter
is thus a convenient word for grouping together those things which have the common
property of weight and form. Air, water, glass, copper, etc., are forms of matter ;
heat, light, electricity, and magnetism are forms of non-matter — energy ; colour,
odour, etc., are specific properties of particular forms of matter.
Energy, Uke matter, is indestructible. — Whenever it has been possible to
make accurate measurements, it has been found that any quantity of one form of
energy is made to disappear, an equivalent quantity of another form, or forms of
692 INORGANIC AND THEORETICAL CHEMISTRY
energy, appears. L. A. Colding, in his Thesis on Energy (Copenhagen, 1843), 8
said :
Energy is imperishable and immortal, and therefore wherever and whenever energy
seems to vanish in performing certain mechanical, or other work, it merely undergoes a
transformation, and re-appears in a new form, but the total quantity of energy still abides.
This is the quantitative aspect of the mutual transformability of the different forms
of energy. No gain or loss of energy has ever been observed in an isolated system.
This is the famous law of conservation or persistence of energy, which appears to
have been foreshadowed by R. Descartes and the Cartesian school as the law of the
indestructibility of momentum or motion; and by Gr. W. von Leibniz as the law of
conservation of vis viva, or force as it was then called. The perdurability of energy
was also foreshadowed by Isaac Newton in 1687,^ and by E. Mohr in 1837, although,
as E. Mach lo has shown, almost all eminent investigators had a more or less confused
idea of it ; and, since the time of S. Stevinus (1605), and G. Galilei in the seventeenth
century, it has served as the foundation of the most important extensions of the
physical sciences ; and adds :
This theorem is usually considered to be the flower of the mechanical world,- — the highest
and most general theorem of natural science, to which the thought of many centuries has
led.
Action and reaction, said Isaac Newton, are equal and opposite ; and further.
If the activity of an agent be measured by the product of the force into its velocity, and
if similarly the counter-activity of the resistance be measured by the velocities of its
several parts, whether these arise from friction, adhesion, weight, or acceleration, etc., then
activity and counter-activity in all combinations of machines will be equal and opposite.
The same principle was recognized to be of universal appUcation by J. R. Mayer,
in a memoir : Bemerhungen uber Krdfte der unhelebten Natur (1842),^^ which was
rejected as eminently heretical by some of the supremely orthodox journals at that
time. This work was followed by that of W. R. Grove, On the correlation of the
physical forces (London, 1843), and almost simultaneously by that of J. P. Joule,i2
in a paper, On the calorific effects of magneto-electricity, and on the mechanical equivalent
of heat, and by that of H. von Helmholtz, Ueher die Erhaltung der Kraft (1847). In
the writings of the brilliant N. L. S. Carnot, published after his death in 1832, there
occur these remarkable words :
Heat is simply motive power or motion which has changed its form, for it is but a move-
ment amongst the particles of a body. Whenever motive power is destroyed, an equivalent
quantity of heat is produced ; and reciprocally whenever heat is destroyed, motive power
is developed. It is therefore possible to establish the these genirale that motive power is a
quantity which is invariable in nature ; that is, to speak correctly, motive power is a quantity
which can neither be produced nor destroyed. True enough, it may change its form, or
produce sometimes one kind of motion, and sometimes another, but it is never annihilated.
Hence, in all chemical changes, two entities — matter and energy — remain quanti-
tatively the same, but qualitatively different. " The transactions of the material
universe," said J. C. Maxwell, in that inimitable work Matter and Motion (London,
1894), " appear to be conducted, as it were, on a system of credit. Each transaction
consists of a transfer of so much credit or energy from one body to another. The
act of transfer or payment we call work." H. St C. Deville, in his Legons sur la
dissociation (Paris, 1864), emphasized the same idea a little differently. He said :
All the labours and all the tendencies of modem science lead to the identification of all
the forces which come into play in physical and chemical phenomena ; all the niunerical
relations which have been obtained, establish their equivalence in the most rigorous manner.
Quantitative relation between the different forms of energy. — Energy in all
its forms can be expressed in terms of one basal unit the erg ; an erg is equivalent
to a force of one dyne acting through one centimetre ; that is, a dyne acting through
one cm. generates one erg of energy. A dyne is that force which acting for one
THERMODYNAMICS AND THERMOCHEMISTRY
693
second on one gram produces a velocity of one cm. per second. The dyne is also
equivalent to a weight of one gram divided by g, the acceleration o£ gravity in cm.
per sec. per sec. — where 5'=980*665 cm. per sec. per sec. at latitude 45° and sea-level.
A pressure of one atmosphere equals 1,013,000 dynes per sq. cm. Weights considered
as forces can be expressed and measured in terms of the dyne. The gravitational
unit of energy or work is the weight of 1 gram through 1 cm., and this is equal to
g dynes per cm. — that is, to (/ergs. For a latitude 45° at sea-level, ^f is 980*617 dynes
per cm. For a latitude A, and height h metres above sea-level, Helmert's formula
is 5r=980-617-2-593 cos 2A-0-0003086A.
The principle of the mutual convertibility of the different forms of energy assimies
that there are measurable relations between the different forms, and that the modes
of measurement are homologous. The possibility of measuring energy when con-
verted from one form into another is dependent on an equation showing in what
ratio the transformation has been accomplished. The ratio between heat and mecha-
nical energy was worked out in a fairly satisfactory way by J. P. Joule in 1846-9,
when he found that " 772 lbs. falling one foot would heat a pound of water 1° ; "
and he called this ratio the mechanical equivalent of heat. Later, more exact
determinations of this constant give a rather higher number than that found by
J. P. Joule ; 13 the best available data range from 4*181 to 4*192 X 10^ ergs per gram-
calorie at 15° ; and the best representative value is taken to be 4*182 xlO^ ergs per
gram-calorie at 15°. If a calorie be defined as the amount of heat required
to raise the temperature of one gram of water at 15° one degree, then 42*670 grams
falling one centimetre will generate one calorie. Hence a calorie is equivalent to
42,670 gram-centimetres of energy. Another ratio commonly employed is the so-
called joule, such that one joule is equivalent to 10,198 gram-centimetres of mechani-
cal energy. Hence,
1 calorie=4-182 joules ; 1 jouIe=0-2423 calorie
Table I summarizes the quantitative relationship between the different forms of
energy, and is convenient for reference.
Table I. — ^Numerical Equivalents of Some Forms
OF Energy.
One
Gram-
calorie.
Gram-
ceutimetre.
Watt-hour.
Erg.
Litre-
atmosphere.
Joule.
Gram -calorie .
Gram-cm.
Watt-hour
Erg . .
Litre-atm.
Joule
Gas constant, R
1
23-41 X 10-«
860-3
24-23 X 10-8
24-54
0-2423
1-9885
42670
1
3-670x10'
0-00101980
10-332x10*
10198
847
11-62x10*
27-24 XlO»
1
27-78x10-12
28-15x10-3
27-78x10-6
41-86 XlO«
80-6
36x10"
1
10-13x108
10'
8-31x10'
41-33x10-3
96-77 X 10-8
35-53
98-70-"
1
0-00987
0-08207
4-186
98-06 X 10-«
3600
10-'
101-3
1
8-316
The index notation is used for representing small or large magnitudes as powers of 10.
Thus 101 = 10, and 10-i=TVh» or 0-1 ; 10^ = 100, and 10-2=^th, or 0-01 ; 103 = 1000,
and 10 — 3— ^^i^th, or 0-001. The positive index, therefore, represents the same number of
cyphers as the index number, and the negative index one cypher less after the decimal
point than the index number.
Perpetual motion. — The law of persistence of energy is sometimes called the
first law o! thermodynamics, or the first law of energetics, and it can be expressed
another way. No machine can generate energy or do work of itself without con-
suming at least an equal quantity of pre-existing energy. Energy cannot be produced
from nothing, something must be consumed. A machine can do no work without
the aid of an external driving force — energy. It is impossible to construct a machine
which will do work without parting with energy ; when all the energy is consumed,
the machine can do no more work until more energy is supplied from without. This
694 INOKGANIC AND THEORETICAL CHEMISTRY
revised statement of the law of persistence of energy is called the law of excluded
perpetual motion. This law does not mean that perpetual motion is theoretically
impossible, but it does mean that work cannot be done without a supply of compen-
sating energy, for no work can be performed without a loss of motion or the expen-
diture of energy. Isaac Newton's first law really postulates perpetual motion as the
normal state of a body moving without constraint in a frictionless medium. Per-
petual motion implies a sustaining and propelling source of energy in order to
compensate the losses necessarily entailed in overcoming friction, etc. No system
can furnish an inexhaustible supply of energy. Neither gravitation nor magnetism
can supply energy which will make good its own loss.
Even as early as 1269, P. Peregrinus alluded to the quest for the perpetuum
mobile as a pursuit where many had " wandered about wearied with manifold toil."
In all times, recent and modern, sanguine seekers after perpetual motion have
attempted to circumvent the law of the conservation of energy. C. E. Benham
has said that " the playful way in which nature presents us with phenomena some-
times seems as if they has been cunningly devised to lure and entrap the human
mind into a belief in the possibility of achieving this unattainable result." The
impossibility of perpetual motion seems to contradict the one phenomenon which is
universal and constant and which is more striking than any other. From the
incessant movements of celestial spheres down to the congeries of rapidly vibrating
atoms and electrons in every created thing, everything appears to be in perpetual
motion. It is therefore inferred that the apparent perpetuity of the movement is
illusory ; that these movements represent but an intermediate stage in a vast uni-
verse which is slowly sinking into a state of final quiescence, when all motion will
cease to be. It must be candidly confessed that we can offer no real proof of the
truth of this law, other than the uncontradicted experience of mankind with finite
systems which admit of observation.^* It is obviously not sound reasoning to infer
that because a phenomenon always has been, it will therefore always be. Knowledge
which has appeared to be certain for hundreds of years may suddenly prove to be
gross ignorance. We assume that if per])etual motion has been possible it would
have been discovered long ago. Of course a similar argument might have been used
in 1890 against the existence of a gas like argon in the atmosphere, and the " uncon-
tradicted experience " would have been contradicted four years later. Conse-
quently, evidence of this kind can never attain certainty, and we can only say that
the wider the uncontradicted experience, the stronger is its testimony that the
empirical law is valid, and the less likely is the necessity to arise for a thorough
revision of the fundamental statement. The search for a perpetual motion through
centuries of laborious work has been fruitless. It has brought nothing but failure.
So great is our faith in the truth of this unproved law that a demonstration showing
that any supposed process would involve a perpetual motion or the creation or
destruction of energy, is considered sufficient proof that the supposed process is
impossible. We assume with M. Faraday (1857) : " No hypothesis should be ad-
mitted nor any assertion of fact credited that denies this principle. No view should
be incompatible or inconsistent with it." Most scientific societies would refuse to
consider seriously papers which violated the assumed law of excluded perpetual
motion.
Algebraic statement of the law of conservation of energy or the first law of
energetics. — According to the principle of the conservation of energy, if an amount of
work W be performed by a body against external forces (say, atmospheric pressure),
when q units of heat are absorbed by the system, the change in the internal energy
dU which the system suffers in consequence of the isothermal change will be :
Decrease of Heat Work
internal energy. absorbed. performed.
dU = (+q) -W
in words, the changes in the internal energy T/, which a system suffers in consequence
THERMODYNAMICS AND THERMOCHEMISTRY 695
of an isothermal change, is equivalent to the amount of heat absorbed less the
external work W done by the system.
In 1862, R. Clausius ^^ drew a clear distinction between the external and internal
work which a body can do when it changes its state. The term external work, W,
refers to work due to the action of external forces on the system, e.g. (i) Expansion
against an external pressure (usually atmospheric) ; (ii) Resistance to changing its
form (i.e. distortion) ; (iii) Changes in surface area against capillary forces ; and
(iv) Electric or magnetic forces when a body is moved from a high to a lower potential.
The term internal energy or work, U, includes : (i) The increased kinetic energy
of the molecules which causes a rise of temperature ; (ii) Intermolecular work done
by or against molecular forces when the volume, cohesion, or elasticity is changed ;
(iii) Intramolecular vibrations, i.e. atomic vibrations within the molecule ; and
(iv) Chemical work as when a body changes its state, etc.
From the energetic point of view, the heat q is conventionally positive if the
system absorbs heat ; and negative if the system evolves heat ; if the external work
W be done by the system, W will be negative, and positive if done on the system ; the
internal energy, U, is negative when U diminishes, and positive when IJ increases.
If the system does no external work, as occurs when a chemical reaction takes place
in a calorimetric bomb, W=0, and U=q, meaning that in an isothermal change
the system will gain an amount of internal energy equivalent to the amount of heat
absorbed. In most calorimetric work, the pressure, not the volume, is constant,
and if Vi and Vg respectively denote the initial and final volumes under a constant
pressure p, and Ui and U2 the initial and final energies of the system, the preceding
equation can be written :
U2~Ui=q—p{v2—Vi) ; or, ^=(C/2-fjP^2)— (f^i+F^i)
Hence, the heat absorbed in the reaction depends on the initial and final states of
the system. In the more general case, if Vi, Vi . . . represent the volumes of the
substances consumed in a reaction under the respective pressures ^j, Pi - ' •,
^2, ^'2' • • v and^2' jPe' • • -j the corresponding values for the substances produced
on the reaction, then, if Ui represents the internal energy of the substances destroyed
in the reaction, and U^ that of the substances produced,
U2--Ui=q-{(poV2+P2V2 + • • ■)-{Pi'^i+Pi\'+)\
which may be more conveniently symbolized, U2~Ui=q—(Zp2V2—^Pi'Viy.
Where q denotes the thermal value of a reaction at ordinary atmospheric pressures,
the symbol Q may be used in place of q, and
0=( £72+2:^)2^2) -(Ui+^PiVi) ; or Q=U-\-i:pv
The magnitude U-\-Epv is called the heat content of a given system.
Examples.- — (1) If one gram of liquid water, whose volume is 1'043 c.c. at 100°, be con-
verted into vapour occupying 1661 c.c. at 100°, against atmospheric pressure in consequence
of the absorption of 537 cals. (heat of vaporization), show that the increased internal energy
in consequence of the vaporization will be 2087 joules. The external work of expansion
is 168 joules; and 537x4-2 = 2255 joules is equivalent to 537 cals. Hence, C7 = 168
less 2255= —2087 joules. If the vaporization could be made to take place without an
increase in volume, no external work would be done, and the increase in internal energy
would be equivalent to 2255 joules.
(2) If a zinc rod be immersed in a solution of copper sulphate a reaction symbolized
Zn4-CuS04=ZnS04 + Cu occurs, and the transformation of chemical energy generates
a calories of thermal energy ; no external work is done, so that W — U. If the system be so
arranged that the chemical energy is transformed into electrical energy- — say, by placing
a zinc rod in a solution of zinc sulphate contained in a porous pot immersed in a solution in
which a copper rod is partly immersed, and connecting the exposed parts of the two rods
by wire — h calories of thermal energy are at the same time evolved. Then U, the electrical
energy produced, will be equivalent to 6 —a cals.
Relation between the laws of conservation of energy and of matter.— In
1902, Lord Rayleighi^ showed that if a real change in weight could be demonstrated
696 INORGANIC AND THEORETICAL CHEMISTRY
during a chemical reaction, perpetual motion would be possible. In the reversible
chemical action A^B, let the system in the state A be initially at a low level, and then
raised (in vacuo) to a higher level when the system is transformed into the state B.
Then let the system be returned to the lower level and transformed into the state A.
The temperature is maintained constant during the whole of the operations. The
reversed chemical reactions compensate one another and there is no gain or loss of
energy ; the operations of raising and lowering the two systems do not compensate
one another unless the weights of the system in the two states A and B are the same.
If the weights be difierent, the cycle of operations can be so executed that work is
gained, and perpetual motion becomes possible. Hence, unless something has been
overlooked, there cannot be a difference in the weights of a system in the two states
A and B. In spite of this demonstration, attempts are not infrequently made to
show that chemical changes involve a real loss of weight too small to be detected by
the balance.
Kinetic and potential energy. — There is an important difference between a stone
lying on the ground and a similar stone lying on the table. Both appear alike to
be motionless, yet the latter possesses more available energy than the former.
For example, the stone on the table, in descending to the ground, could be made
to transfer its energy to the mechanism of a clock, and do work. The available
energy would thus be transformed into mechanical motion. For the same reason,
a wound watch spring possesses more available energy than a similar spring not
wound up. Thus available energy may be conventionally regarded as active
{i.e. kinetic) or passive {i.e. latent or potential). When a marble is rolling along the
ground, it has the power, in virtue of that motion, to change the state of another
marble with which it might collide. A body, therefore, might possess energy in
virtue of its motion. This energy is said to be in a kinetic or active condition.
In 1686, G. W. von Leibniz, in a memorable controversy with R. Descartes, estab-
lished the proposition that the vis viva — the living energy — or, as it is now called,
the available kinetic energy Kf of a body of mass m moving with a velocity F, is
K=imV^. This energy may be transformed into heat when the motion of the body
is arrested. In contrast with this, it is sometimes convenient to use the term
potential energy, suggested by W. J. M. Rankine in 1853, for the available energy
which is potential to, possible to, or latent in a body in virtue of its condition
with respect to surrounding objects ; in G. W. von Leibniz's terminology this would
be the vis mortua of the system. The distinction between potential and kinetic
energy was recognized by Aristotle (c. 320 B.C.), and he called the former iv 8wa/x€t
(dynamic) and the latter iv ivepyeta (energetic) ; and in 1803, N. L. S. Carnot
called potential energy the force vive latente.
When a stone is Ufted above the ground, the energy expended and the work
done depend upon the weight w of the stone, and the height h to which the stone
has been lifted. Consequently, the available potential energy E of the stone will
be E=wh. The meaning is that a measurable quantity of energy is stored up or
rendered passive in some way, and that this same amount of energy can be recovered.
For instance, when the stone returns to the ground, it will in falling acquire an
equivalent amount of kinetic energy. Kinetic and potential energy are here
referred to the earth as constant, for obviously the suspended stone would have
no available potential energy if it could never fall. Again, water in an elevated
position can do work, in virtue of the law that all liquids will flow to the lowest
level that circumstances will permit. Consequently, water at the top of a hill
possesses potential energy. A bent spring, a raised hammer, compressed air, and
a piece of iron in the vicinity of a magnet, all possess potential energy. Substances
which in virtue of their relative condition, or the motions of their constituent
molecules, are capable of entering into chemical actions, are also said to possess
potential energy. Such are gunpowder, a mixture of metallic zinc and sulphuric
acid, etc. The light, heat, sound, and mechanical motion which attend the explosion
of gun-cotton are equivalent to the chemical energy stored in the explosive.
THEKMODYNAMICS AND THERMOCHEMISTRY 897
For convenience, the difEerent forms of energy are usually classified as mechanical
and kinetic energy ; potential energy involving stress or strain ; thermal energy ;
actinic and radiant energy ; electric energy ; magnetic energy ; and chemical
energy. Each of these, in turn, can be regarded either as energy of tension, or
energy of motion.
References.
* R. Hooke, Micrographia, London, 1667 ; I. Newton, Opticks, London, 1704 ; R. Descartes,
Principia philosophice, Amstelodami, 1677 ; R. Boyle, Opera varia de ahsoluta quieta in corporibus,
Oxford, 1680 ; J. Locke, Essay on Human Understanding, London, 1675.
2 Count Rumford (B. Thomson), Nicholson's Journ., 2. 106, 160, 1798 ; Phil. Trans., 88. 80,
1798 ; H. Davy, Elements of Chemical Philosophy, London, 1812.
^ T. Young, A Course of Lectures on Natural Philosophy, London, 59, 1807.
* W. Ostwald, Zeit. phys. Chem., 9. 563, 1892 ; 10. 363, 1892 ; 0. J. Lodge, Phil. Mag.,
(5), 8. 510, 1879.
« L. Buchner, Kraft und Stoff, Leipzig, 1867.
« W. Ostwald, Zeit. phys. Chem., 9. 563, 1892 ; 10. 363, 1892 ; Science Progress, (1), 4. 419,
1896.
' G. F. Fitzgerald, Nature, 53. 441, 1896 ; L, Boltzmann, Verh. Ges. Nalurf. Aezrte, 30, 1895 ;
Monist, 12. 65, 1902 ; Wied. Ann., 58. 595, 1896 ; 60. 231, 1897 ; 61. 790, 1897 ; M. Planck,
ih., 57. 45, 1896 ; P. Volkmann, ih., 61. 196, 1897.
8 L. A. Colding, Phil. Mag., (4), 42. 1, 1870.
^ I. Newton, Philosophies Naturalis, Cantabrigise, 1. 12, 1713 ; F. Mohr, Liebig's Ann., 24.
41, 1837 ; Phil. Mag., (5), 2. 110, 1876 ; M. Seguin, De Vinfluence des chemin defer, Paris, 1839.
^° E. Mach, Die Geschichte und die Wurzel des Satzes von der Erhaltung der Arbeit, Prag, 1872 ;
Chicago, 1911,.
11 J. R. Mayer, Liebig's Ann., 62. 233, 1842 ; Phil. Mag., (3), 24. 371, 1842.
12 J. p. Joule, Phil. Mag., (3), 23. 263, 1843 ; Scientific Papers, London, 1884 ; H. von Helm-
holtz, Ueber die Erhaltung der Kraft, Berhn, 1847 ; Scientific Memoirs, 1. 118, 1853 ; Ostwald' s
Klassiker, I, 1902 ; N. L. S. Carnot, ib.,3'7, 1892 ; Reflexions sur la puissance motricedu feu, Paris,
1824 ; New York, 1897 ; Harper's Scientific Memoirs, 6, 1899 ; P. G. Tait, Sketch of Thermo-
dynamics, London, 1877.
13 E H. Griffiths, The Thermal Measurement of Energy, Cambridge, 1901.
1* F. Ichat, Das perpetuum mobile, Leipzig, 1914 ; A. Daul, Das perpetuum mobile, Wien,
1900 ; P. Peregrinus, De Magnete, sen Rote perpetui motus, Augsburg, 1558 ; C. E. Benham,
Scient. Amer. SuppL, 82. 130, 1916.
15 R. Clausius, Pogg. Ann., 116. 73, 1862 ; Phil. Mag., (4), 26. 81, 201, 1862.
i« Lord Rayleigh, Nature, 66. 58, 1902.
§ 2. Thermochemistry
Heat and cold are nature's two hands by which she chiefly worketh.- — Francis Bacon
(1627).
Sine igni nihil operamur.' — C. Glaser (1663).
It has been shown that matter, as we know it, can be resolved into two ab-
stractions— matter and energy. Neither exists alone. We have no acquaintance
with the one apart from the other. Isolated, matter and energy are pure abstractions.
Each one completes and presupposes the other. The element phosphorus, for
instance, can be regarded as a form of matter which is always associated with a
certain amount of free or available energy, because it is able to do chemical work,
and we cannot conceive of energy coming from nothing. We cannot answer :
How much energy is associated with the phosphorus ? The actual amount avail-
able possibly depends upon the nature of the substance with which it is brought in
contact. Similarly with oxygen. When these two elements — oxygen and phos-
phorus— are brought in contact, under the right conditions for the degradation of
energy, chemical action sets in, and the chemical energy is degraded or transformed
into heat or light. The resulting compound — ^phosphorus pentoxide — still contains
some chemical energy, for if it be mixed with water, a great amount of heat is
developed, chemical energy is degraded, and phosphoric acid results. The phos-
phoric acid still contains chemical energy because more energy is degraded in the
698 INORGANIC AND THEORETICAL CHEMISTRY
form of heat when the phosphoric acid is brought into contact with sodium hydroxide.
Every chemical reaction involves a change both in the form o£ the matter and in
the form of the energy of the system.
Modern chemistry would make C. Glaser's old motto read : Without available
energy nothing can change. What is generally understood by descriptive material
or chemistry deab with matter, not with energy. Chemistry proper is essentially
concerned with both energy and matter ; and hence, it can no longer leave the
development of the energy concept, as was formerly done, exclusively to physics
and mechanics. The fundamental part played by energy in determining the nature
and character of chemical processes was for a long time obscured by the erroneous
notion that all forms of energy are but forms of mechanical energy. This
idea was perhaps the illicit consequence of man's early familiarity with
mechanical energy. First impressions are obstinately persistent. It was not
easy for man to realize that mechanical energy is but one phase of a much wider
and more comprehensive concept which includes all the other different forms of
energy. It has been as difficult for the chemists to recognize that energy is an
entity with which he has to deal, as it was for the alchemists to realize that gases
and vapours are of material importance in the study of chemical changes.
The law of Lavoisier and Laplace. — The free or available chemical energy
of different substances is usually degraded in the form of heat during chemical
action. The system gets hotter because heat is evolved by the reacting substances
— such reactions are said to be exothermic or thermopositive reactions in contrast
with endothermtc or thermonegative reactions which consume heat and thus cause the
system to become cooler. That branch of chemistry which deals with the relation
between thermal and chemical energy is called thermochemistry. The heat evolved
daring a chemical reaction is proportional to the quantity of the reacting substance,
and as a convenient standard, the amount of heat given out at constant volume
and temperature per gram-molecule of the reacting substance is called the heat of
the reaction ; this heat Q is wrongly supposed to be a measure of the decrease in
the internal energy of the substance. Experiment shows that a definite chemical
process generates a definite amount of heat. Every compound has a definite
heat of formation, which is numerically equal to the heat required for the de-
composition of the compound back into its elements, but of opposite sign. Action
and reaction, says Newton's third law, are equal and opposite. This is obviously
a corollary of the law of the conservation of energy. If it were not so, heat would
be gained or lost when a compound is formed and then decomposed back into its
original constituents. Such a result is at variance with the principle of the per-
sistence or conservation of energy. The fact that every compound has a definite
heat of formation which is numerically equivalent to its heat of decomposition
but of opposite sign, is sometimes called the law of LavOisier and Laplace, because
A. L. Lavoisier and P. S. de Laplace ^ first pointed out this generalization, in their
Memoires sur la chaleur, published between 1780-4. .Consequently, under a
given set of conditions, it seems as if each atom and each molecule is charged with
an amount of energy which is as definite as the quantity of matter itself. The
important conclusion of A. L. Lavoisier and P. S. de Laplace was later verified by
the work of H. Hess (1836-40), T. Andrews (1844), P. A. Favre and J. T. Silbermann
(1844-6), T. Woods (1851), and others,^ and data were accumulated showing
the thermal values of various chemical reactions. A large mass of thermochemical
data for various chemical or physicochemical processes has been accumulated by
J. Thomsen, M. Berthelot, and others, and most has been compiled in H. Landolt
and R. Bernstein's Physikalisch-chemische Tabellen (Berlin, 1912). ^
The symbols used in thermochemistry.- — It will be remembered that in physics, the
unit of heat is the calorie, and a calorie represents the amount of heat required to raise
the temperature of one gram of water, at some convenient standard temperature, through
1° C. Consequently, 100 cals. will raise the temperature of 100 grams of water 1°, or of
1 gram of water 100°. Sometimes it is convenient to take a pound of water as the unit,
THERMODYNAMICS AND THERMOCHEMISTRY 699
and a pound-caloiie is then the amount of heat required to raise the temperature of a.
pound of water 1° C. ; if the degree Fahrenheit be the unit of temperature, the amount of
heat required to raise the temperature of one pound of water 1° F.- — say from 60° F.
to 61° F.— is called a British thermal unit, B.T.U. ; the latter unit is used by many engineers.
To convert a big calorie into a B.T.U. multiply by 3*9681 and for the converse operation
multiply by 0-2522 ; to convert a pound-calorie into a B.T.U. multiply by 1*8 ; and for the
converse operation, multiply by 0-555.
In chemistry it is convenient to represent the thermal value or heat of a reaction by
reference to the formula weight or the gram-molecule of the substance concerned in the
reaction. Thus, the heat of formation of phosphorus pentoxide, taken at P2O5, is 370,000
cals. This means that 370,000 cals. are generated when 142 grams of phosphorus pentoxide
are formed by burning 62 grams of phosphorus in oxygen ; or 62 grams of phosphorus
burning in oxygen will give sufficient heat to raise the temperature of 370,000 grams, or
370 kilograms of water, 1°. To avoid dealing with large numbers it will be more convenient
to consider a Calorie as the amount of heat required to raise the temperature of 1000 grams
of water 1° C. This is the so-called kilogram-calorie, or the big calorie, and calorie is
then written with a capital C. Thus cal. refers to the gram-calorie, Cal. refers to the
kilogram-calorie. Hence the energy degraded in the form of heat when phosphorus bums
in oxygen is equivalent to 370 Cals. This is represented in symbols, 2P + 50=P205
+ 370 Cals. Some represent gases by means of italics, solid by clarendon type, and liquids
by ordinary type ; and some put a bar over the symbol for gases, and under the symbol
for solids ; some also abbreviate the first equation (2P, 50) =370 Cals.
Within certain limitations to be discussed later, the heat produced in a chemical
reaction has been attributed solely to the degradation of chemical energy, but in
many cases there are disturbing factors from differences in the physical properties
of the initial and end products of the reaction, etc. For example —
(1) Differences in the states of aggregation or volume of the initial and final products
of the reaction. Thus, a compound formed in a gaseous reaction may become
liquid or solid. If the states of aggregation of the reacting constituents are not
self-evident, they must be represented in the equation, otherwise, latent heats of
fusion or vaporization may lead to ambiguity. Thus, with vapour at 0°, 2H+0
=H20gas+57*82 Cals. means that the union of 2 grms. of hydrogen with 16
grms. of oxygen is attended by the evolution of 57 '82 big calories when the water
produced is in the form of steam ; if the steam be condensed to a liquid at 0°,
2H+0=H20iiq+68-63 Cals. The extra 10-81 Cals. represent the heat given out
when 18 grms. of steam are condensed to a liquid ; for solid ice, another 1'4 Cals.
would have to be added to allow for the heat of solidification of 18 grms. of liquid
water into ice.
It is necessary to distinguish clearly between the observed heat changes and
the real heat changes due to the degradation of chemical energy as heat. The
observed thermal value of a chemical reaction may be greater or less than that
which corresponds with the chemical energy actually degraded during a given
chemical reaction. For example, in the reaction between metallic sodium and
chlorine gas, 2Nasoiid-f-Cl2gas=2NaClsoiid+194:*6 Cals., there is a large contraction,
and work is done on the system by the pressure of the atmosphere ; on the other
hand, when hydrochloric acid acts on zinc, there is a comparatively large expansion,
and work is done by the system against atmospheric pressure. With hydrogen and
chlorine, when one volume of each elementary gas produces two volumes of hydrogen
chloride, H2+Cl2=2HCl-f 22*0 Cals., there is no disturbing factor of this kind.
Suppose that we start with a mixture of two volumes of hydrogen and one volume of
oxygen, and finish with liquid water, there is a tremendous contraction in volume.
This contraction occurs under atmospheric pressure (76 cm.). Hence, the atmo-
sphere does work on the system, and that work appears as heat which raises the
temperature of the system, and makes the observed heat of combination appear
greater than it really is. The work can easily be calculated, and it is equivalent to
0-9 Cal.
One gram-molecule of steam occupies 22-3 litres. The gases from which the steam was
formed occupied 1^ times this volume, i.e. 33-45 litres. A column of mercury 1 sq. cm.
sectional area and 76 cm. long weighs 76x13-59 = 1033 grms.— since the specific gravity
of mercury is 13-59. This pressure exerted along a path of 33*45 cm. will be 33-45 x 1033
700 INORGANIC AND THEORETICAL CHEMISTRY
gram-centimetres, or 33-45 X 1 "033 kilogram-centimetres. Since 42*65 kilogram-centimetres
are equivalent to one calorie, 33*45 X 1033-^4265 = 80 cals., or 0*8 Cal. This discussion
can be generalized. Every kilogram-molecule of any gas at 0°, occupying 22*4 cubic metres,
when evolved during a chemical reaction, will absorb 0*542 Cal., and 0-542(1 +0*00366^)
cal. when the gas is evolved at the temperature 6. The heat absorbed in this way is evolved
again when the gas is absorbed by a liquid or solid.
Correcting the observed heat of combination of hydrogen and oxygen 69 3 Cals.
for the contraction due to the condensation of steam to liquid water, we get,
per kilogram-molecule (18 kilograms) of water :
Apparent energy degraded in the reaction . . . . .69*53 Cals.
Energy due to the contraction . . . . . . .0*81 Cal.
Energy actually due to the reaction . . . . . .68*72 Cals.
Consequently, when the gases are measured at constant pressure, not quite one per
cent, of the heat of the reaction is due to work done on the gas by atmospheric
pressure. Otherwise expressed, the thermal value of the reaction at constant
volume will be 68*72, and at constant pressure 6953. The difference is not great,
and it is within the limits of experimental error when the results of different observers
are compared.
Gas engineers ^ express the calorific power of a gas in terms of the number of
pounds of water which can be raised 1° F. by the complete combustion of one
cubic foot of the gas, at n.t.p., on the assumption that the water formed during the
combustion is condensed to the liquid state at 212° F. This is the gross calorific
value. If the steam formed by the burning gas remains as a gas, the latent heat
of steam must be deducted from the gross calorific value, the result is termed the
net calorific value of the gas.
The apparent failure of chemists to handle the great mass of thermochemical
data satisfactorily is in part due to the fact that the real heats of chemical reactions
are obscured by unknown latent heats, and heat spent in doing work of different
kinds. W. Sutherland (1895) ^ claims that " the ideal condition in which thermo-
chemical data should be presented, is that in which they relate to the heats of
formation at constant volume of the gaseous products from gaseous elements."
By direct calculation he makes an estimate of the heats required to vaporize a number
of metals and non-metals, and also of their binary compounds. He corrects the
heats of formation of a number of binary compounds so as to make them represent
the heats of formation of gaseous compounds as the result of the combination of
gaseous metals and non-metals. It is then found that the atoms in combining
chemically evolve integral multiples of a quantity of heat 3' 8 ; and that each atom
in passing from the elementary to the combined state evolves a definite amount of
heat irrespective of the other atoms with which it combines. This generalization
has not yet been established directly from observed data ; if it be true, allowance
would have to be made for a third factor since the available evidence rather shows
that the amount of heat evolved during a chemical combination also depends upon
the mutual relations of the atoms in the molecules.
(2) Reactions in solution. — ^Again, if the reacting substances are in solution, a
certain amount of heat may, or may not, be dissipated in the act of solution. For
example, 13'7 Cals. are evolved when a dilute solution of sodium hydroxide is mixed
with a dilute solution of hydrochloric acid. The dilute solution is represented by
the suffix aq. Thus, NaOHaq.-f HClaq.=NaClaq.-f H20+13-7 Cals. If the sodium
chloride were prepared by passing hydrogen chloride gas into a dilute solution of
sodium hydroxide, more heat is evolved, because 174 Cals. are evolved when
36*4: grms. of hydrogen chloride are dissolved in water : NaOHaq.+HClgas=NaClaq.
-f H2O+3M Cals.
(3) Effects of allotropism and isomerism. — Again, the physical and chemical
condition of the reacting substances must be taken into consideration. At the
beginning of his thermochemical studies, M. Berthelot « was careful to emphazise
THEKMODYNAMICS AND THERMOCHEMISTKY 701
the fact that la quantite de chaleur degagee dans une reaction quelconque mesure la
somme des travaux chimiques el 'physiques accomplis dans cette reaction. The heats
of combination of hydrogen in oxygen and in ozone would not be the same because
of the reaction 203=302+68*2 Cals. Allowance would have to be made for the
extra energy associated with the ozone. The molecules of ozone are charged with
energy at a higher potential than the molecules of oxygen, and when the ozone
passes into ordinary oxygen, this energy is degraded in the form of heat. The
fact that the molecules of ozone are charged with a large amount of energy is
supposed to explain why ozone decomposes so readily into oxygen — sometimes
with explosive violence. Precipitated silver (108 grms.), dried at 120°, liberates
0'76 Cal. when dissolved in mercury, while the same amount of silver beaten into
a thin plate, and treated similarly, evolves 2"03 Cals. The extra energy stored in
the hammered metal is liberated as heat during the dissolution in the mercury.
If two similar springs, one wound, and the other unwound, be dissolved separately
in acid, it is said that a greater amount of heat is developed during the dissolution
of the wound spring, because the energy stored in the wound spring is degraded as
heat during the dissolution in acid.
(4) Preliminary dissociation of the reacting molecules. — It will be observed that
in the reaction between two gaseous elements A and B, with molecules respectively
A2 and B2, the heat evolved or absorbed in breaking the molecules down into atoms
is ignored. The assumption is virtually made that the heats of formation of the
molecules of the elements from their atoms is zero. This cannot be justified even
if we are ignorant of these constants. The observed heat Q of the formation of the
compound AB from the molecules of its component elements is a resultant efiect.
Let qj, denote the heat of formation of the molecules A2, and qi, the corresponding
value for the molecules B2. It then follows (from Hess' law, vide infra) that the
heat of formation, q, of the compound AB from its elements is really Q-\-\(qa-{-qh)='9.'
An estimate of the thermal values of chemical reactions can be made when the
equilibrium constant K is known. In this way, E. Briner (1914) ^ computed the
heats of formation of a gram-molecule of the following elements :
Temperature .
I2
1390°
1050°
1670°
2177°
2427°
3505°
Equilibrium constant
0-66
0-06
0-01
0-50
0-10
Heat of formation .
32-4
57-0
113-0
120-0
1300
150-0 Cals.
HCl
HBr
HI
22-0
12-4
1-45 Cals.
143-5
105-9
84-8 „
Hence, in the reaction H2+Cl2=2HCl+4:4'0 Cals., the dissociation of the hydrogen
molecules absorbs 130'0 Cals. and the chlorine molecules 113*0 Cals., or jointly,
243"0 Cals., so that the observed thermal value of the reaction 44*0 Cals. must be
increased to 287'0 Cals., if the hydrogen chloride molecules are formed from atoms.
Hence, the real thermal value of the reaction H+C1=HC1 is not 22*0 but 143'5 Cals.
Similarly,
Heats of formation from molecules, Q
Heats of formation from atoms, q .
So far as the evidence goes, E. Briner (1914) concludes : The heats of formation
of all the compounds from the atoms of their elements are exothermal. The
observed heat of formation, Q, of a compound from its elements is therefore the
difference between two magnitudes such that Q—q—\(qa-\-qb), and a compound
will appear to be exo- or endo-thermal according as the heat oiE its formation from
atoms is greater or less than the mean of the heats of formation of the molecules
of its component elements from their atoms. The heat of formation of the nitrogen
molecule from its atoms is comparatively large, and accordingly, this element
forms many endothermal (NO, N2O, NCI3, etc.) or feebly exothermal (NH3, etc.)
compounds. K. Fehrle (1918) has attempted to calculate the heat of a reaction on
the assumption that the atoms of spherical molecules rotate about a common
centre. 8
702 INORGANIC AND THEORETICAL CHEMISTRY
(5) Differences in the specific heats of the initial and final products of the reaction. —
Heat may also appear to be generated during a chemical reaction which is partly-
due to differences in the specific heats of the initial and final products of the re-
action. If the latter be less than the former, some of the heat generated will be
due to the fact that latent heat originally present can no longer be accommodated,
so to speak, owing to the diminished capacity of the system for heat, and there is
an output of heat during the reaction in excess of that corresponding with the
degradation of energy. In illustration, J. Thomson (1882) has shown that when
JiV-sodium hydroxide is neutralized by an equivalent amount of hydrochloric
acid, the thermal capacity of the system increases about 1'37 per cent. Hence,
measurements of the thermal changes which occur during a chemical reaction
should be supplemented by measurements of the heat capacities of the substances
concerned in the reaction.
The temperature coefficient of a reaction. — Consider a reaction in which a
substance A changes into B such that A->B. Let the reaction proceed at the
temperature Ti when Qi units of heat are absorbed ; then heat the product B to
the temperature T2. If Cp and Cp^ respectively denote the thermal capacities or
molecular heats of the initial and final product^ of the reaction, the total energy
absorbed in changing A at T^ to B at T2 is equal to Qi-\~Cp^{T2~Ti). Again, the
same final state can be obtained by heating A to T2, and allowing the reaction
to occur at that temperature. Let Q2 denote the heat of the reaction at T2,
then it follows from the law of conservation of energy : 02~f^p(^2~^i)=Qi
+Cp'(T2-Ti) ; or {Q2-Qi)l(T2-Ti)=Cp'-Cp. If the difference Tg-^i be taken
indefinitely small, say dT, the difference Q2—Q1 will also be indefinitely small, say
dQ, and at the limit, we thus obtain
where dQjdT is the so-called temperature coefficient of the reaction. This equation
is sometimes called— after G. Kirchhoff (1858) 9— Kirchhoff's equation. The
equation can be taken to mean that at any assigned temperature, the change in the
quantity of heat concerned in a reacting system, kept at constant volume, per degree
rise of temperature is equal to the difference in the thermal capacities of the initial and
final states of the system, which for convenience can be written — EC p. From
Mayer's formula, Cp—Cv=R—p, it follows that —l!Cp=—I!(C^-}-R), and hence
The influence of temperature on the heat of a reaction is directly determined by
measuring the heat of the reaction at two different temperatures ; the equation
enables this magnitude to be calculated when the specific heats of the initial and
final products of the reaction are known. If the molecular heats of the initial and
final products of a reaction are the same, Cv'=C/, and the temperature coefficient
will be zero, otherwise expressed, the amount of heat evolved during the reaction
will be the same at all temperatures. If Cp^ be greater than Cp, the molecular
heat of the product of the reaction will be greater than the original initial substance,
the heat of the reaction will decrease with the rise of temperature ; and if Cp be
greater than Cp\ the molecular heat of the product of the reaction will be less than
that of the initial substance, the heat of the reaction will increase with rise of
temperature.
Examples.' — (1) According to L. Holbom and F. Henning (1907), the molecular heat
of hydrogen, Hg, and of oxygen, Oj, is 4-68 + 0-00026T, and of water, HjO, 5-61 +0-0007171'.
Hence, the difference between 6'/ and C„ for the reaction 2H2 + 02 = 2H20 at constant
volume is 3(4-68 + 0-00026T) -2(5-61 +0-000717T) = 2-82 -0-000654T, and this is the
temperature coefficient of the reaction.
(2) According to H. V. Regnault (1862), the molecular heat of hydrogen at constant
THERMODYNAMICS AND THERMOCHEMISTRY 703
volume is 4-82(^-18), of oxygen 4-96(^-18*), and of steam, HoO, 18(^ — 18); and
J. Thomsen observed that at 18°, the thermal value of the retiction H2-1-0=H20 + 67,484
cals. What is the thermal value of the reaction H24-0=H20 at 50° at constant volume ?
Here Ha + 0 = HaO is represented by (4-82 + 2-48-18) (50-18) or -10'7 X 32 = 342-4.
Hence, the thermal value of the given reaction at 50° is 67,484 less 342*4 = 67, 142 cals.,
nearly, when the volume of the system is the same at 50° as it was at 18°.
.(3) According to L. Holbom and F. Henning (1907), the molecular heats of oxygen or
carbon monoxide is 4-68+0-00026r, and of carbon dioxide 5-106 + 0-00334T-7-35
Xl0-'T2. Hence show that dQIdT for the reaction 2CO + 02 = 2C02 is 3-828-0-0059T
+ 0-000000735^2.
According to I. W. Cederberg,!^ the molecular heat of a vapour Cp and of solid
or liquid C^ , at r is (7p=2-5J?+r3125aTi ; Cp'=l-3125aTi, where a and a' are
constants characteristic of particular substances. Consequently, from G. KirchhofE's
equation, for the heat of vaporization A,
^^=2-5i2-r3125(a-a')r* ; or, X=XQ+2'iRT-0'lb(a-a')n
al
where Aq denotes the latent heat of vaporization at absolute zero. Similarly, for
the thermal value of a reaction Q^,,
^=2J2'5R-Ul-3125(a-a')Ti ; Qp=Qoi-2'6ZRT+0'75S{a~a')n
where Qq represents the heat ot the reaction at 0°.
References.
1 A. L. Lavoisier and P. S. de Laplace, Mem. Acad., 359, 1780 ; 387, 1784 ; OstwaWa Klassiker,
40, 1892.
2 H. Hess, Pogg. Ann., 50. 385, 1840 ; T. Andrews, Trans. Roy. Irish Acad., 19. 228,
393, 1842 ; Phil. Trans., 130. 22, 1844 ; 135. 91, 1848 ; Phil. Mag., (3), 32. 321, 426, 1848 ;
P. A. Favre and J. T. Silbermann, Compt. Rend., 18. 695, 1844 ; 20. 1565, 1734, 1845 ; 21. 944,
1845; 22. 483, 1140, 1143, 1846; 23. 199, 411, 1846; 24. 1081, 1847; 26. 585, 1848; 27. 56,
111, 158, 362, 1848 ; 28. 627, 1849 ; 29. 440, 1849 ; Ann. Chim. Phys., (3), 34. 357, 1852 ; (3),
36. 5, 1852 ; (3), 37. 405, 1853 ; T. Woods, Phil. Mag., (4), 2. 268, 1851 ; (4), 3. 43, 299, 1852 ;
(4), 4. 370, 1852 ; (4), 5. 10, 1853 ; Phil. Trans., 146. 1, 1856 ; Proc. Roy. Soc, 8. 211, 1857.
3 J. Thomsen, Thermochemische Untersuchungen, Leipzig, 1882-6 ; Thermochemistry, London,
1908 ; M. Berthelot, Essai de mecanique chimique fondee sur la thermochimie, Paris, 1879 ; Thermo-
chimie — donnies et lois numeriques, Paris, 1897 ; M. M. P. Muir, The Elements of Thermal
Chemistry, London, 1885.
* J. H. Coste, The Calorific Power of Gas, London, 1911.
« W. Sutherland. Phil. Mag., (5), 40. 1, 1895.
* M. Berthelot, Essai de mecanique chimique fondie sur la thermochimie, Paris, 1. 1, 1879.
' E. Briner, Journ. Chim. Phys., 12. 109, 1914.
8 K. Fehrle, Phys. Zeit., 19. 281, 1918; J. Thomsen. Thermochemische Untersuchungen,
Leipzig, 1882.
8 G. Kirchhoff, Pogg. Ann., 103. 454, 1858 ; Gesammelten Abhandlungen, Leipzig, 1882 ;
Ostwald's Klassiker, 101, 1898.
1" 1. W. Cederberg, Die thermodynamische Berechnung chemischer Affinitdten, Berlin, 24,
1916.
§ 3. The Principle of Maximum Work
In exothermic combination the sum of the specific energies of the coAiponent elements
exceeds the specific energies of the compounds formed, whUe in endothermic combination,
the specific energies of the compounds formed is greater than the aggregated specific energies
of the components. — J. B. Stallo.
The heat developed during a reaction represents a certain amount of potential
energy which was associated with the atoms in some way ; and there is a temptation
to generalize, as J. Thomsen ^ did in his paper Die GrundzUge eines thermochemischen
Systems J in 1853, and assume that the total quantity of heat developed during
704 INORGANIC AND THEORETICAL CHEMISTRY
a chemical reaction is a measure of the chemical affinity of the reacting substances
— die ganze durch eine chemische Wirkung erzeugte Wdrmemenge ist also ein Masse
fur die durch dem Process erUbundene chemische Krdfte — and that every chemical
change which can take place without the aid of external energy will be accompanied
by an evolution of heat. Uaffinite etait la cause, said H. St. C. Deville (I860),
la chaleur degagee est Veffet produit par cette force et lui est proportionelle. The same
idea was emphasized by M. Berthelot in his Recherches de thermochimie in 1869,
when he boldly formulated his celebrated principle of maximum work : Every
chemical change which takes place without the aid o! external energy, tends to the
production of that system which is accompanied by the development of the
maximum amount of heat — le principe du travail maximum : tout changement
chimique, accompli sans Vintervention d'une energie chimique etrangere, tend vers le
production du corps ou du systeme de corps qui degage le plus de chaleur. To take an
oft-cited mechanical analogy, the heat evolved by the impact of a falling body on
the ground bears a definite relation to the height from which it fell, and heights
might be measured by the heat developed by falling bodies if it were not that
more convenient methods are available. With chemical reactions, the heat evolved
is assumed by M. Berthelot to measure the mechanical work done, i.e. the loss of
chemical energy ; otherwise expressed, elements with the stronger afi&nity for one
another disengage most heat during chemical action. The production of heat
does not of course explain why the reaction takes place any more than the heat
developed when a falling body strikes the earth explains gravitation.
It follows from Berthelot's principle that reactions which proceed spontaneously,
when once they have started, liberate some form of energy, generally heat, during
the progress of the reaction. In illustration, the heat of formation of calcium
oxide is 131 Cals. ; of lead oxide, 50 Cals. ; and of mercuric oxide, 31 Cals. Calcium
oxide is not decomposed by heating it to redness in a tube, either alone or in a current
of hydrogen ; lead oxide is not decomposed by heating it alone, but it is decomposed
by heating it in a current of hydrogen ; and mercuric oxide is reduced by either
treatment. Again, the heats of solution of the following metals in dilute hydro-
chloric acid, per equivalent of metal, expressed in grams, are :
K
Na
Ca
Mg
Zn
Fe
Cu
61-8
57-2
54-3
54-1
17-4
10-7
-10 cals.
This agrees with the order of affinity for these metals deduced from other con-
siderations. The negative heat of the reaction between copper and dilute hydro-
chloric acid corresponds with the fact that the action does not occur under ordinary
circumstances, and the heat of the reaction has to be determined indirectly from
the action of the acid on the oxide of the element in question.
There are some objections to the principle of maximum work as formulated
by M. Berthelot. Lord Rayleigh, in a paper On the dissipation of energy (1875), showed
that it is not the evolution of heat but the dissipation of energy which determines
whether a chemical transformation is possible or not ; no dissipation of energy,
no transformation. Six years earlier, A. Horstmann 2 had also shown that the
evolution of heat is not the real criterion for the possibility of chemical change.
The main facts which indicate that something is wrong with Thomsen's and
Berthelot's criteria, are as follows :
(1) The principle assumes that reactions proceed completely to an end, whereas
in a balanced reaction, the reaction may be exothermal in one direction, and endo-
thermal in the other. According to the principle of maximum work, the exothermal
change ought to go completely to an end. Hence, the principle is not in agreement
with facts.
(2) A reaction may not always proceed to the stage directly which develops the
maximum amount of heat, as illustrated by successive reactions — e.g. the action
of chlorine on sodium hydroxide gives a mixture of sodium hypochlorite and chloride
which involves a smaller heat of reaction than if all the sodium was converted into
THERMODYNAMICS AND THERMOCHEMISTRY 705
chloride. This objection might be met by insisting on the importance of the word
tends in the enunciation of the principle, were it not for some evidence that the
products are sometimes comparatively stable.
(3) In a series of compounds of the same type, those with the greatest heat of
formation are not always the most stable. Thus, the heats of formation of the
carbonates of silver and lead are respectively 25'96 Cals. and 72*88 Cals. ; and of
the nitrates of silver and lead, respectively 30*06 Cals. and 54*05 Cals. Hence, it
might be concluded that the lead salts are the more stable when heated. This is
not the case. Lead nitrate begins to decompose at 203° (20 mm. pressure), while
the silver salt does not decompose at 350° (in vacuo) to any appreciable extent.
It is probable that lead nitrate is the more stable salt when iij solution.
(4) Several spontaneous chemical reactions are known to be accompanied by
an absorption of heat. The heat of the endothermal reaction between iodine and
hydrogen is nearly —6 Cals. The solution of many salts in water, the action of
lead iodide on potassium sulphate, etc., are further illustrations of endothermal
reactions which proceed contrary to the principle of Berthelot. When a solution
of ammonium nitrate is mixed with a solution of potassium carbonate, —3*1 Cals.
are absorbed in the formation of potassium nitrate and ammonium carbonate.
No measurable amount of heat is absorbed or evolved when the two latter compounds
are mixed, and hence, K2C03+2NH4N03=-(NH4)2C03+2KN03-3-I Cals. This
endothermal reaction has been explained by assuming that heat is evolved by the
reaction between the potassium carbonate and ammonium nitrate, and that the dis-
solution of the products in water accounts for the absorption of heat. It may be
true that the solution of salts in water involves (i) a physical process — the liquefaction
of the salt attended by an absorption of heat ; and (ii) a chemical process — the union
of the salt with water. The cooling effect which attends the solution of many salts
was once thought to be explained by saying that the heat absorbed in the first-
named process exceeded that in the second.
(5) Many systems require a preliminary impulse to start the reaction, and hence,
it would be necessary to introduce a clause to provide for this phenomenon.
The principle of maximum work must therefore be either amended or abandoned.
For example, it has been amended to read : Every change which takes place without
the aid of external energy must do work, and a system which cannot do work is
incapable of spontaneous change and is in stable equilibrium. Hence, the criterion
for spontaneous reactions is not the production of heat, for many spontaneous
reactions absorb heat. The reaction between hydrogen and iodine can do positive
work equivalent to -|-5'0 Cals. per gram-molecule of iodine, but the heat of the
reaction is negative, —6*0 Cals. Further investigations have shown that it is not
at all improbable that all chemical and physical reactions will be exothermal and
complete at absolute zero, —273° ; and consequently, the principle of maximum
work will probably apply at that temperature. At ordinary temperatures, the
principle is only approximately exact.
Explosive compounds. 3 — Chemical union is usually (not always) accompanied
by the evolution of heat, and chemical separation by an absorption of heat. There
are some exceptions — endothermal compounds — which are formed with an absorp-
tion of heat, and hence decompose exothermally. If the thermal value of a reaction
is a measure of the available energy which is degraded as heat during the reaction,
it follows that energy must somehow be stored up in endothermal compounds,
and that such compounds are ready to give up energy to form another state of
things with less potential energy. Just as a bent strip of flexible steel will fly back
to its original position on being released, so does the potential energy of endothermal
compounds tend to *' fly back " so to speak, other compounds with less potential
energy being formed. If a reaction takes place in a very short time it is frequently
explosive. The non-explosibility of endothermal reactions corresponds with the
fact that these reactions are self-cooled and brought to a standstill by the absorption
of heat, whereas in exothermal processes, the reaction once begun, is rapidly
VOL. I. 2 z
706 INORGANIC AND THEORETICAL CHEMISTRY
accelerated by the self-heating which results from the evolution of heat. Other
things being equal, the greater the evolution of heat, the greater the probability
of an explosive reaction. Thus, with the oxalates of the metals — RC2O4 — which
decompose : RC204=R+2C02, where R is the symbol for one of the metals, Zn,
Pb, Cu, Hg, Ag2,
ZnCaO^ PbCjO^ CuCaO^ HgC204 Ag2C204
Heat of decomposition . —49 —17 -f6 +17 +30 Cals.
Non-explosive. Doubtful. Explosive.
The heat of decomposition of a given compound will vary if the course of the
reaction varies. Hence, a compound might decompose with explosive violence
one way and non-e±plosively another — e.g. potassium chlorate. Nitroglycerol,
too, when ignited by a flame burns quietly enough, but if it be subjected to a
mechanical shock, or heated to a high enough temperature, it decomposes with
spectacular violence.
It does not follow that because the heat of formation of a compound from its
elements is positive, therefore the compound cannot decompose exothermally,
since other products of decomposition may be formed, e.g. the heat of formation
of liquid nitroglycerol — C3H5N309^ — from its elements is +415 Cals., and if it be
decomposed back into its elements, the heat of decomposition must be — 415 Cals.
As a matter of fact, when the compound decomposes into carbon dioxide, water
vapour, free nitrogen, and free oxygen, +1580 Cals. are evolved, not absorbed.
The absorption of heat during the formation of an endothermal compound, from
its elements, in general, shows that more energy is needed to tear asunder the atoms
of the reacting molecules, say A2=A+A, and B2=B+B, than is given out by the
union 2A+2B=2AB. Take acetylene or cyanogen in illustration. Endothermal
compounds are not therefore to be regarded as compounds which have been formed
in opposition to the affinities of their constituent elements, because that would imply
the existence of a negative affinity or a negative form of energy which is an idea
quite outside the range of experience. If the atoms in the molecule of an endo-
thermal compound repelled one another, it seems highly probable that the molecule
would break up unless it were continually subjected to an external stress.
Eefeeences.
1 J. Thomsen, Pogg. Ann., 88. 349, 1853 ; 90. 261, 1853 ; 91. 83, 1854 ; 92- 34, 1854 ; Ber.,
64. 23, 1873 ; H. St. C. Deville, Compt. Bend., 50. 534,584, 1860; M. Berthelot, Bull. Soc. Chim.,
(2), 19. 485, 1873; Ann. Chim. Phys., (4), 6. 292, 329, 442, 1865 ; (4), 12. 94, 1867 ; (4), 18.
103, 1869 ; (4), 29. 94, 289, 1873 ; (4), 30. 145, 433, 456, 1873 ; (5), 4. 1, 1875 ; Essai de mecanique
chimv/ue fondee sur la thermochimie, Paris, 2. 422, 1 879.
2 Lord Rayleigh, Proc. Roy. Inst.,1. 386, 1875 ; Nattire, 11. 454, 1875 ; M. Berthelot, Compt.
Rend., 118. 1378, 1894 ; P. Duhem, Thermochimie a propos d'un Hire recent de M. MarceUn
Berthelot, Paris, 1897; A. Horstmann, Liebig's Ann. Suppl.,Q. 61, 1868.
' H. Brunswig, Explosivstoffe, Leipzig, J 909.
§ 4. The Principle of Reversibility
While a transformation of energy is initiated only when equilibrium is unstable, yet
it occurs always in the direction of a recovery of stability.- — S. A. Reeve.
We do not know what is the exact relation between the thermal value Q of a
reaction and temperature although we do know that the heat of a reaction alters
with variations of temperature. It is assumed jpro tempore that the relation between
Q and the absolute temperature T can be represented by an expression of the form
Q—QQ-\-aT-]-^T^, when the numerical values of the constants a and ^ can be
determined from measurements of Q at three or more different temperatures ;
Qo should represent the value of Q at absolute zero. Eor the reaction CO2+H2
=G04-H20,it has been found that the constants assume these values : §=—10232
THEKMOBYNAMICS AND THEKMOCHEMISTRY 707
+01685T+0-0010ir2. When the heat of the reaction is zero, Q=0, and T
must be either 3100° or 2830°. Experiment shows that the second value is nearer
the truth. Hence, the reaction must be exothermal below 2830°, endothermal
above, and thermally neutral near this temperature. The reaction H2+l2=2HI
is endothermal below 320°, exothermal above, and thermally neutral near this
temperature. This shows that some endothermal compounds become exothermal
at higher temperatures — e.g. the formation of hydrogen sulphide, and probably
ozone, hydrogen peroxide, silver oxide, etc. ; and conversely, some exothermal
compounds become endothermal at higher temperatures^e.^r. the formation of
silicon hexachloride, the reaction between carbon dioxide and hydrogen, etc.
These changes correspond with a reversal of the thermal value of the reaction at
the elevated temperature. The consequence is that a compound may be unstable
at low temperatures, and stable at higher temperatures, and conversely, stable
at low temperatures, and unstable at higher temperatures. Hydrogen peroxide
and ozone are examples of the former, water an example of the latter.
This reversal of the direction of a reaction with a change of temperature shows
how necessary it is to indicate the conditions of a reaction when stating the character
of the change. Thus, carbon dioxide is usually a neutral gas, but it oxidizes zinc
vapour at elevated temperature, and steam likewise is an oxidizing agent for iron
and carbon at high temperatures.
The most stable compounds are usually but not always those with the greatest
heats of formation. In a general way, the higher the temperature, the less the
stability of exothermal compounds ; and conversely, endothermal compounds
generally become more stable as the temperature is raised, because an absorption
of heat is necessary for their formation. Here is another illustration of the principle
of reversibility previously discussed. A compound formed with the evolution of
heat is decomposed by the addition of heat ; water, for example, is an exothermal
compound, and steam is decomposed when heated to a high temperature ; the higher
the temperature the greater the amount decomposed, or dissociated into its elements :
2H20^2H2+02. 'Pot instance, W. Nernst and H. von Wartenberg (1906) found :
Temperature . . . 1000° 1500° 2000° 2500°
Amount dissociated . . 0*00003 0-0221 0-5880 3-98 per cent.
This means that if 100 grms. of steam be heated to 2500°, at atmospheric pressure,
the mixture will be in equilibrium when it contains approximately 96 grms. of
steam, 3*55 grms. of free oxygen, and 0'45 grm. of free hydrogen. If the temperature
be lowered some of the hydrogen and oxygen will recombine ; if the temperature
be raised more steam will be decomposed. When a substance decomposes with a
change in the physical conditions— temperature, pressure, etc.^and the products
of decomposition recombine when the original conditions are restored, the process
of decomposition is said to be dissociation. Conversely, a compound formed by
the absorption of heat is decomposed by the withdrawal of heat ; for instance,
ozone is an endothermal compound. The equilibrium conditions at different
temperatures in the presence of oxygen are :
Temperature . . 0° 100° 500° 1000° 2000° 3000°
Per cent, of ozone . 9-5xl0-i« 3-5x10-" 9-6 XlO"*^ 2'2xl0-2 0-9 3-6
Quite an appreciable amount of ozone will be in equilibrium with oxygen at the
higher temperatures, but at ordinary temperatures the amount is inappreciable.
W. Ostwald (1891) i has said :
It is generally believed that at a high temperature, such as that which exists in the
electric arc, and in the sun's atmosphere, all compoimds must be dissociated into their
elements. This view is certainly not justified. On the contrary, what we actually know
about the stability of compounds is that all compounds which are formed with an absorption
of heat become more stable with rising temperatures, and vice versa. Owing to the fact
that the majority of compounds known to us are formed from their elements with the
evolution of heat, and in consequence, become more unstable as the temperature rises,
708 INORGANIC AND THEORETICAL CHEMISTRY
it has been concluded that this is generally the case ; but if we remember that acetylene
and cyanogen- — two compounds formed with the absorption of heat- — are readily formed
in quantity at the high temperature of the blast furnace, and in the arc light, we see the
possibility that spectra occurring at high temperatures may belong to compounds which
exist only at elevated temperatures. •
References.
1 VV. Ostwald, Abhandlungen und Vorlrdgey Leipzig, 41, 1904.
§ 5. Hess' Law
Each element as well as each compoimd embodies a distinct and invariable amount
of energy as well as a distinct and invariable amount of matter, and the energy is as"
constitutive and essential a part of the existence of such element or compound as its
weight. — J. B. Stallo.
The calorimeter is perhaps as necessary for determining the energy communicated to
or from a system undergoing chemical change as the balance is for determining the masses
affected.— C. J. Reed (1901).
G. H. Hess (1840) i measured the heat developed during the formation of a com-
pound made in several different ways and came to the conclusion that the amount
of heat evolved during the formation of a given compound is the same whether
the compound is formed directly or in a series of intermediate stages — wenn eine
Verhindung stattfindet, so Wdrmemenge constant, es mag die Verbindung direct
oder indirect und zu widerholten Malen geschehen — this is called Hess' law. This
law tacitly assumes the law of conservation of energy ; G. H. Hess seems to have
regarded the law as axiomatic or self-evident without proof. It is a direct corollary
from the law of conservation of energy, and is interesting since it came before J. R.
Mayer or J. P. Joule. The principle may be illustrated by making calcium chloride
by the action of quicklime on dilute hydrochloric acid. It is found that :
CaO-f 2HClaq=CaCl2aq+H20+46 Cals.
Instead of this, (i) first slake the quicklime, and Ca04-H20=Ca(OH)2+15 Cals. ;
then (ii) dissolve the calcium hydroxide in water, and Ca(OH)2-|-Aq=Ca(OH)2aq
-f-3 Cals. ; finally (iii) mix the lime with dilute hydrochloric acid, and Ca(0H)2aq
-j-2HClaq=CaCl2aq+H20+28 Cals. These three steps in the formation of the
solution of calcium chloride give a total 28+3+15=46 Cals. as the heat of
formation. The same result was obtained by the direct action of the dilute acid
on quicklime.
Just as in mechanics the work done by a falling body is always the same whatever
be the path described, and whatever be the time occupied in the descent, for the
body may fall perpendicularly, down an inclined plane, down a parabolic or other
path, yet the work in every case is measured by the perpendicular height it
actually falls, so experiments have led to the inference that
(i) The heat of formation of a compound is independent of its mode of
formation. — ^This result is but a particular application of the law of persistence of
energy, and it may be expressed by saying that the change of energy of a system
in passing from one state to another depends upon the initial and final states of the
system, and not on the intermediate states. Starting with given raw materials,
suppose that it were possible to make a compound by two different processes so
that the total heat of formation of the compound formed by one of the processes
were greater than that by the other process, then it would be possible to devise a
process involving the creation or destruction of energy.
(ii) The thermal value of a reaction is independent of the time occupied by
the process.— The thermal value of a reaction is the same whether it takes place
slowly or quickly. In the former case, the heat may have time to be dissipated by
conduction or radiation, and, in consequence, appear to be less than when the
THERMODYNAMICS AND THERMOCHEMISTRY 709
reaction takes place quickly — it is here assumed, of course, that the system is not
affected by external forms of energy. In practice, the risk of error, and consequently
also the experimental errors, are great with very slow reactions.
It also follows as a corollary to Hess' law that the thermal value of a reaction
is the sum of the heats of formation of the final products of the reaction less the
heats of formation of the initial products of the reaction. Let Q denote the thermal
value of a reaction, Qi the heat of formation of the initial products, and Q^ of the
final products of the reaction, then Q='Q2—Qi- This corollary to Hess' law is
valuable because it enables the heat of formation of a compound from its elements
to be computed when a direct determination is either impracticable or very
diflSicult. This may occur when the heat evolved during the mutual action of two
solids is difficult to measure accurately. For instance, if the heat of formation of
carbon dioxide from carbon is C+02=C02+96"96 Cals., and from carbon monoxide
CO+0=C02+68-20 Cals., we have (C+O2)-(CO+O)==96-96-68-20, or 28-76
Cals., and consequently the heat of formation of carbon monoxide C+0=CO
is 28' 76 Cals. Again, it is required to compute the thermal value of
the reaction S03+BaO=BaS04, when measurements show that SO3+H2O
:=H2S04+18-7 Cals. ; BaO+H20=Ba(OH)^soi+13-9 Cals. ; and that Ba(0H)2
+H2S04=BaS04+2H20+18'4 Cals. Hence, BaO+SO3=BaSO4+5r0 Cals.
Again, to determine the heats of formation of hydrogen iodide and hydrogen bromide
when it is known that Cl+HBr=HCl+BriK^+12-5 Cals., and that C1+HI=HC1
+Isoiid+28-2 Cals. ; and that the heat of formation of hydrogen chloride is 22"0
Cals. In the first case, 22-0— 12'5==9'5 ; and in the second, 22*0— 28*2=— 6-2.
Hence, H+Br=HBr+9-5 Cals. and H+I=HI-6-2 Cals.
Similarly, the thermal value of a reaction can be calculated when the heats of
formation of the different substances which take part in the reaction are known.
Thus, by consulting some book of Laboratory Tables we can write the heats of
formation of the substances concerned in the reduction of lead oxide by carbon
monoxide :
CO+PbO=C02+Pb
29-2 50-3 97-3 Cals.
and 29-2+50-3=97*3+a;, where x denotes the thermal value of the reaction ;
consequently, a;=97-3— (29-2+50-3), or 17-8 Cals., and the reaction is accordingly
symbolized : CO+PbO=C02+Pb+17-8 Cals. The heat of formation of silver
chloride, AgCl, is 29 Cals. and of silver bromide, AgBr, 27*1 Cals. Is the reaction
AgBr+Cl=AgCl+Br likely to occur ? The heat of the last reaction is 29—27-1
=1-9 Cals., and hence the reaction is likely to take place. Further, if copper pre-
cipitates silver from a dilute solution of silver nitrate, 2AgN03aq +Cu=:Cu(N03)2aq
+2Ag+25-3 Cals., will zinc precipitate silver from dilute silver nitrate when it is
known that Cu(N03)2aq +Zn=Zn(N03)2aQ +Cu+61-7 Cals.? Probably yes, because
the heat of the reaction 61*7— 25-3=36-4 Cals. Both conclusions are in agreement
with observations. The energy of a chemical reaction is not frimarily inherent in any
one of the reacting components, hut belongs to the system as a whole ; this energy may
be represented as the sum of two or more constants which are peculiar to the
respective elements involved in the reaction.
Examples.— (1) It is required to compute the heat of formation of K+C1=KC1, when
it is known that the heat of formation of K + 0+H+Aq=K0Haq + 117 Cals • 2H + 0
=H20iiq+68-4 Cals.; H+Claq=HClaq + 39-3 Cals.; heat of solution of KCl in
water, -4-4 Cals.; and that KOHaq+HClaq = KClaq + H20 + 13-7 Cals. This last
relation can be written: (H + Cl+Aq) + (K + 0+H+Aq) — (K + Cl + Aq) — (2H + 0)
= — 13-7 Cals. Consequently, after substituting the given data, and transforming alge-
braically, we get K+Cl + Aq + 101-6 Cals. ; and hence, K + Cl + Aq=KClaq + 101'6
Cals. Subtract the heat of solution —4*4 Cals., and we get 101-6 — (— 4*4) = 106 Cals. for
the thermal value of the reaction K + C1=KC1. It will be noticed that the solution of
potassium chloride in water is an endothermal process, and hence, the heat of formation of
KClaq is less than that heat of formation of KCl.
(2) Show that when silver chloride, AgCl, is mixed with hydriodic acid, HI, silver
10
INORGANIC AND THEORETICAL CHEMISTRY
iodide, Agl, and hydrochloric acid, HGl, will probably be formed when it is known that
the heat evolved during the formation of sUver iodide is Ag-|-I=AgI + 18-6 Cals. of silver
chloride, Ag4-Cl=AgCl-f-34'8 Cals. ; hydrochloric acid (aqueous solution), H+Claq=HClaq
+ 39-3 Cals. ; and hydriodic acid in aqueous solution, H+I=HIaq + 13-2 Cals. Ansr. 10-0
Cals. will be evolved during the reaction AgCl+HI=AgI+HCl, and reactions generally
occiu" which are attended by the evolution of an appreciable quantity of heat.
The heat of formation of many substances has been determined from the heat
of combustion, i.e. the heat which is developed when the substance is completely
oxidized. The method is particularly applicable for compounds whose heats of
formation cannot be directly determined, either because the reaction is too slow
or because the compound cannot be formed directly from its elements. It is, how-
ever, necessary to know the heats of formation of the products of combustion as
well as the heat of combustion.
Example." — The heat of combustion of methane, CH4, is 213'5 Cals., and the heats of
formation of the carbon dioxide 96-96 Cals. ; and of water, 68*6 Cals. Hence, since
CHt + 40=CO., + 2H20 + 213-5, Qg — Qi=the required heat of formation of methane is
(96-96 + 2 X 68-6) -213-5 = 20-26 Cals.
The heats of combustion of a few compounds are indicated in Table II. Heats
of combustion are dependent on constitution, so that isomeric compounds may
have different values ; similar remarks apply to the heats of combustion of polymeric
substances. Each radicle in a compound has a definite heat of combustion, called
its thermochemical constant — given the thermochemical constants of the constituents
of a molecule, the heats of combustion follow additively. Conversely, given the
heats of combustion of a compound, the presence of particular radicles can be
inferred.2 The principles just outlined can thus be applied : (1) To the determina-
tion of the thermal values of reactions which cannot be conveniently determined
by calorimetric measurements ; and (2) To the prediction of various chemical
transformations .
Table II.-
—Heats
OF Combustion.
Heat of
Heat of
Heat of
Heat of
combustion.
formation.
18-5
combustion.
formation.
Methane, CH4 .
213-5
0- Xylene, CgHjo
1084-0
20-3
Ethane, CgH, .
372-3
23-3
Naphthalene, CjoHg •
1241-8
-27-4
Propane, CgHg .
528-4
30-5
Anthracene, C14H10 •
1694-3
-.33-3
Butane, C4H10 •
687-2
35-0
Methyl alcohol, CH3OH
170-6
61-4
Ethylene, CgH^ .
3411
-14-6
Ethyl alcohol, C2H5OH
325-7
69-9
Acetylene, CaHg.
313-8
-51-4
Methyl ether, (CHgJaO
344-2
51-5
Benzene, CgHj .
784-1
-4-0
Ethyl ether, (C2H6)20
651-7
70-5
Toluene, CyHg .
933-1
2-3
It must be borne in mind that all deductions from these principles are " subject
to revision " owing to our ignorance of all the factors concerned in the reactions.
It might also be well to emphasize the fact that there is a relatively large error of
experiment in the determination of the heats of chemical reactions. The numbers
obtained by different experimenters vary, sometimes considerably. For instance,
the heats of combustion of acetylene and ethylene are variously given :
Acetylene .
Ethylene .
W. G. Mixter (1901).
. 313-8
. 345-8
M. Berthelot (1893).
315-7
341-1
J. Thomsen (1884).
310-0 Cals.
333-4 Cals.
These discrepancies are sometimes of considerable magnitude. For example,^
J. Thomsen found the heat of the reaction between lead acetate and zinc to be 34*95
Cals., P. Favre gave 31*2 Cals., and T. Andrews, 37-71 Cals.^a total variation between
the extremes of 6-51 Cals. The heat of formation of cupric oxide, CuO, and therefore of
all salts derived from it, is 37-16 Cals. according to J. Thomsen ; 43*77 Cals. according to
P. A. Favre and J. T, Silbermann ; and 3830 Cals. according to T. Andrews. The heat of
THERMODYNAMICS AND THERMOCHEMISTRY 711
formation of ferric chloride, FeCl^aq, is variously given by J. Thomsen at 99*96 Cals.,
by P. A. Favre and J. T. Silbermann at 10676 Cals., and by T. Andrews at 102-06 Cals.
These errors are magnified very much when the thermal value of a reaction is
estimated indirectly by the application of Hess' rule.
References.
1 G. H. Hesa,Pogg. Ann., 50. 385, 1840; Ann. Chim. phys., {[i), 74. 325, 1840; BvU. Acad.,
St. Petersburg, 7. 257, 1840.
^ H. S. Redgrove, On the Calculation of ThermocJi^mical Constants, London, 1909.
3 E. F. Herroun, Phil. Mag., (5), 27. 209, 1889.
§ 6. The Degradation or Dissipation o! Energy
There can be little question that the principle of the dissipation of energy implicitly
contains the whole theory of chemical combination.- — P. G. Tait.
Water may be transported from the top of a mountain to the valley below in a
variety of ways ; it may come down in underground channels, rivers, or rain ; or
in the form of snow, glaciers, or an avalanche. So may energy pass from a state of
high to a state of low potential in many and various ways, giving rise to mechanical,
thermal, actinic, chemical, electrical, or magnetic phenomena. In reality, the
so-called different forms of energy correspond with the tendencies which any given
system may have to change in particular directions. If there is a tendency for the
different parts of a system to come into closer contact, we have gravitation and
cohesion ; if there is a tendency to an equalization of temperature, thermal energy ;
and when there is a tendency to undergo transformation into another substance,
chemical energy. Hence, the definition : a chemical reaction is one mode by which
energy can be transferred from one state to another. Energy cannot be developed
from nothing, but it is derived from certain natural reservoirs — living beings,
falling water, moving air, fuels, etc. — in which energy is accumulated ; and certain
machines — the steam engine, galvanic battery, turbine, etc., draw from the reservoir
and transform one form of energy into another form without changing the total
amount of energy. Thus, heat, light, and electricity may be liberated during
chemical changes. To avoid the assumption that this energy comes from nothing,
it is postulated that the original system contained a definite amount of free or avail-
able energy — chemical energy. As H. Hertz (1894) expressed it : In order to
explain what is palpably before our eyes we are compelled to imagine behind the
things we see, other invisible things, and to search behind the barriers of sense for
a secret hidden accomplice. The hidden factor is here conveniently assumed to
be potential or chemical energy.
Not dead is matter though inert it seems,
A hidden life ensouls the eternal mass. — C. A. Lane.
If a substance can unite with another, it is said to possess chemical energy,
because it can do chemical work ; and conversely, substances which cannot combine
chemically with other substances have no available chemical energy, for they can
do no chemical work. During a chemical reaction, the chemical energy is trans-
formed into an equivalent amount of some other form of energy which is usually,
though not always, heat. Hence, the relation between chemical energy and heat
(thermal energy) is an important subject, which, for convenience, is called thermo-
chemistry ; and the general study of heat as a form of energy is called thermo-
dynamics. Chemical energy may also be transformed into electrical energy during
a chemical reaction, and that branch of chemistry which deals with the relation
between chemical energy and electricity (electrical energy) is called electrochemistry.
Just as chemical changes which are always accompanied by an evolution of heat
712 INORGANIC AND THEORETICAL CHEMISTRY
are called exothermal reactions^ so reactions which are accompanied by an evolution
of electrical energy have been called exo-electrical reactions ; and conversely, for
endothermal reactions and endo-electrical reactions. So far as we can tell, in all
phenomena, the same energy is at work, and the same fundamental principles apply
to all the specialized forms of energy. There is not a set of mutually exclusive laws
for chemistry, another set for electricity, another for heat or for mechanics.
Convenience alone dictates specialized versions of the same fundamental laws for
electricity, chemistry, etc. Strictly speaking, no form of energy can be singled
out and called potential energy, since each form of energy is potential with respect
to the other forms into which it can be converted.
The factors of energy. — -Water will flow from one vessel to another only when
there is a difference in the level of the liquid in the two vessels. The actual volume
of the water in either vessel does not matter. Again, heat will pass from one body
to another only when the temperature of the one is higher than the temperature
of the other. The flow of heat is not determined by the quantity of heat in either
the hot or the cold body, but rather by the difference in the temperature of the two
bodies. The heat in the fire-box of a locomotive can do work, not because it is
hot, but because it is hotter than its surroundings. In his well-known Reflexions,
N. L. S. Carnot (1824)i compared the production of work by une chute d'eau with
the fall of heat from a higher to a lower temperature, and referred to the latter as
une chute du calorique. Again, if two reservoirs of gas be connected by a cylinder
fitted with a sliding piston, the motion of the piston will not be determined by the
volume of the reservoir, nor by the quantity of energy contained in the gas, but it
will be determined by the difference in the pressure of the gas in the two cylinders.
Air confined in a closed vessel at atmospheric pressure might appear to possess no
energy because it can do no work ; but reduce the pressure of the surrounding air,
and the air confined in the vessel is then capable of performing work.
It is therefore possible to show that each form of energy has a dual nature, and
that every form of energy appears as if it were of two dimensions, for it can be
compounded of two factors 2 — mass and difference of level ; thermal capacity (or
maybe entropy) and temperature ; volume and pressure of gas. The one factor
is called the quantity, mass, or capacity factor, and the other, the strength, or
intensity factor. The two factors are combined not as a sum but as a product, for
if one factor diminishes towards zero, the other increases towards infinity :
Available energy = Capacity (quantity) factor X Intensity (strength) factor
To pass from generals to particulars, it is convenient to say that with volume
energy, the factors are pressure and volume ; with surface energy, the factors are
surface area and surface tension ; with distance energy, distance and force ; with
kinetic energy, mass (Jm) and velocity (F^) ; with electrical energy, quantity and
difference of potential ; etc. When the capacity factor is high and the intensity
factor low, more or less work may be got from that form of energy than if the
capacity factor is low and the intensity factor is high — all depends on the relative
magnitudes of the two different factors ; and two different sources of energy with
very different intensity and capacity factors may be able to perform the same
amount of work. If Cj, C2, . . . denote the capacity factors, and Z^, I2, . - • the
corresponding intensity factors of the different forms of energy associated with a
system, then, when the respective intensity factors change by small amounts
dZi, dl2, . . ., the work dW done by the system will be equivalent to
dW=C^dh-\-CM2-\- . . . ; or dW=UC.dI
The degradation of energy. — The law of conservation of energy does not describe
the direction in which a change will occur. It simply states that the amount of
energy lost by one body must be precisely equal to that gained by another ; it does
not say whether heat will flow from a hotter to a colder body or conversely.
Experience answers the question. Heat will be conducted from a hot to a colder
THERMODYNAMICS AND THERMOCHEMISTRY 713
body ; salt will diffuse from a solution of high to one of lower concentration ; and
generally, the trend of natural processes is all in one direction. This general tendency
can be formulated mathematically in terms of the transformations of energy which
come into play, and the relations which determine the final state of equilibrium.
This enables a prediction to be made as to the direction in which any given chemical
or physical process will progress.
In 1856, R. Clausius 3 laid down the hypothesis : Die Wdrme kam nicht von
selbst aus einem kdlteren in einen wdrmeren Korper Hbergehen — heat cannot sponta-
neously pass from a body at a low to a body at a higher temperature, but it can
be forced to do so either (i) by the application of energy from an external supply,
e.g. freezing machines and refrigerators raise heat from a cold to a hotter body by
performing work on the system ; or (ii) by a double transformation first into another
form of energy, say mechanical motion, and back again into heat. The pre-
ceding is one of the protean forms under which the second law of thermodynamics
or the second law of energetics can be stated ; it is also known as Camot's principle,
because N. L. S. Carnot first developed the idea in his celebrated memoir entitled,
Reflexions sur la puissance motrice du feu, published in Paris in 1824. W. Ostwald
(1892) expressed the same idea in his Studien zur Energetik previously cited : The
unUmited conversion of energy without intensity differences is impossible.
Several attempts have been made to deduce the second law from the first by
W. J. M. Rankine, S. H. Burbury, C. Szily, R. C. Nichols, L. Boltzmann,^ etc.
R. Clausius says that the second law is not contained in the first.
Just as water will always run down from a high to the lowest level that circum-
stances will permit, so generally, in all processes with which we are acauainted,
every known form of energy at a high potential always tends to nm down to energy
at the lowest potential circumstances will permit — the law of minimum free energy
— and one of the most interesting facts in connection with all natural changes
is this constant running down or degradation of energy. The law of minimum
free energy is analogous with the counter-statement in mechanics that a body
will always fall as far as it can, and that if it be free to fall, it will fall. The energy
so degraded has no longer a capacity for doing work, and the definition of energy
as a capacity for work is therefore faulty, for the principle of the conservation of
energy cannot be taken to mean that as the result of a given transformation the
capacity of the system to do work has remained constant.
Energy may be degraded slowly in a long series of transformations, or suddenly
in one bound ; in either case, the free energy under the new conditions becomes less
available for doing work. Every change which takes place in nature does so at
the cost of a certain amount of available energy. When we inquire whether or
not a certain transformation can take place, the question to be answered is : Will
the occurrence involve the degradation of energy ? If not, the transformation
will not take place under the given conditions. A moment's reflection will show
that in every transformation, the intensity factor will be diminished, and energy
then becomes less available for doing work. The intensity factor of energy controls
the direction of a given transformation, while the capacity factor largely controls
the quantity of change, that is, the amount of work performed during the change.
Water placed in a series of vessels in communication with one another will come
to rest when the surface of the water is at the same level in both vessels. Difference
of level here means that the gravitational energy has a different intensity in each
vessel. An electrical current will flow whenever there is an inequality of the
intensity factor — i.e. a difference of potential — at different parts of the circuit. If
the intensity factors of any particular form of energy in a system are not equal,
the system will be in a state of unstable equihbrium ; such a condition will
not be permanent, and energy will flow, so to speak, from one part to another
until the different intensity factors become equal.
This principle is true for any closed system, and if the universe is a closed system, it
must also dominate the universe. It has been pointed out that the earth is only part of
714 INORGANIC AND THEORETICAL CHEMISTRY
the universe, and it is continually gaining energy from the siui by radiation, and losing
energy, also by radiation from itself. Accordingly, owing to the universal and unceasing
tendency towards a degradation of energy, the universe is steadily passing from a state
in which energy at a high potential will be uniformly distributed at one uniform low
potential. The universe will then have a dead inert motionless existence at a uniform
temperature. Then follows the so-called thermodynamic paradox' — energy is continuously
being degraded ; the past duration of the universe extends through infinite time, therefore,
unless energy at a low potential is being restored to a higher potential, the degradation of
energy should have been completed long ago. Hence, it has been postulated that by some
hitherto unrecognized phenomenon, unavailable energy at a low potential is being raised
to available energy at a high potential, and that the second law of thermodynamics is being
somewhere and somehow reversed. It is, however, mere speculation to assume that because the
second law of thermodynamics is based on experience, and that there may be localities in
the universe where it does not apply, or that it may not have held good in past times, there-
fore there m^ttst be a source for the restoration of degraded energy. H. Elliot (1895) argued
that it is just as likely that the universe is infinite as that past time is infinite, and that
even the lapse of infinite time would not involve the extinction of all differences of potential.
An ancient philosopher — Heracleitus of Ephesus — has said that Travra pet — all
things are in motion, and it might be added that that motion always involves the
degradation of energy. The transformation of energy in a given system only ceases
when the available energy has run down to the level of its surroundings. The
system is then said to be in a state of stable equilibrium. The stability of a system
thus indicates how the system is related to its surroundings. For stability, a
system must be in equilibrium with its environment. A physical or chemical
change will progress until the different forms of energy which come into play are
exactly balanced, and this determines the final state of equilibrium of the system.
The condition of equilibrium. — In an isolated system, the condition necessary
for the equilibrium of any form of energy is that its intensity shall have the same
uniform value throughout. For instance, if the opposite sides of a bar of metal
have a different temperature, heat will be conducted from the hot to the cold end
until the temperature is everywhere the same ; a mass of gas will be in equilibrium
when it has one uniform pressure (of course neglecting the effect of gravitation,
which may be regarded as extraneous energy) ; etc. When one form of energy is
exactly balanced by another form, the system is in equilibrium, and a virtual
change of one form of energy will be balanced by a corresponding change in the
other form or forms of energy. Virtual change is a convenient term often used to
represent an infinitesimally small change " existing in effect, but not in actuality."
It is not clear at first sight what this phrase means. A virtual change is not a real
change, but rather an abstraction, and in place of " virtual," possible or potential
might be substituted.^ For instance, if a ball be suspended by an elastic string,
gravitation pulls the ball downwards, and elastic energy pulls it upwards ; a
virtual displacement of energy will occur if the ball were pulled an infinitesimally
small distance downwards, and the gravitational energy so expended were exactly
counterbalanced by the gain in the opposing elastic energy. The algebraic sum
of the energies involved in a virtual displacement of equilibrium must be zero
when the system is in equilibrium. This is the so-called principle of virtual work,
which is symbolized :
ZdE=0 ; or, 2]dW=0
where E and W respectively denote the energy and work performed ; and the
summation symbol U is intended to show that the algebraic sum of all the correlated
forms of energy is to be taken. For example, in a reversible chemical reaction in
equihbrium, the reacting substances and the products of the reaction are to be
taken ; the energy of the one increases and of the other decreases by a virtual
displacement of equilibrium. In a system where the mechanical forms of energy
capable of doing work are balanced, the principle of virtual work states that the
sum of the virtual work performed by the forces will be zero, that is, UdW=0.
The principle of virtual work was described by J. Bernoulli in 1717, and developed
THERMODYNAMICS AND THERMOCHEMISTRY 715
by J. L. C. Lagrange in 1788. Its application to chemistry was emphasized by
W. Ostwald in 1892.
The law of mass action has been deduced from this principle. When a system of re-
acting substances is in equilibrium, the volume energy of the component substances is Uvdp
=0. Let p and v respectively denote the partial pressures and volumes of the initial
substances, and P and V corresponding values for the products of the reaction. From
Boyle's law, pv^nR2\ and p^nRT/y; by differentiation, dp = — nRTdv/v^; by substitution
of dp in the condition of equilibrium, —RT{nEd\og v—NUdlog V}=0. Consequently,
at a constant temperature, replacing the molecular volumes v by their reciprocals, the
molecular concentrations c and G, we obtain nUd log c —NUd log (7=0; and on integration,
27 log c^*-f log k=^2J log O-^+log k', where k and k' are integration constants ; And n and
N respectively denote the relative number of gram-molecules of the initial and final products
of the reaction. The last expression can be represented kSc'^^k'ZC^ , or
C7i^i(7a^2. . -k
Expressed in words, in an opposing reaction, at equilibrium, the product of the concentra-
tions of the original substances is equal to the product of the concentrations of the end-
products of the reaction and the equilibrium constant, or, the effect of each reacting
substance is proportional to its concentration. This is the famous law of mass action of
Guldberg and Waage. J. Larmor and T. B. Robertson ' have shown that the mass law of
Guldberg and Waage can be derived from the gas law pv=RT, or p{v—b)=RT, where b is
constant.
Metastable equilibrium. — We are very familiar with systems in which the energy
has not run down to the level of its sm-roundings and yet everything appears to be
in a state of stable equilibrium. The stability is only apparent. As a matter of
fact, available energy does not always o/'tYseZ/' run down to the level of its surround-
ings. For some unknown reason, an influence — conventionally called chemical
inertia, hysteresis, or passive resistance — prevents the initiation of the process of
degradation of energy, a preliminary impulse is needed to start the process of degra-
dation in motion. Passive resistance is here used as a grouping or classification term.
It explains nothing. Just as the throttle- valve of a steam-engine must be moved
before the engine can start on its journey, or some watches, after winding, require
a slight shake before they start, so may a preliminary impulse be required to set the
process of the degradation of energy in motion. The flapping of an eagle's wing may
suffice to start an avalanche rolHng down the mountain side ; with gunpowder, the
preliminary impulse may take the form of heat ; with a mixture of hydrogen and
oxygen, an electric spark, or the mere presence of spongy platinum ; with a mixture
of hydrogen and chlorine, a flash of light, or the addition of a piece of charcoal ; with
fulminate of mercury, a sudden shock ; while the addition of a minute crystal will
start the process of crystallization in a supercooled solution of sodium thiosulphate.
We may thus have a state of metastable, apparent or false equilibrium, as well
as a state of true or stable equilibrium. We naturally inquire : Is there any test to
distinguish between states of real and states of apparent equilibrium ? We know
that if a gas is in equilibrium with regard to volume and pressure, it will satisfy the
conditions of Boyle's law ; volume and temperature, Charles' law ; etc., but we have
not always such useful tests at our disposal.
References.
^ R. Clausius, Pogg. Ann., 120. 426, 1863 ; N. L. S. Camot, Reflexions sur la puissance motrice
du Feu, Paris, 1824 ; New York, 1897.
2 G. Helm, Die Energetik, Leipzig, 253, 1898.
3 R. Clausius, Pogg. Ann., 79. 368, 500, 1850; 93. 481, 1854; 116. 73, 1862; Phil. Mag.,
(4), 2. 1, 102, 1850; (4), 12. 81, 1856.
4 C. Szily, Phil. Mag., (5), 1. 22. 1876; S. H. Burbury, ih., (5), 1. 61, 1876; R. C. Nichols.
ih., (5), 1. 369, 1876; W. J. M. Rankine, ib., (4), 7. 249, 1861; L. Boltzmann, Sitzher. Akad
Wien, 63. 712, 1871.
6 Lord Kelvin (W. Thomson), Proc. Roy. Soc. Edin., 3. 139, 1852 ; 8. 325, 1876 ; Phil. Mag.,
716 INORGANIC AND THEORETICAL CHEMISTRY
(4), 4. 256, 1852 ; (5), 7. 344, 346, 1879 ; (5), 33. 291, 1892 ; Mathematical and Physical Papers,
Cambridge, 1. 511, 1882 ; 5. 1, 1911.
« E. Mach, The Science oj Mechanics, Chicago, 49, 1902.
' J. Larmor, Phil. Trans., 190. A, 276, 1887 ; T. B. Robertson, Journ. Phy.s. Chem., 10.
521, 1906.
§ 7. Bound and Free Available Energy
Matter, whatever it is, must be held to be so adorned, furnished, and formed that all
virtue, essence, action, and motion may be the natural consequence and emanation thereof. —
Francis Bacon.
Energy is the result of a particular state or condition of matter in virtue of which any
definite portion may effect changes in any other portion. — C. F. Barker (1892).
Mechanical and other forms of energy can often be transformed completely into
heat, but the reverse operation is subject to certain limitations since a certain propor-
tion always escapes conversion and is lost. Similarly, when a system undergoes a
chemical or physical change, a certain portion of the energy is simultaneously
transformed into heat. The loss or leakage of energy does not mean that energy
is annihilated ; there is nothing to suppose that the law of the perdurability or
conservation of energy is invalid. The so-called loss of energy means that part
of the energy is degraded — by friction, viscosity, etc. — from a high to a low
potential, and the capacity of the transformed energy for work is diminished. It
is impossible to transform a quantity of heat into work without an accompanying
change in the condition of a portion of the energy of such a nature that its capacity
for work is correspondingly diminished. It therefore follows that the principle of
conservation of energy is incapable of experimental demonstration, for it is only
possible to measure the ratio in which the transformation of energy from one form to
another is accomplished. Again, the definition of energy as capacity for work is
not altogether satisfactory because a fractional part of the heat does not conserve
its capacity to perform work when a transformation of heat into work is attempted.
H. von Helmholtz, in his IJeber die Thermodynamik der chemischen Prozessen
(1882),! pointed out a useful concept by showing that the total available energy
of a system appears as if it were on two different planes — one portion, called the
free energy of the system, is capable of doing chemical, electrical, or mechanical
work ; and a second portion, called the bound energy, is rendered unproductive
during the change, for it is frittered away as heat. The latter can be regarded as
energy which must of necessity be wasted during the operation, or as the energy-
COSt of the reaction. The free energy of a system is a measure of the work which
can be performed by the chemical process ; the bound energy represents the energy
lost by leakage during the reaction, because nature has determined that the rate of
exchange, so to speak, is against the conversion. This must not be taken to mean
that the energy of a body or system of bodies is not homogeneous, and can be sepa-
rated into two parts with different properties ; for the proportion of free to bound
energy in any given process changes with the temperature. The actual proportion
appears to be determined by the changes in the kinetic energy of the molecular or
atomic motions ; by the separation of the molecules against intermolecular attrac-
tions ; by changes in the rotational or vibratory energy of the atoms ; by alterations
in the electrical state or thermal capacity of the molecular systems concerned in the
reaction ; etc. Again, the free energy, or the energy which can perform work
during an isothermal change, is not always derived from the internal energy, thus,
the internal energy of an approximately ideal gas does not change during an isother-
mal expansion when the gas performs no external work.
Free energy. — The maximum amount of work a reaction can do when it is carried
out reversibly at a constant temperature, is called the free energy of the reacting
system. If the free energy of a spontaneously occurring natural process were nega-
tive, it would progress without absorbing energy from its surroundings, and on
being reversed (by the addition of heat), it would transform this heat into work,
THERMODYNAMICS AND THERMOCHEMISTRY 717
and a perpetual fount of energy would be available. This is in conflict with the law
of excluded perpetual motion ; and it is accordingly inferred that every spontaneous
process must do work, and that a system incapable of doing work is incapable of
spontaneous change ; such a system must also be in a state of stable equilibrium,
and accordingly, for stable equilibrium, the free energy o! a system must have a
minimum vsdue. If a system can do no work, it cannot change except by the
application of external energy. The decrease in the free energy of a reacting
system is a measure of the work which can be performed by the chemical
process. Free energy is thus synonymous with the ability to perform work. That
system which can perform the greatest amount of work, when it is carried out
isothermally and reversibly, will be most likely to occur ; i.e. of all possible chemical
changes, that which involves the greatest decrease in the free energy will be
most likely to occur. Given a table of the free energy changes which occur during
the formation of various compounds from their elements, by different processes,
it would be possible to calculate the change in the free energy attending other reac-
tions, just as in ordinary thermochemical calculations, the thermal value of a given
reaction can be calculated from the heat of formation of the various reacting compo-
nents. Tables of the free energy may thus take the place of tables of the heats of
formation of different compounds, and be employed to predict (i) The maximum
work which may be expected from a given reaction ; (ii) The minimum amount
of work which would be necessary to produce a certain reaction; (iii) The circum-
stances under which a given reaction will progress ; and (iv) How nearly a given
reaction will run to an end.
The available, total, or internal energy. — The absolute amount of energy of a
substance cannot be measured because all measurements are concerned with differ-
ences of energy existing between different bodies or systems of bodies, or between
a substance in two different states. This is not particularly a disadvantage, since
it is all that is required in the present state of science. For convenience, the
total available energy involved in any reaction is considered to be the algebraic
sum of the free and bound energy. The total intrinsic or internal energy U —
sometimes called the potential energy — transformed in a reaction is measured in a
calorimeter as the heat of the reaction Q when due allowance is made for external
work done on or by the reacting system against atmospheric pressure. In the latter
case, the potential energy of the system is the difference between the heat energy Q
and the external work. The total energy must not be confused with the free
energy. The whole of the free energy in any given system may disappear without
diminishing the total energy of the system. If U represents the total energy of an
isolated system, W the free, and q the bound energy, then, by the law of the conser-
vation of energy, U is constant ; and if W becomes zero, q= U. In that case, the
Q units of energy of the system can do no work. On the other hand, if q=0, then
U=W, and all the energy of the system is free and capable of doing work. If
q=iU, half the total energy can do work, and half will not be utilizable for doing
work. M. Berthelot's principle of maximum work assumes that the total energy of a
chemical reaction is equal to the free energy, that is, to the energy available as work,
when the work is carried out reversibly. There is nothing to show that ?7— Tf =0,
and accordingly, heat may be absorbed or evolved when a reaction is carried out
under these conditions. It is not the maximum production of heat energy which
determines if a reaction will necessarily occur, since some spontaneous reactions
absorb heat.
Reversible and irreversible processes. — When a body has fallen a certain dis-
tance in vacuo, it would rise again the same distance as it fell if its velocity could be
reversed ; and by relieving the pressure on liquid water confined in an air-tight
cylinder, fitted with a frictionless piston, and maintained at a constant temperature,
the water could all be vaporized, and by reversing the pressure, the vapour could be
condensed back to its original liquid state. In reality, a frictionless piston is impos-
sible, and only when the pressure required to compress the vapour is exactly equivalent
7l« INOEGANIC AND THEORETICAL CHEMISTRY
to that required for the expansion is the process considered to be strictly reversible.
Again, by raising the temperature of a closed vessel containing calcium carbonate,
the compound will dissociate, and by lowering the temperature, the products will
recombine to form the original compound ; similarly, if a cyHnder contains a satur-
ated solution of, say, potassium nitrate in the presence of some of the undissolved
salt, the solution will alter its strength by diffusion, if the temperature be slowly
raised ; and if slowly cooled to its former temperature, the original condition will
be restored. If there were any leakage of energy due to the viscosity of the solution,
so that more heat energy were required for the heating than for the cooling, the pro-
cess would not be considered strictly reversible. A process is considered to be strictly
reversible only when it can be made to pass back from its final to its initial stage
successively, and in the reverse order through all the stages traversed in the direct
process by the application of external agents which are equal in magnitude but in
opposite directions. In illustration of an irreversible process, if a cylinder containing
a saturated solution of magnesium sulphate, MgS04, in presence of an excess of the
same salt, be treated as in the case of the cylinder of potassium nitrate, the hepta-
hydrate, MgS04.7H20, separates during the cooling. Again, if potassium chlorate
be heated in a closed vessel, the oxygen and potassium chloride which are formed
will not recombine to reform potassium chlorate when the system cools, and hence
the process is irreversible. G. H. Bryan 2 has stated :
While students of reversible phenomena have had fairly straightforward problems
to solve, the problem of irreversibility still remains to a great extent a mystery and nobody
seems to have got to the bottom of it. The irreversible phenomena of the universe all have
a certain definite trend, and lead to the transformation of energy into certain definite
forms. We say that certain forms of energy are less available than others, but why the less
available forms are those associated with what are commonly called heat phenomena is a
riddle . . . still unsolved.
No known natural process is strictly reversible, because, if a process goes in one
direction, experience shows that it cannot be made to go in exactly the opposite
direction by reversing the same outside agency. Among the various causes which
make real processes more or less irreversible are viscosity in liquids, imperfect
elasticity in solids, friction, diffusion, radiation, radioactivity, conduction, and types
of electrical, magnetic, and chemical action. It is, however, possible to imagine two
transformations — direct and inverse — to be conducted by a continuous series of
infinitesimally small changes so that the system is all the time infinitely near being
in a state of equilibrium ; and it has been agreed to call such a process reversible.
Hence, in a reversible transformation or process, two imaginary operations are
performed on a system whereby the system traverses the same intermediate
states in a continuous series of indefinitely small stages, but in the reverse order,
so that the transformation is attended by the production of as much external work
W as would have been expended in restoring the system to its original condition
W\ and the total work performed in a reversible cycle is zero, for If =1^'. More
work cannot be produced by a direct transformation than is required for its reverse,
or perpetual motion would be possible ; and if less work is required, the system is
irreversible. Hence, if a change is reversible, it works under the most favour-
able conditions, for it furnishes the maximum amount of work which it is
capable of producing, or else the minimum amount of work is expended in bringing
about the change — otherwise expressed, in a reversible transformation, the free
energy is a maximum, the bound energy a minimum.
References.
1 H. von Helmholtz, Sitzber. Akad. Berlin, 22, 825, 1882; 647, 1883; Ostwald's Klassiker.
124, 1902; Physical Memoirs, 1. 43, 1891 ; R. A. Lehfeldt, Electrochem. Met., 3. 126, 1903;
M. M. Garver, Journ. Phys. Chem., 15. 20, 613, 1911.
2 G. H. Bryan, Proc. Roy. 80c. , 80. 13, 1908.
THERMODYNAMICS AND THERMOCHEMISTRY 719
§ 8. The Amount o! Heat which can be Utilized for doing Work
Every change in the distribution of matter in a given system, under given conditions
is accompanied by a definite energy change. Therefore the laws which govern changes of
energy, are the laws which govern transformations of matter.— A. J. Lotka (1913).
To what extent can heat be converted into work ? The maximum quantity of
heat q which can be reversibly converted into work W under ideally perfect condi-
tions, working between the temperatures T and T-\-dT, at constant volume without
doing external work, was shown by N. L. S. Carnot (1824) i to be equal to the product
of the change of temperature into the quantity of heat q absorbed, divided by the
absolute temperature. In symbols,
dq=dW=q^-^; or, c?If=|c^r . . . (1)
Hence, said R. Clausius (1850),2 when the temperature of a quantity of heat q is
changed by a small amount dT, the fraction dTjT of q is transformed into work
provided no heat is lost. N. L. S. Carnot illustrated the principle by referring to
an arrangement for utilizing a fall of water. Suppose water to be in a reservoir T^
feet above sea-level feeding a mill Tg f^et below ; let w denote the amount of water
which falls in unit time. If the disposition of the system be ideally perfect, the energy
of the falling water available per minute at the mill will be the weight of water
multiplied by the fall, or WT2, or, referred to sea-level, the free energy of the water
at the reservoir, per minute, is wTi. The free energy of the water starting on its
downward journey from the mill is WT1—WT2, or w{Ti—T2), so that the amount
of free energy utilized at the mill is w{1!i—T2)lwTi or (Ti—T^jTi per minute.
If q represents the total quantity of free energy of the water in the reservoir, the
amount actually utilized will be q{Ti~T2)ITi. Making the necessary changes in
the meaning of the terms, this same result is obtained by integrating equation (1),
if W is put equal to W1—W2 or to qi—q2,
W=q
T1-T2
meaning that the theoretical maximum quantity of work W which can be obtained
from a quantity of heat q working between the absolute temperatures Tj and To,
is equal to the product of the quantity of heat into the change of temperature divided
by the higher temperature, and is independent of the nature of the working sub-
stance. If the combustion of carbon under a steam boiler furnishes a quantity of
heat q, then, under the very best (ideal) conditions, the amount of work W which
can be derived from the steam engine will be q{Ti—T2)ITi, where Tj denotes the
temperature of the boiler, and T2 that of the condenser. This shows that the ideal
limit to the efficiency of a thermal process, or the fraction of the total energy q
capable of doing work W/q, is equal to the ratio of the difference of the absolute
temperatures between which the operation is performed and the maximum absolute
temperature.
In processes available industrially, the lower temperature limit is fixed by the tempera-
ture of the cooled water, and the upper temperature limit is fixed by the temperature
generated by the combustion of the fuel in air — about 1500°. The efficiency of a process
dependent on steam and cold water is increased by using higher pressures in the boiler, but
the pressure of steam in a boiler rises very rapidly with rise of temperature, and this limits
the efficient use of high pressure steam boilers.
Carnot's theory has been reasoned out another way : The amount of energy
given out by a cooling gas is proportional to the change of temperature, so that the
maximum amount of energy can be obtained only by cooling the gas to absolute
zero, and consequently, if the gas be cooled one-tenth of its way to zero, it will yield
but one-tenth of the available energy.
Examples. — (1) The explosions in the cylinder of a gas engine raise the temperature to
927°, and the temperature of the exhaust is 127°, what is the theoretical eflftciency of the
720 INORGANIC AND THEORETICAL CHEMISTRY
process ? Since the efficiency of a process indicates the fractional part of the total energy
utilized in doing work, the required efficiency will be dq/q=dT/2\ or (1200— 400)/l 200,
or two-thirds of the heat would be converted into useful work under ideal conditions, and
the remaining third would be wasted.
(2) The temperature of the boiler of a steam engine is 127°, and the condenser 17°,
what percentage of the heat is theoretically wasted ? Here the theoretical efficiency is
(400— 290)/400 =0-275, or 27*5 per cent. Hence, 72*5 per cent, of the heat energy is
wasted.
Equation (1) may be called Camot's equation. It can be extended to chemical
reactions by assuming that the reaction takes place in one direction at the tempera-
ture T, and in another direction at a temperature T-^-dT ; and if W and Q respec-
tively denote the free and total energy at the temperature T, q in the expression
dq/q, 01 dWlq=dT/T, can be replaced by W^Q, where If represents the free energy
or maximum work, and Q the thermal value of the reaction, i.e. the heat absorbed
(+) or evolved (— ) in the reaction as measured in the calorimeter. Hence,
W—Q=q, and W—Q represents the bound energy absorbed or evolved during
the reaction. Consequently,
dW
W-Q=t'^ (2)
This form of the expression is called Helmholtz's eauation, after H. von Helmholtz
(1882), although A. Horstmann ^ had previously deduced a similar equation in 1870.
Equation (2) means that the difEerence between the free and total energy in a chemical
reaction under the given conditions is equal to the product of the absolute tempera-
ture T and the temperature coefficient of the free energy, dW/dT, which represents
the change in the amount of work performed per unit change of temperature.
It will be evident that Berthelot's rule — that the total energy of the thermal
value of a reaction is numerically equal to the free energy or maximum work which
the reaction can perform — will be strictly valid only at absolute zero, when T=0,
for dWjdT always has a finite value, and the term T.dW/dT can be suppressed. At
absolute zero, therefore, W—Q=0, or W=Q, and the thermal value of the
process measures the driving force of the reaction. At ordinary temperatures, which,
after all, are not very far removed from absolute zero, the difference between W—Q
and T.dW/dT is not usually great, and Berthelot's rule becomes a useful approxi-
mation, particularly in chemical technology, where a measure of chemical eneigy
in terms of the thermal energy of a process is desired. There are so many cases
where the rule is approximately correct, that one is apt to forget that it is after all
but a " half truth " and not the " whole truth."
H. von Helmholtz's equation is used in the interpretation of a great many pheno-
mena. W. Nernst (1907) ^ was so impressed with its importance that he said :
" It contains in a general manner all that the laws of thermodynamics can teach
concerning chemical processes." In illustration, it can be applied (1) to show the
influence of temperature on the thermal value of a reaction — J. H. van't Hoff's
equation ; (2) to show the relation between pressure and processes of volatilization,
melting, and allotropic transformations — Clapeyron's and Clausius' equations ;
(3) to show the effect of temperature on the electromotive force of a cell — H. von
Helmholtz's equation ; (4) to show the influence of temperature on the surface
energy of a liquid — Lord Kelvin's equation ; etc.
Example.' — How does the pressure of an ideal gas kept at a constant volume change
with temperature ? For an ideal gas, Q=0. Let the gas at a temperature T change to
a temperature T-\-dT, when the pressure changes from p to p+dp. Then, W=pdv,
and dW=dp.dv. Substituting these results in Helmholtz's equation, pdv = T{dp.dv/dT)f
or p/T=dpJdT, which is virtually Charles' law.
References.
* N. L. S. Carnot, Reflexions sur la puissance motrice du feu, et sur les machines propres a
developer cette puismnce, Paris, 1824 ; New York, 1897 ; Harper's Scientific Memoirs, fi, 1899 ;
THERMODYNAMICS AND THERMOCHEMISTRY 721
OstwaWs Klassiker, 37, 1892 ; E. Clapeyron, Jmirn. J^cole Polyt., 14. 153, 1834 ; Taylor's Scien-
tific Memoirs, 1. 347, 1837 ; Fogg. Ann., 59. 452, 1843.
2 R. Clausius, Pogg. Ann., 79. 368, 1850 ; Phil. Mag., (4), 2. 1, 102, 1857 ; Harper's Scientific
Memoirs, 6. 1899.
3 A. Horstmann, Liehig's Ann. Suppl, 8. 112, 1872; Liebig's Ann., 170. 192, 1873; Ost-
wald's Klassiker, 137, 1903 ; H. von Helmholtz, Sitzber. Akad. Berlin, 22, 825, 1882 ; Wissen-
schaft Abhandlungen, Berlin, 2. 958, 1884; OstwaWs Klassiker, 124, 1901 ; Physical Memoirs,
London, 1, 43, 1891 ; F. Massieu, Journ. Phys., 6. 216, 1877; J. W. Gibbs, Trans. Conn. Acad.,
3. 343, 1878 ; ScierUific Papers, London, 1. 56, 1906.
* W. Nernst, Experimental and Theoretical Applications of Thermodynamics to Chemistry,
London, 1907.
§ 9. Non-productive Energy. Entropy
Nature never undertakes any change unless her interests are served by an increase of
entropy, while man endeavours so to make use of those changes allowed by nature that his
own interests- — namely the acquisition of available energy^ — are served as completely as
possible.-— M. Planck (1903).
No change occurs in nature without incurring waste. — J. Swinburne (1904).
The driving force of all natural events is the difference between the existing entropy and
its maximum value. — L. Boltzmann.
While work may be wholly converted into heat, only a definite fraction of heat
can be converted into work ; otherwise expressed, if a quantity of heat Qi is received
by a system at a temperature Ti, only the portion ft— Q2=0i(^i~^2)/^i
can be converted into work by a reversible process ; the remaining quantity of
heat ^2 is simply transferred to the lower temperature Tn. The reversibility of the
process implies that the portion of heat which was not transformed into work can be
restored to its former temperature level Tj, with the same expenditure of work
Qi~Q2~Qi{Ti — ^2)1^1- If *h® process is not reversible, and no known natural
process is reversible in this sense, more heat flows to the lower level Tg than in the
reversible process, and accordingly less heat is transformed into work ; the surplus
cannot be brought up to its former level T^ except by the expenditure of more
work than is represented by Qi{Ti—T<2)ITi. Hence, in all real thermal processes,
a fraction of the heat is lost for mechanical work ; and there is a constant wastage
of mechanical energy. The expression ft— Q2=ft(^i~^2)/^i> indicated above,
reduces to
Ti T^ ^^^
meaning that when a body changes from one state to another by the addition of a
quantity of heat ft at Tj, the change can be reversed by taking away a quantity
of heat ft ^^ ^2> provided these four quantities are related so that QilTi=^Q2lT.^.
A gas, therefore, which receives ft units of heat at Ti will retain its energy until
it has given up the equivalent of ^2 units of heat at Tg. If the process is not reversible,
Q1/T1—Q2IT2 is not zero, but some greater magnitude.
In 1854, W. J. M. Rankine 1 called the magnitude Q/T the thermodynamic
function, but R. Clausius' term entropy — from Ivrpoirifj, a change — has alone been
adopted. R. Clausius said :
I hold it better to borrow terms for important magnitudes from the ancient languages
so that they may be adopted unchanged in all modern languages ... I have intentionally
formed the word entropy so as to be as similar as possible to the word energy ; for the two
magnitudes, to be denoted by these words, are so nearly allied in their physical meanings,
that a certain similarity in designation appears to be desirable.
Entropy is here symbolized by <j). When a body at the absolute temperature T
receives a small amount of heat dQ, the entropy is said to have increased by an
amount d<f>=dQjT. If a gas of entropy (/>, pressure f, temperature T, and volume v
receives heat and performs work so that it undergoes various changes and returns
VOL. I. 3 a
722 INORGANIC AND THEORETICAL CHEMISTRY
to its former values of T, p, and v, its entropy will be the same as it originally-
possessed — viz, <l>. Neither the heat added to and taken from the gas nor the
work done by and upon the gas are necessarily the same ; but the entropy given
to and taken from the gas is exactly the same.
It can be shown (i) that what is true of a gas is also true of a liquid and of a solid —
homogeneous or heterogeneous ; (ii) that the entropy <^ of a complex system is the
sum of the entropies of its component parts 0i+<^2+</>3+ • • • =^ I and (iii)
that if the entropy of unit mass of a material in a given state be </>, the entropy of
a mass m of the same material in the same state is m<f) ; accordingly also (iv) that
^i<^i+^2<^+ • • • =wi^j where nii^ m^ . . . denote the masses of the component
parts of the system of mass m ; </>i, ^2 • • • ^^^^^ entropies.
The measurement of entropy. — When a body receives 100 units of heat at a
temperature 50° K., its entropy will be increased 100/50 or 2 entropy units. This
unit was called a claus by J. Parker (1891), and a rank by J. Perry 2 — respectively
after R. Clausius and W. J. M. Rankine. Hence entropy can be measured as
_, ^ Units of heat
Entropy =-pr — 1-7—7 r— units
^'' Absolute temperature
The measurement of the entropy of a system is made under such conditions that,
excluding changes of volume, the system gives up no energy other than heat. The
external pressure of a changing body may not be equal to the internal pressure :
for instance, in most chemical reactions, the volume of the system alters quite
independently of the heat changes. The total change of energy is then equivalent
to the heat energy Q evolved or absorbed from outside minus the amount of energy
W consumed in the work of expansion against an atmospheric pressure f, plus the
change jt.dv in the volume energy. The total change of energy divided by the abso-
lute temperature T represents the change expressed in entropy units, or
Q-W-Yjpdv .^
Entropy = j=j-^ units
Here, Q—W represents the change in the internal energy. If the conditions are
such that the external pressure is equal to the internal pressure required to pro-
duce the change in volume, W will be equal to pdv, and the entropy will then be
(t>=QIT. This condition is not usual ; in general, the entropy depends on the total
change of internal energy, plus the change in volume energy. For example, when a
gas expands into a vacuum, its temperature does not change, and its internal energy
is not changed, for Q and W are both virtually zero, but the volume energy is increased
by jpdv. The entropy of the change is therefore pdv/T. To bring the gas back
to its former condition requires an expenditure of energy equivalent to pdv. The
entropy of the two operations can change only by the gas receiving or giving up heat
to its surroundings. Hence, the entropy of the gas at the end of the operation is
the same in magnitude as it possessed just after expanding into the vacuum, and this
value is greater than it possessed at the beginning by the magnitude of the volume
change pdv. The change of ice into water at 0° in vacuo is measured by the heat
rendered latent, A/T, since both W and pdv are zero.
Example.- — If a gram of water at the absolute temperature T receives dQ units of heat as
its temperature is raised to T-\-dT, then, since dQ for water is very nearly equal to dT, the
gain of entropy d<l> will be d<f)=dQIT, or d<f)=dT/T ; and if the entropy of liquid water at
0° or 273° K. be conventionally regarded as zero, the entropy j>, at T° K., will be, in
natural logarithms,
/ d<f> =
Jo J ',
If the water at T° be also converted into steam at T°, it will be necessary to add on the
term A/T, where A denotes the latent heat of vaporization, at T°.
If a body absorbs an amount of heat dQ so that its energy increases by an amount
THERMODYNAMICS AND THERMOCHEMISTRY 723
C^dT, where C„ denotes the thermal capacity at a constant volume, and the work
pdv is done against atmospheric pressure, dQ=CidT-{-j)dVj and the entropy of
the body will be
^^^,,,,^^.^1+2^^ .... (2)
From the gas equation pv=RT, it follows that fjT^Rjv, and supposing that Cv is
independent of the change of temperature, the entropy ^ of a perfect gas, per
gram-molecule, is
=Cvf^+Rf~+I; 01 cl>=C,logT+Rlogv+I . . (3)
Several different but equivalent forms of this relation for ideal gases can be obtained
by means of the relation jpv=RT ; and Cp—C^=R. The integration constant, /,
can be evaluated only by the introduction of another assumption. To overcome
this difficulty, W. Nernst 3 postulates what he calls the third law of thermodynamics :
In an isothermal process involving liquid or solids, the change of entropy approaches
zero as the temperature approaches absolute zero; and at the limit, the absolute
zero of temperature, the entropy of every substance is zero. Consequently, if the
entropy at absolute zero be taken as zero, the entropy at any assigned temperature
will have a definite positive value determined solely by the thermodynamic state of
the system. This means that when a body is heated from the absolute temperature
T, the entropy will be given by
when the changes of volume are negligibly small, provided also that the body suffers
no change of state — e.g. from solid to liquid ; or from one crystalline form to another.
It will be observed that the integration constant has been eliminated in virtue of
Nernst's assumption. It cannot be assumed that the specific heat at constant
volume is independent of temperature ; and before the integration is performed,
the relation between the specific heat at constant volume, C^, and temperature must
be known.
For very low temperatures P. Debye's limiting equation ^ Cv='a(Tld)^ can
be used for evaluating (f) ; here « is a universal constant, and ^ is a characteristic
constant for each substance. For ordinary temperatures, P. Debye's formula is
too complicated for practical use. S. Pagliani (1915), and G. N. Lewis and G. E.
Gibson (1917) have evaluated the integral for a number of elements ; the latter,
using a method of approximation, obtained the values for the atomic entropies of
the 48 elements indicated in Table III.
In order to determine the free energy of formation of any compound from its
elements, it will be sufficient to know the entropy of the compound and of its ele-
ments at the same temperature, as well as the heat of formation of the compound
from the equation dE—dQ=—Td(f), where dE denotes the increase of free energy ;
dQ the thermal value of the reaction ; and d(j> the increase in entropy at the constant
temperature T. G. N. Lewis and G. E. Gibson add : The atomic entropies of the
elements appear to show the same sort of periodicity with respect to atomic weights
that occurs with other important atomic properties ; and according to S. Pagliani
the entropies at the melting point calculated from (4), for Cp in place of C^, diminish
as the atomic weights increase — particularly for the groups in Mendeleeff's system
when the metals are separated from the non-metals. As a result of his examination
of G. Tammann's statement ^ that the entropy per gram atom of the metals before
fusion is the same for all the metals, S. Pagliani found that the products of the
entropies and the atomic weights of the elements are not constant.
Physical analogies of the mathematical concept of entropy. — It is extremely
724
INORGANIC AND THEORETICAL CHEMISTRY
difficult to get a physical apart from a mathematical conception of entropy. The
work W performed by an idea,l gas expanding a volume v against a constant pressure
p, is W=pv ; so is the entropy related to the quantity of heat and temperature by
the formally analogous expression Q=T(j>. We are here in the same position with
Table III. — The Atomic Entropies
J OF THE Elements ai
'25°.
Element.
Entropy.
Element.
Entropy.
Element.
Entropy.
Aluminium .
6-9±*
Iodine
15-7zt:l
Potassium .
19-7±i
Argon (gas)
36-4±2
Iridium
8-7
Rhodium
7-6
Beryllimn .
7-3
Iron
6-6
Ruthenium
6-9
Bromine (liquid) .
18-5±2
Lanthanum
13-7
Silicon (metal) .
4-7±i
Cadmium .
11-6
Lead
15-4dbi
Silver
10-2±i
Calcium
ii-o±j
Lithium
7-6
Sodium
12'2±i
Carbon-diamond .
0-6±i
Magnesium
8-3±i
Sulphur (rhom.) .
7-6±l
Carbon-graphite .
l-3±^
Manganese
7-3
Sulphur (monoc.)
7-8±l
Cerium
13-8
Mercury (liquid)
17-8±1
Thallium .
14-6±J
Chlorine (gas)
25-7±2
Molybdenum
7-5
Thorium
13-6
Chromium .
5-8
Nickel
7-2
Tin .
11-5
Cobalt
7-2
Nitrogen (gas) .
22-8±l
Titanium
6-6
Copper
8-OdbJ
Oxygen (gas) .
24-l±l
Tungsten
8-4
Gold .
110
Osmium .
7-8
Uranium
IM
Helium (gas)
29-2il
Palladium .
8-9
Zinc .
9-8±^
Hydrogen (gas) .
15-9±1
Platinum .
10 0
Zirconium
9-5
regard to the entropy <j> as if in the former case we had no conception of volume v.
We might then speculate as to the physical significance of volume just as to-day
we wonder what is the physical significance of entropy. Volume would then be
defined as something whose change is measured by the ratio TF/^, just as entropy
is a magnitude whose change is measured by the ratio QjT. Again, just as volume
is regarded as the capacity factor and pressure the intensity factor of the volume
energy of a gas, so it has been suggested that entropy is the capacity factor and
temperature the intensity factor of heat energy. G. Zeuner ^ called entropy, heat
weight — Wdrmegewicht — ^probably as a result of an analogy between the part entropy
plays in the energy developed by the fall of heat from a state Ti^i to the state
T2<^2» s-^d ^^® P^r* played by weight in the energy developed by the fall of a mass
from a state hiWi to a state ^2^2 — when h represents the height, and w the weight
of the bodies indicated by the subscripts.
Following a suggestion by 0. E. Meyer, in 1872, L. Boltzmann^ showed an
intimate relation between the entropy of a system and the probability of a particular
state in which the system can exist. It is assumed that the state of the system can
be defined by the volume, temperature, etc. When an isolated system can pass
from one state A to another state B, the probability of, say, the state A represents
the relative chance that the system will exist in that state and not in the other. If
the probabiUty of the state A were J, there is an even chance that the system will
exist in either state just as there is an even chance of heads or tails appearing after
the toss of a penny. In all spontaneous natural processes, the probability of the
state of the system increases and tends towards a maximum probability (unit
probability is certainty). The entropy of a system, said L. Boltzmann, is determined
by the probability of the state of its molecular motion ; and M. Planck has
shown that if P denotes the probability of the state of a system, and (f> its entropy,
(l>=k log P, where A; is a universal constant which is independent of the chemical
nature, and other variations in the condition of the system. In illustration, it is just as
probable that heat will pass from a cold to a hotter body^ — say that a kettle of water
over a fire will become colder — as that the entropy of a system will decrease by a
spontaneous natural change. With R. Clausius, the transformation value of a
particular state is equal to the difference between the entropies of the initial and final
THERMODYNAMICS AND THERMOCHEMISTRY 725
states ; with L. Boltzmann, the entropy of a physical system is a definite condition
which depends solely on the probability of that state.
The law of maximum entropy. — If a system of two bodies A and B, respectively
at the absolute temperatures Ti and T2, be in thermal communication, and so iso-
lated that no heat is lost or gained from external sources, the total energy remains
unaltered although heat passes from the hotter body A, at the temperature Jj, to
the colder body B, at the temperature T2. When the infinitesimally small quantity
of heat dQ has passed from A to B, the entropy of A is diminished by dQjTi, and the
entropy of B is increased by dQjT^- Consequently, the entropy of the whole system
will be increased so that
Gain of entropy =
<i-i>«
This expression must be positive, since T^ is greater than T^- The effect of the
transfer of an infinitesimally small quantity of heat dQ from a hot to a colder body is
to increase the entropy of the system. The transfer of heat, and consequently also
the gain of entropy, will continue by a series of infinitesimally small stages until
the bodies A and B have one common temperature, when the entropy of the system
will have attained its maximum value. While the first law of thermodynamics
says that in an isolated system, all possible changes of state leave its energy unaltered,
the second law may be understood to mean that in an isolated system all possible
spontaneous changes of state produce an increase in its entropy. The law of maxi-
mum entropy holds good however many bodies be included in the system, and unless
something has been overlooked, it will apply to the whole universe considered as
one system. In all changes, nature seeks to realize the condition of maximum
entropy ; every spontaneous change is accompanied by an increase of entropy, and
the tendency to change ceases only when the increase of entropy is the greatest
possible. If the entropy of a system were to decrease, heat would pass from a cold
to a hotter body ; the equalization of tempeiature which attends the passage of heat
from a hot to a colder body increases the entropy oi a system.® The very process
which increases the entropy of the system is attended by a loss of available energy,
so that, in this sense, an increase of entropy and a decrease of available or free energy
are correlated phenomena. The degradation of energy will be complete when the
entropy acquires a maximum value, and the free energy is zero.
Entropy and work of diffusion. — From equation (3), if a series of gases of mole-
cular weight Ml, M2 . . . which have no chemical action on one another, be con-
tained in separate compartments, respectively of volume -y^, ^2 • • •» ^^^ entropy
of the gas of molecular weight Mi contained in the first compartment will be
<f)i=Mi{Cr^ log T-\-R log Vi), and the total entropy of the gases in all the compart-
ments will be
(f>=UMC, log T+EMR log V . . . . (5)
Similarly, the entropy of an intimate mixture of a volume V, of the same gases when
<i>m=SMC,\ogT+EMR\ogV . . . (6)
The difference between these two magnitudes represents the change of entropy
which occurs on diffusion. The difference 0^— <^ with a mixture of two gases,
where V ^=Vi-\-V2, reduces to
Increase of entropy, <^«, — <^=R los . . (1)
where M=Wi4-^?«2 and F=Vi+V2. The numerator of this expression is always
greater than the denominator, and therefore the entropy of a uniform mixture of
gases is greater than the entropy of the same gases before they were mixed by
diffusion. Hence, the diffusion of gases is an irreversible process, for work will have
to be expended in separating the mixture into its constituents ; and work should
726 INOKGANIC AND THEORETICAL CHEMISTRY
be obtained if the gases are allowed to mix in suitable vessels.^ In 1875, Lord
Rayleigh showed that the work performed during the physical mixing of volumes
Vi and V2 of two different gases at the same temperature and pressure is the same as
that which would be gained during the expansion of the first gas from the volume
Vi to the volume F=ii+V2 — namely, pvi log V/vi, where p denotes the partial
pressures of the two gases of volumes Vi and V2 when V=Vi-\-V2, vide supra,
together with the work gained during the expansion of the second gas from a volume
Vg to a volume 7=^1+^2 — namely, pv2 log F/vg — when the expansions are made
in a vacuum. Consequently, the work gained by mixing two gases of volume Vi
and i'2 respectively at the constant pressures p, when Vi-\-V2=V, is
V
Work of diffusion =1) log ; . , . (8)
Thus the rule is brought under Dalton's principle that each gas behaves towards the
other as a vacuum. In all cases the gases are supposed to follow Dalton's partial
pressure law, where the total pressure is the sum of those pressures which would be
exerted by each gas in the absence of the other. The result does not depend upon
the physical nature of the gases, and there appears no reason why the argument should
not be valid for two portions of the same gas. Hence arises the so-called Gibbs'
paradox, because it follows that if the gases are chemically identical, there will be no
change of state, and no change of entropy, accordingly </>»i— <^ will be zero and
<l)^=<f). Consequently, M. Planck regarded the increase of entropy on diffusion
as a sign that the gases are chemically different, and he suggested that the
chemical difference between two substances cannot be represented by a magnitude
which varies in a continuous manner, because the magnitude varies discontinuously ;
on the other hand, the physical differences between two bodies can be represented
by a continuous function. This is taken to establish a fundamental difference
between the chemical and physical properties of a substance. ^^
The relation between the laws of maximum entropy and the degradation of
energy. — Imagine a system of three bodies, A at a temperature T^ which is higher
than T2, the temperature of the second body B ; and let C be the third body at a
temperature T, the lowest temperature of all. If a small quantity of heat dQ
performs work in passing A to C by a Carnot's cycle, the maximum available
work is {Ti—T)dQITi. Similarly, when a portion of heat dQ is transferred from
B to C, the maximum available work is {T2—T)dQIT2. By subtraction, it will
be seen that there is a decrease dE in the availability of the energy of the system
which is equivalent to
''^=<i-F>
but the increase in the entropy of the system is represented by the product of
the bracketed and succeeding term. Consequently,
dE=Tdj>
A finite quantity of heat can be so transferred by the summation of a succession of
infinitesimal instalments. This means that when heat passes from one part of an
isolated system to another part, in consequence of a difference of temperature,
there is a degradation of energy which is equal in amount to the product of the
increase of the entropy of the system, and the absolute temperature of the coldest
part. An increase of entropy thus corresponds with a decrease of the available
and an increase of the unavailable energy. For stable equilibrium, the entropy of a
system must have a maximum value.
The relation between entropy and free energy. — Under what conditions is
it possible to predict the direction in which a particular process or reaction will
proceed ? In answer, the change will take place in that direction which involves
THERMODYNAMICS AND THERMOCHEMISTRY 727
a decrease in the free energy or an increase in the entropy of the system. The so-
called intrinsic or internal energy U represents the work the system can do in virtue
of its actual condition without any supply of energy from without. For example,
when heat energy is communicated to a gas, it may separate the molecules or the
atoms further apart against molecular or atomic attractions ; it may change the
kinetic energy of molecular or atomic motion ; it may change the rotational or
vibratory energy of the atoms ; and it may change the electrical state of the mole-
cules. If a system with the total intrinsic energy XJi and entropy <f>i changes or
reacts so as to produce another system with the total intrinsic energy TJ^ and entropy
<^2> at any given temperature, T, a reaction can take place only with the expenditure
of free energy, and Ui—JJ^ must be greater than T{<j>i—<j)2). In reversible
isothermal changes, a system with a positive value of {Ux—U2)—T{(j)i~<l>2) will
tend to react ; and whether or not a reaction will take place in a particular
mixture depends on whether what H. von Helmoltz ^ called the
Free energy of a system = ( Ui — V<j^ — ^(^1 — ^2) • • (9)
Free energy of a system =(C/i — ^<^i) — {U^ — ^^2) • • (1^)
has a positive, negative, or zero value. This function U—Tcf), is sometimes symbo-
lized ip ; so that 0=01—^2 5 if »/f be negative, the system will be stable ; if i/j
has a positive value, the system will be unstable ; and if i/j be zero, the system
will undergo no change. Hence, this expression may be regarded as a stability
function — ^it is usually called a potential, or thermodynamic potential, and i/i may be
taken to represent the driving force of a reaction. The entropy itself is a test of
the stability of a system, for equilibrium is stable when the entropy has a maximum
value. As a rule, the application of the entropy test to a system is more difficult
than the free energy or the thermodynamic potential test. It was not possible to give
an exact definition of chemical affinity until thermodynamics had been developed.
The thermodynamic potential or free energy can be employed as a measure of
chemical affinity.
In ordinary thermochemical calculations, the quantities of energy absorbed and
evolved during the reaction are made to balance. The balancing is quite illusory.
The fallacy underlying the assumption that the driving force of a reaction i/fi— ^2
represents the thermal value or heat of a reaction as measured in a calorimeter,
rests on the fact that the energy T{<l)i—(f)2) which is not available for doing work
is ignored. In formulating the principle of maximum work it was tacitly assumed
that the driving force of a reaction which occurs without change of state is equiva-
lent to U1—U2 units of heat, and that if U-^ — U^ be a positive quantity, the reac-
tion is necessarily exothermal. It will be obvious that the term T(^i— (^2) ^an
be suppressed only at absolute zero, for only when T=0 will T(<j>i—<f><2) be also
zero. Consequently, the thermal value of a reaction can be a measure of the driving
energy of a reaction, only at absolute zero. At ordinary temperatures, which are,
after all, not far removed from absolute zero, the difference between {Ui—Vo)
and (Ui—U2)—T{(t)i—(f)2) is not very great, and M. Berthelot's rule is a useful
approximation — particularly in chemical technology — where a measure of chemical
energy in terms of heat energy is required.
If a reaction can take place, with the absorption of heat, the internal energy of
the system U2, after the change, will be greater than the original internal energy Ui ;
and Ui — JJ2 will have a negative value ; but in order that free energy may be
available for doing work during the change, </>£ must be so much greater than (^1
that when T{<f>i~<f>2) is subtracted from U1—U2 the driving force of the reac-
tion will still have a positive value. If the external work be involved in the change,
as will be the case when work pv is performed on or by a uniform and constant
pressure p, the term ^(^2— ^'i) must be included in the expression for the free energy
of the system, and consequently,
Freeenergy of asystem=(C7i — r<^l-f-^Vi)— (r/2— r^2+/^^2) • • (H)
728 INORGANIC AND THEORETICAL CHEMISTRY
The fundamental equation for the free energy of a system thus assume? the form
dE=dV-Td<t> (12)
meaning that in any isothermal process, or chemical reaction, the increase dE in
the free energy of the system is equal to the difference between the increase dU
in the internal energy of the system and the product of the absolute temperature,
Ty and the increase, d^, of the entropy.
J. W. Gibbs (1878) ^* in his important paper, On the equilibrium of heterogeneous sub-
stances (1875), used these functions with the object of facilitating mathematical operations.
He called \f/=U — T<f>, the force function for constant temperature; ^ = U — T<f)-\-pv, the
force function for constant temperature ; and x=^ +pv, the heat function for constant
pressure. They are often styled the psi, zeta, and chi functions of Gibbs. P. Duhem (1895)
called Gibbs' ^-function the internal thermodynamic potential ; and the ^-function, the
thermodynamic potential for constant pressure, or the thermodynamic potential or the
potential of the system.
It will be observed that while the difference {Ui—U2)—T(<j>i—<i)<2) repre-
sents the change in the free energy of the system during the isothermal change,
(C/i— 1/2) represents the change in the total intrinsic energy of the system ; the
remainder, T{(f)i—(f>2^ represents the unavailable energy which H. von Helmholtz
called the hound energy of the system. J. Swinburne 13 has further emphasized
H. von Helmholtz's conception by showing that the degradation of energy is a liability
which is incurred when any form of energy is converted into heat, for only part may
be available for work, the remaining part which is eventually and unavoidably
produced or left as heat at the lowest possible temperature is of no use ; it is waste.
Entropy furnishes a measure of this waste, for
An increase of entropy is a quantity which, when multiplied by the lowest available
temperature gives the increased waste. In other words, the increase of entropy multi-
plied by the lowest available temperature gives the energy that either has been already
irrevocably degraded into heat during the change in question, or must, at least, be degraded
into heat in bringing the working back to the standard state.
The product of the absolute temperature of a process conducted at a constant
uniform temperature T, into the difference between the entropy at the beginning
and end of the operation, viz. T{(j)i—<f)<^, represents the energy cost of the reaction,
or the amount of energy rendered unproductive during the change ; it is in some
respects analogous with the so-called latent heat of a change of state, and hence has
also been called the latent energy of a reaction. The term latent suggested for say
the heat of liquefaction by J. Black implied that the heat was still present, but latent,
dormant, or held in abeyance. The term latent energy for that portion of energy
which is degraded into a lower plane of availability during a process or reaction is
quite inappropriate and should be avoided. The loss of work- value suffered by energy
during a physical or chemical process may be due to a change in the constitution or
the state of aggregation of the body, or to a change in the thermal capacity of the
initial and final substances concerned in the reaction.
A reversible change is an idealized phenomenon which is supposed to take place
in every respect, in an opposite direction, so that when completed, the system is in
the same condition as it was at the beginning. Actually, no change takes place in
nature without incurring some waste. Consequently, a reversible change is an
imaginary cyclic phenomenon in which there is no waste of energy. If energy were
wasted during a reversible change, there would be, on reversal, a decrease of the
incurred waste, and a perpetual motion would be possible, for, in spite of friction,
etc., the available energy would not diminish. The entropy of a reversible system
remains constant, so that if there is an increase of entropy in one part, this must be
compensated by an equal decrease in another part.
Consider an indefinitely large quantity of each of two reacting gases A and B
in a state of equilibrium, to be confined in a suitable compartment, and further,
suppose a compartment containing an indefinitely large quantity of pure A, and
THERMODYNAMICS AND THERMOCHEMISTRY 725
another compartment with pure B, each at the same concentration as it has in the
equilibrium compartment. Let the component B be transferred by a semi-permeable
membrane or other reversible method from the equilibrium compartment to its own
special B-compartment, and replenish the equilibrium compartment with one equiva-
lent of A taken from the A-compartment. Let the two operations be carried on
simultaneously, so that the concentrations remain unchanged throughout the process.
The temperature remains uniformly constant at T. If <f>i denotes the entropy of
an equivalent mass of A and <j>2 that of B,
^1-^2= -| (13)
where Q denotes the heat absorbed per equivalent at constant pressure, and
the negative sign shows that the change of entropy is equal to the heat absorbed
divided by the temperature. This is regarded as the symbolic expression of the
second law of thermodynamics applied to chemical reactions. From equation (3),
^1=0,1 log T+R log Ci+/i ; i>o=C,2 log T+R log C2+/2. Accordingly, by
substitution in (9), and rearranging terms.
The equilibrium constant K can be used in place of the ratio of the concentrations
CijC^- The sign E is used for the summation sign for the specific heats and the
integration constants. The argument can be extended to include any number of
reacting components on the assumption that the specific heats of all the gases taking
part in the reaction are independent of temperature, and that the gases follow the
ideal law 'pv = RT, so that their total energies are independent of the pressure.
Equation (14) is the general form of the law of mass action.
Again, if equation (14) be differentiated, and G. Kirchhoff's equation dQp/dT
=~E{Cj,-\-RT) be substituted, and remembering that Qp-{-I!RT=Qr, J. H. van't
Hoff's important equation
^g K__Q,
dT RT^
showing the relation between the equilibrium constant and the thermal value of the
reaction at constant volume, is obtained — vide infra.
References.
1 W. J. M. Rankine, Phil Trans., 144. 115, 1854 ; R. Clausius, Pogg. Ann., 125. 353, 1865.
^ J. Perry, Th^ Steam Engine, London, 343, 1904 ; J. Parker, Elementary Thermodynamics,
Cambridge, 1891.
3 W. Nernst, Gott. Nachr., 1, 1906 ; Experimental and Theoretical Applications of Thermo-
dynamics to Chemistry, London, 1907.
4 P. Debye, Ann. Physik, (4), 39. 789, 1912 ; G. N. Lewis and G. E. Gibson, Journ. Amer.
Chem. Soc, 39. 2554, 1917 ; S. P^gliani, Atti Accad. Lincei, (5), 24. i, 835, 1915 ; Nuovo Cimento,
(6), 10. 5, 1915.
5 G. Tammann, Zeit. phys. Chem., 85. 273, 1913.
* G. Zeuner, Grundzilge der mechanischen W drmetheorie, Leipzig, 1865 ; London, 1. 45, 1907 ;
J. E. Trevor, Journ. Phys. Chem., 3. 339, 1899 ; 4. 514, 529, 1900.
' L. Boltzmann, Sitzber. Akad. Wien, 66. 275, 1872 ; W. Thomson (Lord Kelvin), Phil. Mag.,
(5), 33. 291, 1892 ; Proc. Roy. Soc. Edin., 8. 325, 1874 ; M. Planck, Vorlesuiigen iiber die Theorie
der Wdrmestrahlung, Leipzig, 135, 1906 ; J. F. Klein, Physical Significance of Entropy, New York,
1910 ; J. H. Jeans, The Dynamical Theory of Gases, Cambridge, 83, 1916 ; A. Einstein, Ann.
Physik, (4), 33. 1276, 1910 ; L. S. Omstein, Proc. Akad. Amsterdam, 15. 840, 1918 ; J. W. Gibbs,
Elementary Principles in Statistical Mechanics, New York, 165, 1902.
8 Lord Kelvin (W. Thomson), Phil. Mag., (4), 4. 256, 1852 ; R. Clausius, Pogg. Ann., 93.
481, 1854 ; M. M. Carver, Journ. Phys. Chem., 15. 613, 191 1.
9 Lord Rayleigh, Phil. Mag., (4), 49. 311, 1875 ; J. W. Gibbs, Scientific Papers, London, 1.
166, 1906; Trans. Connecticut Acad., 3. 218, 1876 ; M. Planck, Vorlesungen Uber Thermodynamik.
730 INORGANIC AND THEORETICAL CHEMISTRY
Leipzig, 203, 1897; London, 2. 4, 1903; Wied. Ann., 31. 189, 1887; 32. 462, 1887?
G. H. Bryan, Thermodynamics, Leipzig, 121, 1907.
^° N. Schiller, Ludwig Boltzm^nn's Festschrift, Leipzig, 350, 1904 ; P. Langevin, Ann. Chim.
Phys., (8), 5. 245, 1905.
11 H. von Helmholtz, Sitzher. Akad. Berlin, 22, 825, 1882 ; 647, 1883 ; OstwaWs Klassiker,
124, 1902 ; Physical Memoirs, 1. 43, 1891.
12 J. W. Gibbs, Amer. Journ. Science, (3), 16. 441, 1878; Trans. Connecticut Acad., 3. 108,
343, 1876-8 ; Scientific Papers, London, 1. 33, 1906 ; P. Duhem, Traits iUmentaire de mecanique
chimique, Paris, 1. 90, 1897 ; Le potential thermodynamique et ses applications, Paris, 1895 ;
J. J. van Laar, Sechs Vortrdge vber das thermodynamische Potential, Braunschweig, 1906.
1' J. Swinburne, Eniropy, London, 8, 1904.
§ 10. The Work done by Afl&nity during a Chemical Reaction
The doctrine of chemical affinity is unquestionably the great and distinguishing principle
of the science of chemistry as the laws of motion are of mechanical philosophy.- — J. Black
(1803).
Every chemical change performs work which is equivalent to a certain amount of
mechanical energy.. — W. Stillie (1865).
Chemical action usually produces changes in the state of aggregation or density
of the reacting substances, and so performs work as well as produces heat. As indi-
cated above, H. von Helmholtz (1882) emphasized the necessity for distinguishing
between that part, dE, of chemical energy, which can do work (free energy) from
that part, Td^, which is degraded solely as heat (bound or latent energy), since,
as previously indicated, the total chemical energy dU=dE-\-Td<j). Every
spontaneous reaction involves a decrease in its free energy, and an increase
in the bound energy. It is the decrease in the free energy, and not the develop-
ment of heat, which determines the direction of a chemical reaction. Hence, the
decrease in the free energy of a system is a measure of the work which can
be done by chemical affinity during a chemical reaction ; and is equivalent to
the maximum work gained when the process is conducted reversibly at a constant
temperature. ^
J. H. van't HofE, in his Etudes de dynamique chimique (Amsterdam, 1884), con-
siders that the magnitude of chemical affinity is equivalent to that of the work the
reaction can do when it is carried out in such a manner that the driving forces are
always balanced by equal and opposite external forces ; otherwise expressed, the
affinity of a chemical reaction can be measured in terms of the amount of work
the reaction can do when it is carried out at a constant temperature in a reversible
manner. The work in maximo which can be obtained in a chemical change is
closely related to the free energy of the reaction ; this work cannot always be deter-
mined by direct measurement, although in the case of gases, and also in the case of
dilute solutions when the gas laws are applicable, it can often be calculated from
(i) The change in the vapour pressure, or, in the case of dilute solutions, the osmotic
pressure ; (ii) The chemical equilibrium constants ; or (iii) The electromotive force
of the reaction; for example, the chemical affinity of the reaction H2+Cl2=2HCl
can be determined by measuring the electromotive force of a cell whose electrodes
are chlorine and hydrogen' gases, with hydrochloric acid as electrolyte.
When a gas at a pressure j>i changes its volume isothermally at T°, so as to corre-
spond with a pressure ^2> ^^^ work, W^, done by the gas is Wv=^RT log {jhlv^)^ ^^
previously indicated, provided that Boyle's and Charles' laws are applicable. Since
the volume of a gas varies inversely as its concentration C, usually expressed in gram-
molecules per litre, it follows that j)=CRTj and hence, the free energy Wv involved
in changing a gas from a pressure pi to a pressure p2> at constant volume, is
W,=RT log^ ; or, W,=RT log ^' ; or, W,=RT log K . (1)
for gaseous reactions of the type A^B, when kiCi--=k2C2, where ki and k2 are
THERMODYNAMICS AND THERMOCHEMISTRY 731
constants such that CilC2=K=kilk2. The equations, therefore, can be used to
calculate the maximum work or the free energy of physical or chemical processes
which pass isothermally from A with an initial concentration Ci or vapour pressure
Pi to a state of equilibrium with B which has a concentration C2 or vapour pressure
j)2, without doing any external work, that is, the volume is supposed to be constant.
These relations are generally applicable to reactions in gaseous systems or in dilute
solutions at constant volume and temperature. One of the simplest illustrations
occurs during the transformation of rhombic sulphur — vapour pressure pi — to mono-
clinic sulphur — vapour pressure ^2-
Example.' — The affinity of a gram-molecule of water for a gram-molecule of an aqueous
solution of sulphuric acid, H2SO4.H2O, which has a vapour pressure 0"0184 cm. of mercury,
at 25°, is given by the first of equations (1), when the vapour pressure of the acid with the
addition of another gram-molecule of water, H2SO4.2H2O, is 0*1125 cm. at the same
temperature. Hence, Wv=RT log {pjpi), or Tl'i, = 2 x 298 X 2-3 X log (O'l 125/0-0184), or
1371 Xlog 6-135, or lOSOcals. The factor 2-3 transforms natural into common logarithms.
The observed heat of the reaction, according to J. Thomsen (1870), is 1874 cals.
It was shown by J. H. van't Hoff, in his Etudes de dynamique chimique (Amster-
dam, 1884), that if hydrogen and oxygen, of the concentrations C^^ and Cq^^ pass
isothermally and reversibly into water vapour when the initial concentration of the
water vapour is CHgOj and the equilibrium concentrations of these three gases are
respectively cHj, cq^,, and CHjO at the absolute temperature T, the maximum work
If is a measure of the affinity of the reaction 2H2+02->2H20 per gram-molecule
of oxygen or per two gram-molecules of hydrogen, where
Tf=/JT(log5«££i+log|5^) . . . (2)
By Guldberg and Waage's law, for equiUbrium, c^B„o=Kcji2Co^j and the last equation
can be reduced to
=«r(.og/f-iog^^a^^) .... (3)
or generally
W=RT log K~RTUn log C .... (4)
where C refers to the concentration of the saturated vapour of each constituent ;
and the term UnlogC refers to A'*! log CB^+iVa log Cb2+ . . .— % log Ca^
— 7^2 log Ca2— in the chemical equation WiAj— W2A2+ . . .=NiBi+N2B2+ . . !
If the concentrations of the initial and the final products of the reaction are arbitrarily
made unity, the affinity of the reaction is simply expressed by the relation
W=RT log K . . . . . (5)
The affinity of the process 2H2+02->2H20 per gram-molecule of hydrogen is
therefore JTf .
Examples.— (1) The equilibrium constant in the reaction Hj-f I2 = 2HI at 300° is
nearly 80. What is the affinity of the reaction Ha4-l2->2HI at this temperature when the
initial concentration of the three components of the reaction is in each case unity ? From
(4), 2x573x2-3 logio 80=4631 cals. nearly. The thermal value of the reaction is
negative — 6000 cals.' — so that there is a marked difference between the free energy and
the thermal value of the process.
(2) In the reaction 2H2 + 02 = 2H20 at 727°, the equilibrium constant is l-25xlO-2o.
Hence, the free energy of the reaction 2H2^-02->2H20 at this temperature, when the initial
concentration of the hydrogen, oxygen, and water vapour are imity, is —2 X 1000 X 2-3 log 1-25
X 10-20, oj, 90,600 cals. nearly.
(3) In the reaction CaO+H20->Ca(OH)2, if p^ be the initial and p the equilibrium
pressure of the steam at a temperature T, the affinity of the reaction is W^RT log {Po/p) ;
and if the initial pressure of the steam be one atmosphere, W=—RT log p. Except at a
very high temperature, p is less than unity, so that W will be positive, meaning that steam
732 INORGANIC AND THEORETICAL CHEMISTRY
will unite with the lime, and if W be negative, the hydroxide will dissociate. At 25°, the
dissociation pressure of calcium hydroxide is 9 x 10~^^ atm., and therefore, the affinity of
the water vapour for lime is 1F = 2 x 2-3 x 298 xlog (9 X lO'^^) or -2 x 2-3 X 298 X —13-95
= 19,120 cals. Similar remarks apply to the reaction CaO + COa— ^-CaCOg.
(4) What is the affinity of iron for oxygen under the partial pressure of oxygen in the
atmosphere when the dissociation pressure of ferrous oxide at 1000° abs. is 3*1 x 10"^® mm. ?
The equilibrium constant K for the reaction 2Fe + Oj = 2FeO is K = l/p, where p is the
dissociation pressure of the gas from the ferrous oxide. The oxygen of the atmosphere is
under the partial pressure of one-fifth of an atmosphere, and a pressure of 3-1 x 10~^* mm.
is 3-lxl0-i«/760=4-lxl0-" atm. Hence, W=RTilog^-log p), or 23x2x1000
(logio 0-2-log 4-1x10-21), or 4600 (1-3010-21-6128), or 4600 x 19-6882, or 90,000 cals.
nearly. H.le Chatelier's number for the heat of the reaction of two gram-molecules of iron
with a gram-molecule of oxygen is 129,900 cals.
It is sometimes possible to calculate the free energy of a reaction indirectly as
in the case of heats of reactions. Thus, the equilibrium constant in the reaction
2CO=2C+02 is very small even at high temperatures ; the free energy of the
reaction at 1000° abs. can be computed from the free energy of the reaction CO
+ J02->C02, which is 47,200 cals. — and the free energy of the reaction 2CO->C+C02,
which is —610 cals. Subtracting the latter algebraically from the former, and
CO+JO2-2CO->CO2-C-CO2+(47,200+610) cals., orC+J02->CO-47,810 cals.;
or the free energy of the reaction 2CO->2C+02 is —95620 cals.
Examples.— (1) If the free energy of the reaction 2CO = 2C + 02 at 1000° is —95,620
cals., and of the reaction 2FeO = 2Fe + 02, —95,400 cals., show that the free energy of the
reaction FeO+C->Fe+CO is 110 cals.
(2) According to G. Bodlander,^ assuming that the thermal values of the reactions are
independent of the temperature, the free energy of formation of zinc oxide is Zn + ^Og
=ZnO + (85,800 — 30-8r + 2-29T log po) cals., where p^ denotes the partial pressure of the
oxygen expressed in atmospheres. Similarly, the free energy of formation of water is
B.^-{-^0^ = B.fi-{- {51600 -22'4T-\-2-29T log (p^^pjp^^)] cals., where p^ denotes the partial
pressure of the oxygen, p^ that of the hydrogen, and p^ that of the water expressed in
atmospheres. By subtraction of the first from the second equation, the free energy of
the reaction, ZnO+H2=Zn+H20, at 1000° C, or 1273° K., will be 57,500-85,800
— (30-8- 22-4)1273 + 2-29xl273xlog plp^^ cals. This shows that zinc oxide is reduced
at 1000° by hydrogen at atmospheric pressures only when the partial pressure of the water
vapour does not exceed ^2=0-001 atm. or 0-76 mm., because only under these conditions
is free energy available for the reaction. Analogous results are obtained with the reduction
of ferrous oxide by hydrogen, for the free energy of the formation of ferrous oxide is
Fe + |02=FeO + (64600 -25-9?^ + 2-29r log po) cals.
Refbeences.
1 G. Bodlander, Zeit. Elektrochem., 8. 833, 1902.
§ 11. The Effect of Temperature on Chemical Equilibria
Le probleme de I'afiinite est le probleme central de la chimie.- — S. Arrhenius.
There are four important stages in the evolution of the modern idea of chemical
affinity : (i) The growth of the concept connotated by the term affinity ; (ii) The
discovery of the law of mass action as a result of the work of C. F. Wenzel (1777),
C. L. BerthoUet (1803), L. Wilhelmy (1850), A. V. Harcourt and W. Esson (1866),
C. M. Guldberg and P. Waage (1867), etc. ; (iii) The recognition of the significant
part played hy free energy in determining the character and state of chemical pheno-
mena as a result of the work of J. W. Gibbs (1876), H. von Helmholtz (1882), J. H.
van'tHofi (1884-7), W. Ostwald (1892), etc.; and (iv) The effect of temperature
on chemical equilibria as a result of the work of J. H. van't Hoff (1884-7), H. le
Chatelier (1884), F. Haber (1905), W. Nernst (1906), etc.
The two great principles of thermodynamics- — embodied in the statements that
all the changes which take place in an isolated system produce a decrease in the free
energy (second law), and leave the amount of energy unchanged (first law) — have
THERMODYNAMICS AND THEEMOCHEMISTRY 733
furnished Helmholtz's equation, W—Qp=T.dWldT, which shows a relation between
the affinity, W, the thermal value, Qp, and the absolute temperature, T, of a
chemical process. With the aid of the calculus, this equation can be written :
where I is the so-called integration constant. The last equation embodies one solu-
tion of the main problem of chemical affinity, namely, to predict what will occur if
a number of substances are mixed together under given conditions of temperature,
pressure, electromotive force, etc. The affinity of a chemical reaction can be calcu-
lated for any temperature T when (i) the relation between the thermal value of the
reaction and the temperature is known, and when (ii) the integration constant /
has been evaluated. Since W=RT log K^ it follows, substituting for W,
,4(logi^)=-^,;or,logi^ = -/;;g,.r+/ . . (2)
The first of equations (2) can be written in the alternative form
lfdK\ Qp
dT\K)~ RT^
which shows that the fractional change, dK/K, in the value of the equilibrium
constant, per degree, dT, (i) is proportional to the heat Qp of the reaction ; and (ii)
inversely proportional to the square of the absolute temperature. The equation also
shows that a rise of temperature, dT, will displace the equilibrium conditions in the
same direction as the reaction which absorbs heat. For endothermal reactions
where Qp is negative, the equilibrium constant increases with rise of temperature ;
and conversely, for exothermal reactions, where Qp is positive, the equilibrium
constant decreases with rise of temperature. This is the principle of reversibility
previously discussed. The above result was obtained by J. H. van't Hoff in 1887,
and has been called the equation of the reaction isochore (to-os, equal, x^P^^^ place)
by W. Nernst (1889), because the volume is kept constant during the change.
J. H. van't Hoff's equation, (2), also represents the influence of temperature on the
system kept at a constant volume ; Qp represents the heat emitted during the forma-
tion of a substance whose concentration appears in the numerator of K=CilCo.
There is a formal relation between Clapeyron-Clausius' and van't HofE's equations :
^(log P)=^2; ^(log K)=-^, . • . (3)
The difference lies in the interpretation of the symbols ; in the one case, p repre-
sents a pressure, while the corresponding K in the other case is a product of concen-
trations, but in dilute solutions and gases, concentration can be expressed in terms
of osmotic or gaseous pressure. The term A=— ^^ refers to the diminution in
the internal energy, or the heat evolved by the reaction under a constant pressure p.
If the numerical value of the integration constant / could be deduced from the
known laws of heat, the problem concerning the effect of temperature could be
solved, but the two laws of thermodynamics per se leave the problem undetermined.
Although the two laws of thermodynamics can be applied generally to a great variety
of phenomena, they fail to yield precise conclusions, appHcable to particular cases,
without the use of certain experimental data to evaluate the integration constants i
which arise because of our ignorance of the absolute values of the energy of the system
at any temperature. Nernst's theorem is an attempt to solve this problem by
assuming that at absolute zero, the entropy is zero, and this is taken as a standard of
reference. There are two cases to consider before the integration of equations
(2) can be performed :
734
INORGANIC AND THEORETICAL CHEMISTRY
/. The thermal value Qp of the reaction does not alter appreciably with changes of
temperature.'^ If the equation be integrated on the assumption that Qp is constant
over a small range of temperature, and that K^ and K^ respectively denote the
equilibrium constants at the two temperatures T^ and T^, then, since R is approxi-
mately 2,
log.„|=115Q4-i) .... (4)
By means of this equation it is possible to compute the thermal value of a reaction
which changes but little between the temperatures Tx and T^ when the equilibrium
constants for these temperatures are known, Qp may represent the heat of sublima-
tion, the heat of vaporization, the heat of solution, the heat of dissociation, as well
as the thermal value of strictly chemical reactions. The thermal values calculated
by means of this equation are in close agreement with the observed values when the
necessary conditions obtain ; this is illustrated by Table IV.
Table IV. — The Thehmal Values of Physical and Chemical Changes.
Q in
Cals.
Heat of
Calculated.
Observed.
Vaporization of water .....
10-10
10-30
Solution of boric acid in water ....
5-2
5-6
Sublimation of ammoniiun sulphide .
21-55
21-64
Combination of BaClg -I- 2H2O ....
3-82
3-83
Dissociation of nitrogen peroxide
12-90
12-50
Precipitation of silver chloride ....
15-99
15-85
When the heats of formation of a substance at any temperature are known, the per-
centage dissociation at any assigned temperature can be calculated on the assumption
that this equation is valid, since if any four of these five magnitudes are known, the
fifth can be computed.
Examples.- — (1) Calculate the heat of solution of mercuric chloride from the change of
solubility with temperature when the solubility at the absolute temperature 283° is 6-57,
and 1 1 -84 when the temperature is 323°. Substituting these numbers in the above equation,
Q is 2700 (nearly) cals. The observed value is nearly 3000.
(2) At 670°, the dissociation pressure of barium dioxide is 80 mm., and at 720°, 210 mm.
Show that the heat of the reaction, 2Ba02 = 2BaO + 02, approximates —36-1 Cals., and
that the maximum work furnished by the formation of two gram-molecules of barium
dioxide from the monoxide and oxygen at 670°, and at the same partial pressure as it
occurs in the atmosphere, is equivalent to 1203 cals.
When but one value of K at the temperature T is available, the integration of
(2) furnishes the expression :
logio K
-0-22% +7
(5)
where I is the integration constant, and Qp the thermal value of the reaction at
constant pressure, and does not vary with changes of temperature. F. E. C. SchefTer
considers that the expression log K=aT~^-\-b is sufficiently in agreement with
measurements of the equilibrium for all reactions which have been studied. If
the reaction takes place at a constant pressure, an allowance can be made for the
work done, and the thermal value of the reaction at constant volume, Qi, is related
to that at constant pressure, Qp, by the expression Qp=Q,.~£nRT, where EnRT
represents the work done by the system against this pressure, and En denotes the
difference between the simi of the molecular coefficients of the products of the reaction
THERMODYNAMICS AND THERMOCHEMISTRY 735
taken negative, and of the initial products taken positive — for instance, in the
reaction 2H2+02=2H20, i:n=2+l-2=l.
II. The thermal value Q of the reaction changes with variations of temperature. —
It has been assumed that Q is constant, but if Q varies with temperature, the relation
between Q and T must be known before the integration of (2) can be performed.
It is usual to represent the relation between Q and T by the empirical formula of the
type Q=QQ-\-aT-\-pT^ . . ., where Qq, a, ^ . . . are constants whose numerical
values are calculated from the observed values of Q and T, and Qq is the value of
Q when the temperature is at the absolute zero. If this expression for Q be substi-
tuted in the second of equations (1), and the integration be performed,
W=QQ+IT-aT log T-pT^-lyT^- (6)
which represents the affinity W in terms of the heat of the reaction at the tempera-
ture T.
Liquid and solid systems. — In his memoir Veher die Berechnung chemischer
Gleichgewichte aus thermischen Messungen (1906), W. Nernst ^ introduced the hypo-
thesis that in the case of condensed systems — that is, systems involving only liquid or
solid substances — the temperature coefficients of the free and total energy — viz.
dW/dT and dQ/dT — decrease to indefinitely small values as absolute zero is
approached ; and, the limiting value of
dW do
y™ = ,^=0, at absolute zero, when T=0
al a J.
Consequently, the curves showing the relation between the affinity W, or the total
energy Q, and temperature T, will coincide at absolute zero, and the equality will
be usually maintained only for a short region of temperature in the neighbourhood
of absolute zero. The measurement of Q and of W cannot be performed in the
vicinity of absolute zero, and consequently, Nernst's hypothesis — called Nemst's
heat theorem — cannot be directly verified. The hypothesis, however, has been
verified indirectly, and the results are in satisfactory accord with experiment. It
will be observed that if dQjdT be zero when T=0, the term involving a must
vanish from the expression Q=QQ-\-aT -\-fiT^-[- . . ., because otherwise, at
absolute zero, when T=0, dQldT=a. The differential coefficient, dQ/dT,
represents the specific heat of a gas, and consequently W. Nernst's assumption also
includes the assumption that the specific heat of a gas is zero at absolute zero.'* If
very high temperatures be not under consideration, the higher powers of T can be
neglected because their numerical coefficients are very small, and
g=Oo+i8T2 (7)
For similar reasons, if dW/dT be zero, I and a in the expansion (6) must vanish,
and
W=Qo-pT^ (8)
Consequently, for systems involving only liquids and solids the integration constant
is zero, in agreement with results previously obtained by G. N. Lewis (1899),^ and
others. Again, for condensed systems, the nmnerical value of the affinity of a reaction
can be computed from heat measurements alone. Experiments show that for many
reactions, where the evolution of heat is great, the coefficients j8, y . . . are very
small, and W and Q have nearly the same value Qq ; and in these cases, M. Berthelot's
rule — the principle of maximum work — will apply.
According to G. Kirchhoff's equation, the variation of Q with temperature, viz.
dQjdT, is equal to the difference in the specific heats, or rather the thermal capacities
at constant volume of the initial (Cj) and final (C2) products of the reaction ; in
symbols, dQIdT^C^—Ci ; and if y, S . . . are negligibly small, dQldT=2^T'^
=C2—Ci, which gives the value of ^ ; hence, if Q be determined for any
736
INORGANIC AND THEORETICAL CHEMISTRY
temperature T, Qq can be calculated from (7), and hence the free energy or affinity
W, or log K can be calculated when ^o> P> ^^^ ^ ^^^ known.
Example. — From H. V. Regnaiilt's measurements (1844) of the specific heats of rhombic
and monoclinic sulphur, the numerical value of C^ — C^ is nearly 0'1840— 0*1764=0*0076.
Consequently, 329^=0-0038, or j8=l'15 X lO"*^. The heat of transformation at 368° abso-
lute,accordingtoG.Tammann(1903),isnearly313cals. Hence,313 = (?o+l'09 x lO-^ x3682;
or (2 = 1-57 +0-00001 15^2. According to J. N. Br6nsted,« the results obtained by com-
paring the observed and calculated values of Q at different temperatures are satisfactory.
Similarly, IF = /?r log iC = 1-57-0-0000115^2. The curves obtained by plotting corre-
sponding values of Q and T and W and T are indicated in type II, Fig. 2. Here, W
decreases, and Q increases with increasing values of T, and at absolute zero the two
curves coincide.
Equations (7) and (8) show that the two curves change symmetrically for solid
and liquid systems as indicated in Fig. 2, for the coefficients j8, y . . . may be
positive or negative, and there are two possibilities for the slopes of the curves
representing the changes of W and Q with temperature T. Deviations from sym-
metry occur when the change in the specific heat of the reacting substance is not a
linear function of temperature.^ For the reaction CuS04H-H20ice->CuS04.H20,
H. Schottkys found ^=4520+0-00408^2, and If =4520~0-00408T2, and this
reaction is therefore representative of type II, Fig. 2, where the free energy W de-
creases, while the thermal value Q of the reaction increases with rise of temperature,
and at 291° K. the free energy is less than the thermal value of the reaction. Again,
for Clark's cell, Zn+Hg2S04+7H20ice->ZnS04.7H20+2Hg, W. Nernst^ finds that
^=38505-0-0017^2, and If=38505
+0-0017^2. Here, the free energy W
increases with temperature, while the
thermal value Q of the reaction decreases
as illustrated by the curve type I, Fig. 2.
At the transition point, say, in the
passage of rhombic to monoclinic sulphur,
the free energy must be zero, and W=0.
Consequently, the transition temperature
T={Qo/B)i from (7) and (8). The free
energy of the change of rhombic to
monoclinic sulphur is represented by the
consequentl}^ , the transition temperature is
273°) -:95°. The observed value is 94-4'
IT
1 J
Z'
: iz
a W^^ -
D^^
a W^
9 i^±
J<E-=#
/. ■g^-'^ Type//.
' I ^^ -4-
^ 0^ -
^ ^^
^^-
^^
S.
S
0
Fig.
Temperature 0 Temperature
— Variations of Q and W with
Temperature.
relation F=l '57 -0*00001 15^2
approximately 368° K., or (368°-
Gaseous systems.— The preceding discussion refers to solid or liquid systems,
and it has been extended to gaseous systems when the molecular heat of the reaction
at one temperature and the molecular heats of the gaseous substances at a few other
temperatures are known. If Q denotes the heat of the reaction, it has been shown
that Q==Qo+ar+^J2^yj3_|_ , , ^ and this value of Q may be substituted in
van't Hofi's equation (2). On integration,
RT log K.^Q^J^a log T-f j3T+iyr2+ . . . +Z . . (9)
To evaluate the integration constant 7, turn to Clapeyron-Clausius' equation,
^=JtT^{d log p)ldT, a solution of which is possible when A can be represented as
a function of the temperature. If A=— Ao+a'r+j8T2-f- . . ., where Aq represents
the internal molecular heat near absolute zero, and a', j8' . . . are numerical con-
stants. It follows, after the substitution and integration, that
£n log C=-^^(^~a' log T~P'T-~T^- . . .)Unr . (10)
R\T
:go-i3"T2-Jy"J3.
■P'T-'^T^- . . .y
Substituting W=^Qo~P''T^-iy'T^- . . ., and the above value for 27% log C,
in van't HofE's equation W=RT log K—RTnlogC, the result can be reduced to
the form
RT log K={Qo'-i:nXo)-\-i:na'T log T-{P"-Zn^')I^
+RTZnr
THERMODYNAMICS AND THERMOCHEMISTRY 737
This deduction from Clapeyron-Clausius' equation has a formal analogy with the
deduction (9) from van't HofE's equation, and both refer to the same quantity
RT log K. Assuming that the equations are identities, the coefficients of like
powers of T in both equations can be equated each to each : Qq^^Qq—Eti^q ;
■~a=I!na ; — j3=— j8"+i7wj8' ; and J=I!nr. Hence, the integration constant
of the equations for the vapour pressures of the reacting components can be
used to evaluate the constant Z of a given reaction, for the revised equation (9) may
be written
It will be remembered that if,., the equilibrium constant of the chemical equation
HiAi-^-n^A^-^- . . . =iVx5i4-iV252+ . . ., is equivalent to
when the partial pressures are used. From Boyle's law^ p=CRT it follows
that P]^'=C^l(RT)^i, etc., and consequently, Kj,=K,(RT)En, or log i^p=log K^
-ySn log R-\-I!n log T. Substituting for log K^ in equation (10) ; using common
instead of natural logarithms ; and
logio Kv-^T^fj,- ~:b- log 2'-4.57^- • • • 2^3 (12)
The constant En{J.-\-\og R)I2'3 is represented by UnC, where C is called by
W. Nernst the chemical constant of the substance. When the different terms of the
vapour pressure equation (10) have been introduced, the constant /' for any given
substance can be calculated, and thus C can be determined. To evaluate the chemical
constants, it is therefore necessary to have a great number of accurate measurements
extending over a wide range of temperature : (1) The thermal value of the reaction
at a given temperature ; (2) the equiUbrium constant of the same reaction at the
given temperature ; and (3) the specific heats of all the substances which take part in
the reaction, from the given temperature T down to absolute zero. The observation
data available are not very accurate, and a number of empirical formulae have been
recommended. For example, W. Nernst recommended C=1'33 log Tb—0-0G098Tb,
where T^ denotes the absolute boiHng point of the substance. This equation is
obtained from the empirical observation, that the chemical constant is nearly 0'14
times Trouton's constant X/T, and W. Nernst's observation that A/T— 9*5 log Tf,
-O'OOITi,.
It follows from what precedes, that the chemical constant
c='^+§-^=^Aj,-i-75ioa„r+^5^T+iogi„p . (13)
where Aq is the molecular latent heat at absolute zero ; and ^ is a constant depend-
ing on the change of specific heat with temperature, The number 1*75 was
obtained on the assumption that at the lowest temperatures, the molecular heats
of all gases are 1*5 greater at constant volume, and 3"5 greater at constant
pressure than the molecular heats of the corresponding condensation products.
"W. Nernst's later work showed that these assumptions cannot be justified, and the
formula there becomes an empiricism. i^ A. C. Egcrton deduced an expression from
X=RTHd \og p)ldT; X=Xo+SoCpdT-SoCpdT', c^=c„+«r2, or c^=c,+9a2rT/j3,
whence 9a-vT/p=aTi ; and W. Nernst and F. A. Lindemann's molecular heat
formula. His expression enables the chemical constant to be evaluated from fi^., a.
VOL. I. 3b"
738
INORGANIC AND THEORETICAL CHEMISTRY
and two values of p, or one value of p and the value of Aq.
obtains for the chemical constant of
A. C. Egerton thus
Hg
Cd
Zn
W
Mo
Pt
A
H
•633
1-42
1-49
3-5
4-4
1-5
1-65
1-68
The chemical constants for a few substances (pressure in atmospheres) indicated
in Table V have been calculated mainly by W. Nernst.
Table
v.— Numerical Values <
OF
Chemical Constants C.
C
C
C
He . . .1 -M5
I2 .
3-9
HgO
3-6
Ne .
1 -0-10
HCl
30
CCI4
31
A .
-0-35
HBr
3-2
CHCI3
3-2
Kr .
. i 0-82
HI
3-4
CH4
2-8
X .
1-10
NO
3-5
C^He
2-6
H^ .
1-6
N,0
3-3
C2H4
2-8
N, .
2-6
H2S
30
C2H2
3-2
O, .
2-8
SO2
3-3
CeHe
30
CO .
3-5
CO,
3-2
C^N,
3-4
C12 .
31
CS2
31
C2H5OH
41
Br,,
3-2
NH3
3-3
Hg
1-4
The quotient obtained by dividing W. Nernst's chemical constant by the
logarithm of the critical pressure in atm. varies from 1*5 to r8. Thus, the critical
pressure of chlorine is 92*5 atm. and log 92*5=1*971 ; hence 3*l-i-l*97=l*6. This
means that the chemical constant C of a liquid is proportional to the logarithm of the
pressure p, so that C=l*65 log^o Jp- For monatomic gases, C=— 1*62+1*5 log M,
where M. represents the molecular weight of the gas.
The methods of evaluating the chemical constants are: (1) Direct comparison
with the vapour pressure curve ; (2) The method of chemical equilibrium ; (3) Some
empirical formulae — e.g. S. Young's modification of Clapeyron-Clausius' equation,
above. It must be added that the evaluation of the chemical constant is the
weakest part of the discussion, and the results obtained by the different methods are
not always concordant, probably owing to the inaccuracy of the available data. For
example, two equally satisfactory empirical formulse for the specific heat of a gas
may lead to widely different values for the chemical constant. Thus A. Langen
gives the chemical constants of nitric oxide, +0*92 ; oxygen. 1*021—0*539 ; carbon
monoxide, —0*04 ; nitrogen, 0*05 ; carbon dioxide, -0*406 ; water, —1*930 ; and
ammonia, —2*454. However, as B. Weinstein has emphasized, although the results
which have been obtained die Zuldssigkeit des Gleichungssy stems zweifellos feststellen,
yet the uncertainty in the numerical values of the constants can be removed only
by observations extending over wide ranges of temperature.
Remembering that a=—Zna, where a' is a constant in the vapour pressure
curve, and therefore a—2JnR=—Un(a'-}-R), W. Nernst puts a'+72=3*5 cals. as
a first approximation, so that a—I!nR=—Un3'5 ; or (a--IJnR)IR~Unl'7b. Sub-
stituting these results in equation (12), and
logioKj,=§l^+^nl'76 log T-^
_ ^ j_
•57
. . +UnC
(14)
the values of a, j3, y, . . . can be determined from the relation dQ/dT^a-{-2pT
+37T2+ . . . from G. Kirchhoff's equation, where dQ/dT is equal to the difference
between the molecular thermal capacity of the initial and end products of the
reaction ; ^0 i» evaluated from the relation Q=QQ-\-aT -]-pT^-{- . . . For gaseous
reactions in which the initial and final products occupy the same volume, Un is zero,
and Unlld log T is then a zero term, but neither £nC nor jST is necessarily zero.
THERMODYNAMICS AND THERMOCHEMISTRY 739
These relations, obtained by the application of W. Nernst's theorem to
J. H. van't HofE's equation, enable the equilibrium constant of a gaseous reac-
tion, and consequently also the free energy or affinity of a reaction, to be calculated
from three sets of data : (i) The thermal value of the reaction at the temperature T ;
(ii) The molecular heats of the reacting substance over the range of temperature in
question ; and (iii) The chemical constants of the reacting substances. As H. le
Chatelier ii stated in 1888, the chemical constants are definite functions of certain
physical properties of the reacting substances. Consequently, the indeterminate
integration constant in J. H. van't Hoff's equation can also be expressed as a sum of
constants which are characteristic of each reacting substance. This can also be done
quite independently of W. Nernst's heat theorem (1906), as was demonstrated by
M. Planck (1897) and by F. Haber (1905).i2 Another important feature of this
investigation is that it enables the integration constant I to be calculated from the
characteristic function Unl', that is, the chemical constants of the substances con-
cerned in the gaseous reactions can be computed without making any observations
on the reaction itself. As H. le Chatelier predicted in 1888, the evaluation of the
nature of this function will lead to a complete knowledge of the laws of chemical
equilibrium, and it will enable chemists to determine the conditions of equilibrium
of a given chemical reaction, a priori, and independently of new experimental data.
References.
1 J. W. Mellor, Higher Mathematics for Students of Chemistry and Physics, London, 194, 1913.
2 F. E. C. Scheffer, Proc. Acad. Amsterdam, 19. 636, 1917.
3 W. Nernst, Gott. Nachr., 1, 1906 ; Sitzber. Akad. Berlin, 52, 1906 ; Theoretische Chemie,
Stuttgart, 1916 ; Experimental and Theoretical Applications of Thermodynamics to Chemistry,
London, 1907 ; F. Pollitzer, Die Berechnung chemischer Affinitdten nach dem Nernstschen Wdrme-
theorem, Stuttgart, 1912; 0. Sackur, Ann. Physik, (4), 31. 455, 1911; Lehrbuch der Thermo-
chemie und Thermodynamik, Berlin, 1912 ; London, 1917 ; Die chemische Affinitdt und ihre
Messung, Braunschweig, 1908 ; F. Jiittner, Zeit. Elektrochem., 17. 139, 1911 ; I. W. Cederberg,
Die thermodynamische Berechnung chemischer Affinitdten von homogenen und heterogenen GaS'
reaktionen, Berlin, 1916.
* F. Juttner, Phys. Zeit., 8. 147, 1907.
^ G. N. Lewis, Proc. Amer. Acad., 35. 3, 1899 ; Journ. Amer. Chem. Soc, 35. 1, 1913 ; Zeit.
phys. Chem., 32. 364, 1900 ; T. W. Richards, ib., 42. 129, 1902 ; J. H. van't Hoff, Festschrift
Ludwig Boltzmann, Leipzig, 233, 1904.
8 J. N. Bronsted, Zeit. phys. Chem., 55. 371, 1906.
' J. N. Bronsted, Zeit. phys. Chem., 56. 645, 1906 ; J. H van't Hoif, Festschrift Ludwig
Boltzmann, Leipzig, 223, 1904 ; P. Debye, Ann. Physik, (4), 39. 752, 1913.
8 H. Schottky, Zeit. phys. Chem., 64. 415, 1908.
» W. Nemst, Theoretische Chemie, Stuttgart, 728, 1907.
10 W. Nemst, Zeit. Elektrochem., 20. 185, 1914; W. D. Treadwell, ib. 23. 270, 1917;
O. Sackur, Aim. Physik,{4), 40. 67, 1913 ; H. Tetrode, ib., (4), 38. 434, 1912; 0. Stern, ib., (4),
44. 123, 1914 ; Phys. Zeit., 14. 629, 1913 ; A. Langen, Zeit. Elektrocliem., 25. 25, 1919 ; W. Nemst,
Ver. deut. phys. Ges., 11. 247, 313, 1909; 12. 565, 1910; Recent Applications of Thermodynamics
to Chemistry, London, 1913 ; F. Haber, Thermodynamik technischer Gasrealdionen. Miinchen,
1905 ; A. B. Lamb's trans., London, 88, 1908 ; A. March, Phys. Zeit., 18. 53, 1917 ; B.Weinstein,
Thermodynamik der Kinetik der Korper, Braunschweig, 3. 1059, 1905 ; M. Planck, Vorlesungen
uber Thermodynamik, Lcijizig, 275, J 911; K. J ellinek, Physikalische Chemie der homogenen und
heterogenen Gasreaktionen, Leipzig, 1913; S. Young, Phil Mag., (5), 34. 505, 1892; A. C. Egerton,
ib., (4), 39. 1, 1920 ; F. A. Lindemann, ib., (4), 39. 21, 1920.
1^ H. le Chat«her, Recherches exptrimentales et theoriques sur les equilibres chimiques, Paris,
184, 1888 ; Ann. Mines, (8), 13. 157, 1888.
12 M. Planck, Vorlesungen liber Thermodyrmmik, Leipzig, 205, 1897; London, 214, 1903;
F. Haber, Thermodyruimik techniscJier Gasreaktionen, Miinchen, 38, 1905; London, 38, 1908;
O. Sackur, Lehrbuch der Thermochemie und Thermodynamik, Berhn, 235, 1912; London, 307,
1917.
CHAPTER XIII
THE KINETIC THEORY OF ATOMS AND MOLECULES
§ 1. The Molecular Theory of Matter
If wo would become imbued with the spirit of the new philosophy of chemitstry, we must
begin by believing in molecules.- — J. P. Cooke.
For purely chemical reasons, which culminated in Avogadro's hypothesis, chemists
have invested matter with an imaginary structure which explains, very well, the
various transformations which matter undergoes. Matter — which in bulk appears
to the eye continuous and perfectly uniform in all its properties and parts — is
supposed to be made up of extremely small discrete particles called molecules.
Molecules are the imaginary units which make up matter en masse. Molecules
are made up of one or more atoms. Atoms are the imaginary units which make
up individual molecules.
Molecular structure of matter. — Matter must be either a discrete or a continuous
medium. The phenomena which attend diffusion in solids, liquids, and gases
lead us to reject the hypothesis that matter is continuous, for how can two continuous
media occupy the same period of time ? A study of the compressibility of gases —
Boyle's law — leads to the same view. There appears to be no limit to the expansion
or dilatability of a gas ; and therefore, says A. W. Eiicker (1901), it is inconceivable
that any real substance or thing which can at the same time be present in every
part of a given space would also be present in every part of a space a million times
as great. How can a continuous medium on rarefaction (that is, diminution of
pressure) expand indefinitely ? How can compression diminish the volume of
matter itself ? If matter be discrete, we can readily answer these queries. Com-
pression involves a closer packing or a crowding together of the molecules by
diminishing the space between them. This very explanation was given by Hero
of Alexandria i circa 177 B.C.
The particles of air do not fit closely in every part, but void spaces are left between them
as in the sand on the seashore ; the grains of sand must be imagined to correspond to the
particles of air, and the air between the grains of sand to the void spaces between the
particles of air. Hence, when any force is applied to air, the air is compressed, and, contraiy
to its nature, falls into the vacant spaces from the pressure exerted on its particles ; but
when the force is withdrawn, the air returns again to its former position from the elasticity
of its particles, as is the case with horn shavings and sponge, which, when compressed and
set free again, return to the same position, and exhibit the same bulk.
Conversely, rarefaction involves an increase of the space between the molecules,
so that the molecules become less closely packed and less crowded together. If
matter be discrete we can also understand how one substance can diffuse into
another — hydrogen into air ; and anihne dye into water. There are also numerous
examples of the diffusion of one soHd metal into another, the penetration of solid
metals by gases, etc., which show that solid metals are porous to certain elements.
Mercury will pass through a piece of tin, a centimetre thick, in half a minute.
The porosity of metals was recognized by G. Homberg 2 in 1713, and, in his
Observations sur des 7natieres qui 'penHrent et qui tr aver sent les metaux sans lesfondre,
he gave examples of substances which will pass through the pores of the metals.
Again, the volume of a mixture of two liquids is not necessarily the same as the
740
THE KINETIC THEORY OF ATOMS AND MOLECULES 741
joint volume of the separate components, even though no chemical reaction,
recognized as such, occurs.
A mixture of 500 c.c. of alcohol and 500 c.c. of water occupies 940 cc, which is much
less than 1000 c.c, and the difference is still greater in the case of sulphuric acid and water
provided the temperature at which the volume of the solutions is measured is the same.
Conversely, a solution may occupy a greater volume than the joint volume of its separate
constituents. Thus 111-1 c.c. of j9-nitrotoluene and 100 c.c. of carbon disulphide give
not 2iri c.c. but 224-7 c.c. of the mixture- — an expansion of 13-6 c.c. Similarly, the
volumes of solutions of the ammonium halides in water are greater than the joint volumes
of salt and water.
Many other examples might be quoted which indicate that (1) one substance may-
be actually penetrated by another ; or that (2) the molecules of one substance
may be so disposed that the molecules of another substance can be accommodated
between them much as a scuttle of coal might at the sam.e time accommodate a
bucket of sand; (3) matter is compressible or expansible so that it occupies a different
volume in contact with another substance than it does alone. The molecules seem
to lead a more or less independent existence, and the space between the molecules
furnishes accommodation for the introduction of other particles.
There are many other lines of argument pointing in the same direction : If
transparent substances like glass or water were infinitely homogeneous, the velocity
of propagation of light through them would be independent of the period of vibration
or the wave-length of the ray of light. A. L. Cauchy (1836), ^ therefore, inferred
that transparent substances are not infinitely homogeneous because the velocity
of propagation of light does depend on the period of vibration ; and the coarse-
grainedness of liquids and transparent solids is comparable with the wave-length
of light.
W. Ostwald in his Grundriss der allgemeinen Chemie (Leipzig, 1904) did not
accept the interpretation of the evidence for the granular structure of matter, for
he confessed that he did not then know any facts which could not be described
without this assumption ; but in a later edition (London, 1912), he said that after
a fruitless search extending over a century, a final proof of the grained, atomistic, or
molecular nature of matter has been obtained by studying the properties of colloidal
systems, and the effects of electrical discharges in gases. A study of the physical
and the chemical properties of matter has thus led to the conclusion : Matter is
discrete not continuous ; and it is made up of minute particles called molecules.
This hypothesis is called the molecular theory of matter.
Are the molecules stationary or in motion ? Here again the phenomenon of
diffusion has led to the further assumption that the molecules of matter are in
rapid motion. How could gases diffuse one into the other in such a remarkable
way if the molecules were at rest ? Again, in Rumford's celebrated experiment
(1798) it was proved that mere friction produces heat in unlimited quantities, and
hence it was argued that it is " extremely difficult, if not quite impossible, to form
any distinct idea of anything capable of being excited and communicated in the
manner heat is excited and commimicated in this experiment, except it be motion."
Heat must be a mode of motion. Again, if heat be a mode of motion the motion is
not apparent ; it is not a motion of the body as a whole, but rather a motion of the
fundamental particles of matter. This internal motion, too, must be more rapid
the higher the temperature, a conclusion which is in harmony with the phenomenon
of diffusion. Diffusion is produced by the internal movement of the particles of
matter, and this is th-e more rapid, the higher the temperature. The fact that
gases rapidly fill a confined space, however large, has been explained by assuming
that the molecules repel one another, but, as H. Davy emphasized in his Essay ov.
heat, light, and the combinations oflight,^ in 1799, the so-called repulsive force can be
identified with the thermal oscillations of a body :
Hoat may be defined as a peculiar motion, probably a vibration of the corpuscles of
742 INORGANIC AND THEORETICAL CHEMISTRY
bodies, tending to separate them. ... To distinguish this motion from others . . . the
name repulsive motion has been adopted.
The phenomena which attend the expansion and diffusion of matter lead to the
assumptions : (i) that matter has a granular structure, for it consists of discrete
parts of molecules ; and (ii) that the molecules are in a state of incessant independent
motion, and that they are travelling to and fro in all directions. The physical
evidence here outlined is altogether independent of assumptions as to the nature
and properties of the molecules ; it can be supplemented by an enormous mass
of other facts from diverse sources ; and it is strongly supported by chemical
phenomena as interpreted by Dalton's and Avogadro's hypotheses.
According to the kinetic theory as expounded by P. Gassend in 1658, and others,
the difference between solids, liquids, and gases is due to a difference in the average
distances between the molecules, and in the mobilities of the molecules. The
physicists' definition of a molecule is : a minute portion of a substance which moves
about as a whole, so that its parts, if it has any, do not part company during the
motions.
References.
^ B. Woodcroft, The. Pneumatics of Hero of Alexandria^ London, 2, 1851.
2 G. Romberg, Mem. Acad., 306, 1739.
' A. L. Cauchy, Memoire sur la dispersion de la lumiere, Prag, 1830 ; A. W. Reinhold, B. A.
Rep., 986, 1885 ; Lord Kelvin (W. Thomson), Proc. Roy. Inst., 10. 185, 1883.
* H. Davy, Collected Works, London, 2. 20, 1839.
§ 3. The Kinetic Theory of Gases— Boyle's Law
What is must be studied before what was can be inferred. Precedent states remain
visionary unless they can be linked to actual and observable conditions.- — ^A. M. Clerke.
The molecules of a gas seem to lead to a more or less independent existence ;
and their average distance apart is much greater than with liquids or solids. The
molecules of a gas appear to be continually moving with a great velocity in approxi-
mately straight lines in all directions. The molecules spend most of their time
travelling about like missiles, without the kinetic energy of the motions predomi-
nating in any one direction. There is an interchange of energy during the collisions
of the molecules, and the immense number of collisions leads to a rapid distribution
of any excess of energy which the motions of any one molecule might possess, and
thus the pressure, etc., is rapidly equalized. The molecules in their travels are
not only colliding with one another but they are also bombarding the walls of the
containing vessel ; in consequence, the molecules are continually changing their
speeds and directions. It is clear that an outside pressure must be exerted on the
walls of the vessel every time a molecule strikes the boundary walls, but every
bombardment known to human experience involves several losses— e.g. energy,
velocity, and momentum are lost. The molecules are supposed to be perfectly
elastic so that after each collision they rebound with the same velocities as before,
otherwise, it is said, their momentum would decrease with each collision and the
pressure of the gas would decrease with time, which it does not. Gases have been
confined many months under pressure without sign of loss ; but attempts to so
detect a diminution of pressure are foredoomed to failure since any slackening the
average speed of the molecules would probably be immediately restored by collision
with the boundary walls, if the prevailing temperature determined the average
speed of the molecules.
Hence, the kinetic theory of molecules postulates : (i) The molecules are perfectly
elastic ; and (2) in spite of the law of excluded perpetual motion, we have what
H. Poincare called un eternal paradox, for it is inferred that the molecules of a gas
THE KINETIC THEORY OF ATOMS AND MOLECULES 743
are in a state of perpetual motion. A hj^othesis is weak when it is based upon
more or less arbitrary fictions, and not on something about which we have experience.
Still there are hypotheses which are strong and vigorous in spite of their explaining
the known in terms of the obscure — ad obscurum per obscurum. For instance, we
have no experience of an interstellar oBther, and yet the undulatory theory of light
has thriven on such a medium ; nor yet have we any experience of perfectly elastic
solids, and yet the kinetic theory of molecules has grown about this fiction. The
preceding assumptions suffice for some important deductions which enable the
condition of the molecules of a gas to be inferred with some degree of probability.
Boyle's law,— Assume that a closed vessel contains an indefinitely large number,
n molecules, and that the ceaseless cannonade of these molecules on the walls of
the vessel produces an average pressure p. Imagine n similar molecules to be
squeezed into the same vessel. This will double the number of impacts on the
sides of the containing vessel so that the pressure will rise from p to 2^. The
concentration of the gas will also be doubled. This is nothing but another way of
stating Boyle's law. The argument is due to Robert Hooke i in 1678. R. Hooke's
own words are :
If therefore a quantity of this body be enclosed by a soUd body, and that be so contrived
as to compress it into less room, the motion thereof (supposing the heat the same) will
continue the same, and consequently the Vibrations and Occursions will be increased in
reciprocal proportion, that is if it be condensed into half the space the Vibrations and
Occursions will be double in number. If into a quarter the Vibrations and Occursions will
be quadruple. . . . Again, if the containing vesesl be so contrived as to leave it more space,
the length of the Vibrations will be proportionally enlarged and the number of Vibrations
and Occursions will be reciprocally diminished, that is, if it be suffered to extend to twice
its former dimensions, its Variation will be twice as long, and the number of its Vibrations
and Occursions will be fewer by one half, and consequently its endeavours outward will
be also weaker by half.
The further mathematical study of a system of elastic spheres, ceaselessly moving
at different speeds in all directions is based upon the principle of averages. It
does not consider the motion of an individual molecule, but rather the average
motions of the entire system of spherical particles.
Boyle's law can then be obtained in another manner : Suppose a gas containing
n molecules, each of mass m, be confined in a cube with edges each I cm. long, and
that the molecules are moving with a mean velocity F. Although the molecules
travel about in every conceivable direction, it is fair, for purposes of calculation,
to consider the molecules are divided into three equal sets with velocities parallel
to three adjacent sides of the cube. At any instant, therefore, we assume that
}^n molecules are travelling with a mean velocity V parallel to any particular
edge, and therefore perpendicular to the two corresponding faces of the cube.
One molecule moving with a velocity V will take IjV seconds to pass from side to
side, and it will therefore strike a side ^Vjl times per second. At each collision
with the face of the cube, the velocity of the molecule is reversed in direction
that its momentum changes from mF to — mV ; that is, its momentum changes
2mF. The total change of momentum by \n molecules striking a side JF// times
per second will therefore be the product \Vjl X 2mF X \n, or JwmF^//. This measures
the total force or pressure exerted on one face of the cube. But the total surface
of one face of the cube is l^. Hence, the total pressure per unit area is p=^^nmV^ll
-L.i2—:^fi^yiy2iis^ But ^3 represents the volume v of the cube. Hence
pv=lnmV^ . . . . . (1)
The product of the pressure and volume of a gas is equal to one-third the sum of
the masses of all the molecules into the square of the mean velocity of the translatory
motions of all the molecules. For unit volume also, the average kinetic energy
of the molecules of the gas is equal to the pressure p. If the number n, the mass m,
and the mean velocity V of the molecules does not change, the expression InmV^
744 INORGANIC AND THEORETICAL CHEMISTRY
will be a constant ; and hence also the product fv will be constant. This is in accord
with Boyle's law. Since j)v=RT it follows that
RT=lnmV2; oi RT=INV^ .... (2)
where N denotes the number of gram-molecules per c.c. Dalton's law of partial
pressures follows ks a corollary, because the total pressure exerted by a mixture of
gases must be the sum of the partial pressures exerted by the individual molecules
provided they exert no physical or chemical action upon one another. By
definition, the mass M of any substance is equal to the product of the density D
into the volume v, and accordingly, the density of a mass of n molecules each of
mass m occupying a volume v will be D=nm/v. Substituting this relation in the
preceding equation, and
P=hDr^ (3)
which is sometimes called Bernoulli's equation, and which shows that the pressure of
a gas is equal to one-third the product of its density into the square of the mean
velocity of the translatory motions of the molecules.
The mean kinetic energy of the molecules is K=lnmV^, and accordingly,
equation (1) can be written ^pv=inmV^, showing that the product ^pv is equal
to the mean kinetic energy of the molecules, or
Kinetic energy =^RT (4)
since pv=RT. Hence, the pressure of unit volume of a gas is two-thirds the
kinetic energy — expressed in proper units — whatever be the temperature. This
enables the molecular energy of a gas to be expressed in terms of a magnitude
which can be measured directly. Both magnitudes have their origin in molecular
motion ; and both change proportionally with the absolute temperature, so that
T^'T'o' orT^^i^oCl+al?); and f = j^ or K=Ko(l+ae) . (5)
whereto ^^^ -^o represent the values of the pressure and kinetic energy per unit
volume at 0°, or at Tq=273 ; and p and K the corresponding values at the absolute
temperature T. It follows therefore that the kinetic energy of molecular motion
is the mechanical measure of the temperature.
The mean kinetic energy is also equal to \MV^, when M=nm. Consequently,
3RTIM=V^, and since the numerical value of R is 83*15 XlO^ ergs, the so-called
mean velocity V of the molecules is
F=15800Ay^ cm. per second . . . (6)
From (3), when the pressure is constant, the velocity F will be inversely proportional
to the square root of the density D, for V^=Jc/D, where k represents a constant ;
or Vl^/Dl=V2\/iy2' This is Graham's diffusion law. If V denotes the mean
velocity of the molecules of a gas, the average kinetic energy is \MV^, and by
Graham's law, the velocity is equal to a constant, say ^/2k, divided by \/Z), or the
mean kinetic energy is JcM/D ; but by Avogadro's rule, M=k'D, and hence the mean
kinetic energy is equal to a constant k. The temperature and pressure are supposed
to be invariable. Hence, the mean kinetic energy of the molecules of all gases
at the same temperature and pressure is the same. Since the mean kinetic energy
a gram-molecule of a gas, JMF2,.is equal to ^RT, it follows that if there are
6*062 X 1023 molecules per gram-molecule of the gas, the kinetic energy per
molecule at 0° and 760 mm. is 5-62x101* erg.
How fast do the molecules move ? — Bernoulli's equation makes possible an
THE KINETIC THEORY OF ATOMS AND MOLECULES 745
extraordinary calculation— no less than the mean velocity of the translatory
motions of the molecules. The other two magnitudes which occur in the equation
can be directly measured. J. P. Joule made this calculation in 1848 ; although
J. J. Waterston presented an analogous calculation to the Royal Society nearly
three years earlier. 2
A gram -molecule of hydrogen^ — that is, 2-016 grms.— at 0°, and imder a pressure of 760
mm., occupies very nearly 22400 c.c. The density of hydrogen is therefore very nearly
2-01 0/22400 =0-0000896. Again, a pressure of 760 mm. of mercury is equivalent to a
weight of 1033-3 grms. per sq. cm., and since a weight of one gram falling freely acquires
an acceleration of 981 cm. per second, owing to gravitational attraction, it follows that
1033-3 grms. will acquire an acceleration of 981 x 1033-3 cm. per second. By substituting
these results in Bernoulli's equation : 1033-3 X 981 =^ X 0-0000896 X V\ or F = 184100 cm.
per second.
It is not to be assumed that all the molecules of hydrogen have this particular
speed, nor that any single molecule retains this speed over any lengthened period.
Some of the molecules no doubt have a greater velocity, others a smaller velocity.
A molecule of hydrogen starting off with a velocity above the average will soon
have both its speed and direction changed by encounters with other molecules.
The velocity F under consideration represents the mean velocity, or the velocity
of mean square of the whole of the molecules of the given gas at the temperature
of melting ice. If the gas contains n molecules, and the velocities of the different
molecules are Vi, V2, . . . there must be a quantity 72such tha.t nV^=Vi^-{-V2^ +^'3^
+ • • • +'^/t^ ; this quantity V is called the velocity of mean square of the moving
particles, and V is called here the mean velocity. At 0°, therefore, the mean velocity
of the molecules of hydrogen is nearly 1'84 kilometres per second, that is, about
6100 ft., or just over a mile per second. The arithmetical mean U of the velocities
of all the molecules is rather less than the mean velocity F, such that C/ =0*921 7,
or F=l-08C7or,
Z7=14550.
M
cm. per second
(7)
The speeds of the molecular motions of other gases can be calculated in a similar
manner, or Graham's law can be used. The results for seven typical gases at 0°
and 20° are indicated in Table I.
Table I.- — The Mean Velocities of Different Gases.
Molecular weight, M.
Mean velocity V
cm. per second.
2-016
AtO°.
At 20".
Hydrogen, Hg
1-84 xlO^^
1-90x106
Oxygen, Oo .
3200
0-46
0-48
Nitrogen, Ng
28-02
0-49
0-51
Argon, A
39-88
0-41
0-43
Water, H2O .
18-016
0-61
0-64
Carbon dioxide, COg
44-00
0-39
0-41
Mercury, Hg .
200-6
0-18
0-19
M. Cantor-'' has made an experimental demonstration of the speeds of molecular motion.
When a copper plate is inserted in a vessel of chlorine, part of the molecules of the gas
which bombard the plate rebound, and part unite with the metal to form the chloride.
If p denotes the pressure of the gas on the surface of an inert substance, and p' the pressure
on the surface of the copper, p—p^ = ^wV, where w denotes the mass of chlorine absorbed
per sq. cm. per second, and V the velocity of the molecules which react with the copper,
by suitably suspending plates of glass with the right half of one face and the left half of the
opposite face coated with copper. M. Cantor was able to measure the difference of pressure
p —p' from the forward movement of the coppered faces ; w could be determined by direct
746 INORGANIC AND THEORETICAL CHEMISTRY
weighing. When m' = ^x10~* grm., and p—p' =10'7 xlO~* dynes per sq. cm., the mean
velocity of the molecules of chlorine absorbed by the copper is 48 metres per second. The
mean velocity of all the molecules at 0" is 310 metres per second.
0. E. Meyer, in his Die kitietische Theorie der Gase (Breslau, 1877), has shown
that in a gas at rest as a whole, the number of molecules which strike unit area of
the containing vessel in unit time is \nmV, or since nm=D, the density of the gas,
the number of grams of gas molecules which strike unit area in unit time is fjL=lI)U,
and since D=M/v=Mp/RT from the gas equation, it follows that ix=MpUI4:RTy
where p is expressed in bars. Substituting the value of U from (7), when R=S3'lb
X 10^ ergs per degree.
/M
^=43'7xlO— 6^/y/ grms. per sq. cm. per sec. . . (8)
A gram-molecule of gas contains 6*062 X 10^3 molecules, and therefore, the number
of molecules of gas which strikes a sq. cm. of surface per second is
iV=2-652xl0i9^>>yy .... (9)
For hydrogen, at a pressure ^=10^ bars, and 20°, jLt=13*8 grms. per sq. cm. per
second. That is, the total number of molecules striking a sq. cm. of surface is
equivalent to the number of molecules contained in 154 litres of hydrogen gas —
although, of course, the same molecules may strike the surface many times.
1. Langmuir * has applied this equation to compute the vapour pressure of
tungsten, molybdenum, and platinum.
Consider a surface of metal in equilibrium with its saturated vapour. According to
the kinetic theory, equilibrium is looked upon as a balance between the rates of vaporization
and condensation. These two processes are conceived to be going on simultaneously at
equal rates. At temperatures so low that the vapour pressure of a substance does not
exceed a millimetre, the actual rate of evaporation of a substance may be considered to be
independent of the presence of the Vapour aroimd it. That is, the rate of evaporation in
a high vacuum is the same as the rate of condensation in presence of a saturated vapour.
Similarly, the rate of condensation may be considered to be determined only by the pressure
of the vapour.
1. Langmuir therefore argues that the vapour pressure of a metal like tungsten can
be calculated, by means of equation (8), from the observed rate of evaporation or
loss of weight at constant temperature when heated in vacuum tubes. At 2800° K.,
for example, the loss of weight of a tungsten filament was observed to be
0'43xl0~* grms. per sq. cm. per second. Consequently, from (8), the vapour
pressure is 38*1 X 10-3 bar, or 28-6 XlO-^ mm. of mercury.
I. Langmuir has further compared the energy lost by convection from tungsten
wire heated in hydrogen gas with the loss calculated from the kinetic theory, and
found that the two results are in agreement at temperatures below 2100° K. ; at
higher temperatures, the observed losses are greatly in excess of the theoretical
results. The increased loss of heat is supposed to be expended in the dissociation
of the hydrogen molecules as they strike the tungsten wire. The augmented loss
does not occur below 3500° K. in nitrogen or mercury vapour. The calculated
dissociation of hydrogen H2^2H at 2500° K. is 37 per cent. ; at 3330° K. 5*3
per cent. ; and at 4500° K, the dissociation is almost complete. The number of
hydrogen molecules which strike the hot wire at very low pressures can be calculated
from (7), and the heat losses from the surface of the wire can be measured ; conse-
quently, the heat carried away by each molecule of the gas can be computed.^ At
high temperatures, the heat losses become approximately constant ; and, assuming
that every hydrogen molecule which strikes the filament is then dissociated, it is
possible to calculate a lower limit for the heat of the reaction. In this way, it was
found 2H->H2+130 Cals. at constant volume and 3000° K. ; and similarly,
2H->H2+136 Cals. at constant pressure.
THE KINETIC THEOKY OF ATOMS AND MOLECULES 747
Again, I. Langmuir found that tungsten wires heated in oxygen gas are
chemically attacked, and the tungstic oxide so formed volatilizes and condenses on
the walls of the bulb, and leaves the surface of the filament clean. Equation (6)
again enables the rate at which the oxygen molecules strike the wire to be calculated,
and on comparing the result with the rate at which the filament is actually attacked,
I. Langmuir found that only a fraction of the total number of molecules of oxygen
which strike the filament are chemically fruitful. About 0"033 per cent, of the oxygen
molecules which strike the filament react chemically with the tungsten at 800°,
and 1 5 per cent, at 2500°. At this latter temperature, therefore, only one molecule
in every seven of those which strike the tungsten actually form tungstic oxide, WO3.
References.
* R. Hooke, Lectures de poientia restitutione or of Spring, London, 1678.
2 J. J. Waterston, Phil. Trans., 183. A, 1, 1892 ; J. P. Joule, Manchester Memoirs, 9. 107,
1848 ; Phil. Mag., (4), 14. 211, 1857 ; R. Clausius, ib., (4), 14. 108, 1857 ; Pogg. Ann., 100. 376,
1867.
3 M. Cantor, Wied. Ann., 62. 482, 1897 ; Zeit. phys. Chem., 26. 568, 1898.
* I. Langmuir, Phys. Rev., 34. 401, 1912 ; (2), 2. 329, 1913 ; Journ. Amer. Chem. Soc, 34.
261, 1912 ; 35. 107, 1913 ; 37. 417, 1139, 1915 ; I. Langmuir and C. M. J. Mackay, ib., 36. 1708,
1914.
6 W. Nemst, Boltzmann's Festschrift, 904, 1904.
§ 3. The Kinetic Theory of Gases— Charles' Law and Avogadro's
Hypothesis
Nihil enim viveret sine calore.— L. A. Seneca (c. 50).
Heat is motion which in its strife acts upon the smallest particles of bodies. — Francis
Bacon (1620).
That all kinds of fiery burning bodies have their parts in motion ; that heat argues a
motion of the internal parts, and that in all extremely hot shining bodies, there is a very
quick motion that causes light, will be generally granted.— Robert Hooke (1665).
Heat consists in a minute vibratory motion of the particles of bodies. — Isaac Newton
(1675).
Heat is considered by men of science to be a mode of motion of the molecules
of matter ; and the texts at the head of this section show that the idea is very
old. The dynamical theory of heat was advanced as an hypothesis by F. Bacon,
I. Newton, and R. Boyle, and established by the experimental work of Count
Bumford, H. Davy, J. P. Joule, etc.
The kinetic theory and Charles' law. — The average speed of molecular motion
is assumed to determine the temperature. Each change of temperature, how-
ever slight, is supposed to be attended by a corresponding change in the average
speed of the moving molecules. From Boyle's and Charles' laws, pv=RT ; when
V is the mean velocity of the molecules of a gas, fv^^nmV^, or RT=\nmV'^;
and since R, n, and m are constant, it follows that V^^^hT, where ^ is a constant ;
otherwise expressed, for a given gas at a given pressure, the mean molecular
velocity F^ depends only upon the one variable — temperature. J. J. Waterston i
first developed this hypothesis in 1845 — the mean kinetic energy of the mole-
cules of a gas is proportional to the temperature — Waterston's hypothesis.
The same postulate was made independently by R. Clausius, and shown to be
a necessary consequence of J. C. Maxwell's law of the equipartition of energy.
It is sometimes called Clausius' postulate. The hypothesis that temperature is
a quantity of the same kind as molecular kinetic energy is one of the most
important fundamental postulates of the kinetic theory. From this it follows
that at the same temperature, the mean kinetic energy of the molecules of all gases
(which obey the ordinary gas laws) is the same. The total kinetic energ)^ of a gas
is proportional to the product jpv. Hence, if the temperature be altered, pressure
remaining constant, the kinetic energy (i.e. temperature) must alter to the same
748 INORGANIC AND THEORETICAL CHEMISTRY
extent, and hence also the volume. Otherwise expressed, if the pressure remains
constant, the same alteration of temperature will alter the volume to the same
extent. This is Charles' law. Since the mean kinetic energy of the translatory
motions of the molecules of any gas is constant, and when two or more substances
at difEerent temperatures are in intimate contact, their temperature will assume
one constant value, it is accordingly assumed that equality of temperature means
equality of the kinetic energies of the molecules, and Waterston's hypothesis takes
the form : Two gases are in thermal equilibrium when the mean kinetic
energies of their molecules are the same. Since real gases — which exert no
chemical or physical action on one another and which are under the same con-
ditions of temperature and pressure — can be mixed without change of pressure or
temperature, it is assumed that the molecules of equal volumes of two gases at the
same temperature and pressure possess the same total kinetic energies. The sense
impression we call temperature is our mode of perceiving the kinetic energy of the
translator}^ motions of the molecules.
The kinetic theory and Avogadro's hypothesis — From what has just been
stated, it follows that equal volumes of two gases at the same temperature and
pressure have the same value for the product fv. Hence, also, the total kinetic
energy of the one gas will be equal to the total kinetic energy of the other ; or
the product w^rnxFi^ for one gas will be equal to the product n^m^V^ for the other.
Again, the average kinetic energy, \MV^, per molecule in the two systems will be
equal if the temperature is the same ; and hence, \miV-^=\m2V<^ ; or, by sub-
stitution in the preceding relation, ^1=^2. This is the symbolic way of saying
that equal volumes of two gases under the same physical conditions contain the
same number of molecules, or that the pressure of a gas at any temperature depends
on the number of molecules and not on their kind ; that is, Avogadro's hypothesis.^
It is possible to argue backwards from Avogadro's hypothesis, and deduce the
assumption indicated above. The one is dependent upon the other. Remember,
therefore, that, contrary to what some enthusiastic writers have asserted, Avogadro's
hypothesis has rendered it necessary to introduce an unknown and unverifiable
assumption into the reasoning. According to J. C. Maxwell (1879) :
If the system is a gas, or a mixture of gases not acted on by external forces, the theorem
that the average kinetic energy for a single molecule is the same for molecules of different
gases is not sufficient to establish the condition of equilibrium of temperature between
gases of different kinds, such as oxygen or nitrogen, because when the gases are mixed we
have no means of ascertaining the temperature of the oxygen and nitrogen separately.
We can ascertain the temperature of the mixture only by putting a thermometer in it.
The kinetic theory should not be quoted as a proof that Avogadro's hypothesis
is true. Avogadro's rule is a corollary of J. J. Waterston's hypothesis.
The average distance traversed by a molecule between two colUsions — the
average free path. — The term average free path, L, of a molecule denotes the
average distance traversed by a molecule between two successive collisions : that
is, the sum of the free paths of all the molecules divided by the total number of
molecular paths. The average free path is a pure length, and is determined by
the concentration or number of molecules per unit volume, and it does not
depend upon the average speed of the molecular movements. The magnitude
in question for liquids at ordinary temperatures and pressures is of the same
order as the millionth of a millimetre, for Lord Kelvin 3 has shown by several
independent lines of argument — electrification of metals by contact, the tension of
soap films, and the viscosity of air — that in ordinary solids and liquids the average
distance between contiguous molecules is less than the hundred-millionth, and
greater than the two-thousand-milUonth of a centimetre. The length of the path
L which a molecule, moving amid a swarm of molecules at rest, can traverse without
collision is nearly
A3
Average free path, Z= _
77-r2
THE KINETIC THEORY OF ATOMS AND MOLECULES 74^
where A denotes the mean distance between any two neighbouring molecules,
and r is the radius of the sphere of action of each molecule, so that nf^ is the sectional
area of one of these spheres. It is assumed that the sphere of action of a molecule
cannot contain the centre of that of another molecule.
Let unit volume of a gas contain N molecules ; let this volume be divided into N small
cubes each of which on the average contains only one molecule ; let A denote the length
of the edge of one of these imaginary cubes, and accordingly, only one molecule is contained
in a cube of capacity A^. When a molecule A moves forward a distance x, it traces out a
cylindrical space of voliune irr^x, and if perchance * the centre of a molecule happens to
be inside this cylinder, a collision will take place. Assume that the probability of a
collision is in the ratio of these two volumes when x is small. Hence, the probability
P of a collision when a moleciile moves a distance x is P=i{r^xX^. If there are n mole-
cules moving like A, when n is very large, Pn molecules will collide in the distance x, and
Pn{\—P) molecules will pass on. Of the latter, Pn{\—P) will collide in travelling another
distance x, and Pn{\—P)^ will pass on. Hence, Pn travel a distance between 0 and x ;
P«2(l— -P) travel a distance between x and 2x ; Pn{\—P)^, a distance between 2x and
3a; ; etc. The total paths traversed by all the n molecules before collision will therefore
be Pnx + Pn{\—P)2x-\-Pn {\—Py^x-\- . . . When this series is summed by the rules of
algebra, it becomes nearly equal to nx/P, meaning that a distance nx/P is traversed by
n molecules before collision ; or a distance x/P is traversed by one molecule before collision ;
but P—Trr^x/X^, so that the average free path L—X^/nr^ as indicated above.
It has been assumed that all but one of the molecules in the enclosure are at rest.
It will be obvious that if all the particles are in motion, the probabiUty that the
particle A will collide with others will be increased, for the particles can also be
struck by others from the sides and which would not be encountered as a result
of its own motion. The increase in the probability of a collision shortens the free
path to say I in accord with R. Clausius' demonstration ^ that the number of
collisions will be increased in the proportion 4 : 3, and therefore the average free
path of a particle in a swarm of molecules in uniform motion is
, r 3 A3
Average free path, L=- .
4 7rr^
J. C. Maxwell, R. Clausius, and others have deduced relations between the
viscosity t^ of a gas and the average free path of the constituent molecules. If a
fluid — liquid or gas — streams through a tube of narrow
bore, it experiences a resistance to flow so that the
velocity of flow decreases uniformly from the centre
outwards until it reaches the walls of the tube where
the velocity is zero. Each layer of the fluid, parallel
to the direction of flow, exerts a tangential force on the
adjacent layer tending to decrease the velocity of the faster-moving layers, and to
increase that of the slower-moving layers. The property in virtue of which a
fluid exhibits this phenomenon is called the viscosity of the fluid. From the
analogy between viscosity and friction some — ^particularly the Germans — employ
the term internal friction — innere Reibung — of the fluid. Isaac Newton assumed
that the viscosity is directly proportional to the rate of decrease of velocity V in
the different layers of the fluid. If the tangential force be F, and the area of the
plane be 5, then F=^7]sV, where rj is the constant of proportion, the so-called
coefficient of viscosity, which can thus be defined as the tangential force per unit
area required to maintain unity velocity gradient between two parallel plates in
the fluid, at unit distance apart.
If a plane layer of fluid, CD, Fig. 1, moves with a velocity V^ parallel to another plane
layer of the same fluid, AB, at a distance d from it, and moving in the opposite direction
with a smaller velocity V^, the change of velocity per unit distance- — the so-called velocity
gradient— \B F = (Fi — Fg)/^. Let the force acting on an area F required to produce this
velocity gradient be F ; this force must act in the direction CD on the upper plane and in
the opposite direction on the lower plane. The simplest assumption regarding the force
F was made by Isaac Newton, who assumed that the force F is proportional to the velocity
gradient F, in the immediate neighbovirhood of the plane ; this force is also proportional
to the area 8 of the plane. Hence, P = ij«F.
750
INORGANIC AND THEORETICAL CHEMISTRY
The viscosity of gases is assumed to be an effect of the interchange of molecules
between different layers of gas, and the tangential force per unit area is measured
by the rate at which momentum is transferred per unit area between adjacent
layers. The molecules which travel from a faster into a slower moving layer possess
more momentum in the direction of flow than those moving in the opposite direction.
The net rate of transfer of momentum across unit area of a plane is, according to
Isaac Newton's second law, a measure of the force F^
F=ikLqMV
where A; is a constant.
Let the velocity V of the layer CD, Fig. 1, be taken as Vi=Jcd, where A; is a constant ;
and F = r]v/d per unit area. Assume that the motion of the molecules produces the same
effect as would be obtained by resolving the particles into three groups- — one-third move
perpendicular to the plane CD, and the other two perpendicular to one another and parallel
with the plane CD ; and further on the average half the molecules moving vertical to the
plane CD pass upward, and the other half downwards. Hence, one-sixth of the molecules
moving with an average velocity V, that is In V, traverse L, the mean free path, and their
average velocity is the same as that of a layer at a distance L below CD, namely k{d—L).
These molecules carry a momentum \k{d—L)nM V, where M denotes the mass of each of the
n molecules. A stream of the same number of molecules comes from AB, with the velocity
of a layer d-\-L, and their momentum is lk{d -\- L)nM V . The difference, ^/jLnM F, repre-
sents the total momentum lost by the layers above the plane CD to those below.
The loss of momentum causes a drag on the adjacent layer such that F^rjvld
per unit area. Equate these two values of F together, and reduce the result to
its simplest terms, remembering that k=vld,
r)=iLnMV', or, rj^iLDV
where the density D of the gas is equal to nM. For instance, taking the density
of air, Z)=0*001293, and the experimentally determined viscosity t7=0*00018,
and F=48,000 cm. per second ; the mean free path is nearly L=SxlO~^ cm. at 0°.
In deducing this equation, it has been assumed that the molecules all possess the
same velocity F, introducing the correction for Maxwell's distribution of velocities.
L. Boltzmann represented the viscosity of a gas at 0° by the expression :
V
-.0'3502DUL
(10)
where U denotes the average velocity of the molecules. Consequently, the average
length L of the free path — that is, the average distance traversed by a molecule of
the gas between two collisions — can be calculated, from the relation between viscosity
and the free path. Some results at 0° and 760 mm., with L. Boltzmann's equation,
are indicated in Table II. A molecule of nitrogen, for instance, travels about
Table II. — ^Average Free Path and Collision Frequency of Gases.
Gas.
Viscosity, r).
Average
velocity U cm.
per S3C.
Density, D.
Average free
path, L, cm.
Collision
freaue^ncy.
Hydrogen, Ha
843 X 10-'
1-696x105
88-73 X 10-6
16-00 X 10~«
10-6 Xl0»
Oxygen, 0, .
1905
0-425
1414
9-05
4-7
Nitrogen, Nj
1670
0-454
1234
8-50
5-3
Argon, A
2107
0-381
1758
8-98
4-2
Water, H2O.
904
0-566
798
8-91
6-3
Carbon dioxide, CO 2
1375
0-362
1951
5-56
6-5
Mercury, Hg
1620
0-170
4200
12-88
1-3
450 metres per second, it collides with other molecules about five thousand million
times per second, and the average distance between two consecutive collisions is
about eight ten-thousandths mm. The density of the gases is known, and their
THE KINETIC THEORY OF ATOMS AND MOLECULES 751
viscosities have been experimentally determined from observations on the damping
effects of the various gases on vibrating pendulums (0. E. Meyer), or oscillating
discs (J. C. Maxwell), 6 and by other methods.
R. Clausius' theoretical value for the effect of temperature on the viscosity of
a gas, is
,,=20-43£v/g; or, ,=20-43i(g)"'
where T denotes the absolute temperature of the gas ; and D, the density at 0° C,
and atmospheric pressure — air unity. Observations show that this expression
gives better results if the exponent be empirically taken nearer 0'77 than 0*5.
W. Sutherland's formula has an empirical constant dependent upon the nature of
the gas.
From L. Boltzmann's equation, it follows that the average free path is inversely
proportional to the pressure of the gas, so that while the average free path of the
gas at 760 mm. is about 16 Xl0~6 cm., the value increases to about 0*025 cm. when
the pressure falls to 0*5 mm. Consequently, the average free path of a molecule
of a gas is about 700 times its diameter, and over a million times the diameter
when the pressure is reduced to about half a miUimetre of mercury. The size of
a molecule is therefore small in comparison with its average free path. With
increasing pressures, the increase of density is proportional to the diminution of
the average free path, so that the product Dl is constant. Consequently, the
velocity depends only on the temperature of the gas. For any gas, therefore, the
viscosity is independent of the pressure if the temperature is constant, and the
viscosity increases with the temperature. The converse obtains with liquids, for
here the viscosity usually increases with rise of temperature. These results were
deduced from the kinetic theory for gases, and the prediction was subsequently
confirmed experimentally. The relation did not obtain with very high and very
low pressures — in the former case, attractive forces come into play, and in the
latter case, the molecules might travel the whole distance between the walls of the
confining vessel without encountering another molecule. In deducing the formula,
the assumption is made that the free path is so small that terms involving higher
powers of L can be neglected. When the gas is rarefied, the molecules have more
free space for their movements, and they do not collide so frequently. When the
gas is sufficiently rarefied to make the higher powers significant, the formula is no
longer applicable.
Direct determinations ^ of the average free path have been made by J. Franck
and G. Hertz, who measured the maximum distance at which two plates must be
placed apart in a gas in order that the ions passing from one plate to another may
produce fresh ions by collision one with another. The results with hydrogen agree
but approximately with the computed values :
Pressure 45 81 152 1670 bars
Average free path (Obs.) . . . 0-436 0256 0'149 0-014 cm.
Average free path (Calc.) . . . 0-388 0-215 0-115 0-011 cm.
The number of collisions between the molecules per second — the collision
fcreOLUency. — If the average free path of the hydrogen molecule is 16xlO~6cm.,
and the average velocity of translatory motion is 16*96 X 10* cm. per second,
it follows that the number of collisions per second must be the quotient of the
distance TJ traversed per second, into the length of the free path L, provided
the paired time of the colliding molecules be negligibly small. Hence, a molecule
undergoes nearly 10,000,000,000 collisions per second. The last column in
Table II shows estimates made with a number of gases. Why does a gas not
diffuse through another gas with a speed comparable with the velocity of translatory
motion of its molecules ? The hydrogen molecule, for instance, has a velocity
of r696XlO^ cm. per second, but the average free path is only 16x10— ^ cm.,
752 INORGANIC AND THEORETICAL CHEMISTRY *
so that the molecule collides with other molecules nearly ten thousand million
times per second. A molecule is therefore continually abutting against other
molecules, and being deflected from a straight path, so that its rate of progress in
any particular direction is comparatively slow.
The average diameter o! the molecules. — It is difficult to define pre-
cisely what is meant by the size of molecules primarily because their shape is
unknown. Whatever be the structure of the molecule, the diameter of the molecule
is understood to be a number which represents the smallest distance apart to which
the centres of two molecules can approach one another. It is highly improbable
that the molecules are elastic spheres, and the assumption that the molecules are
spherical will therefore introduce an error which affects the different magnitudes
concerned in free path and collision formulae approximately the same, and accord-
ingly J. H. Jeans ^ defines the molecular diameter as the diameter of a sphere such
that spheres of this diameter undergo the same number of collisions as occur in
actual gases. With solids, however, where free paths and collisions do not come
into consideration, he regards the molecular diameter as the diameter of a sphere
which occupies the same space as the molecule, and the more the molecules differ
from the spherical shape, the more will the value so obtained differ from the former
value. Calculations of the molecular diameter based on the volume occupied by
matter in the solid or liquid states of aggregation, with the additional assumption
that the molecules are packed as close as is physically possible, must give results
.too large- — they are, how*ever, regarded as useful in fixing an upper limit to the size
of the molecules.
SoHd hydrogen at 13'2° K. has a density of 0-0763, or one c.c. weighs 00763
grm. Since the mass of a molecule of hydrogen is 3*27 x 10"^* grms. the number of mole-
cules per c.c. of the solid will be 2-33 x 10^^^ jf thig number of spheres be packed as closely
as possible, they will occupy a volume 2-33 x 10^^ Xd^-^N c.c, where d represents the
diameter of each molecule. This volume must be less than 1 c.c. If it be 1 c.c,
d = 3*93xl0~^ c.c. This is therefore the upper limit to the molecular diameter. For
some unknown reason, in some cases the upper limit so obtained is less than the values
obtained by other methods of calculation presumably more exact- — e.g. xenon, benzene,
chlorine, carbon dioxide, ethylene, etc.
0. Loschmidt, in his memoir Zur Grosse der Luftmolecule, made the first estimate
of the actual size of the molecules of a gas in 1865. R. Clausius, J. C. Maxwell,
J. H. Jeans, fe. Chapman, and W. Sutherland have deduced expressions for this
constant. R. Clausius obtained TrndrL—O'lb ; J. C. Maxwell made the constant
0"707 ; and J. H. Jeans made it 0*933, to allow for the persistence of the original
velocity for a small period of time after a collision. J. H. Jeans, after making an
allowance for the persistence of the velocity after a collision, found
1"2547
Mean free path, Z=—, .... (11)
V'liTnd^
where d denotes the diameter of the molecule ; n, the number of molecules per c.c. ;
and L, the average free path. From (10), r)=02>bODUL, when Z)=0-001293 ;
t/=45100 cm. per sec. ; and on substituting the value of //from (11), it follows that
w(i-=3306 sq. cm. The result now depends upon what value is assigned to the
constant n, values for this constant are estimated in the next section. Estimates
of the maximum and minimum diameter of a molecule have been made by several
different methods— the viscosity of gases, the thermal conductivity of gases, the
rate of diffusion, the deviations from Boyle's law, liquid films, contact electricity,
refractive dispersion, and the dielectric constant or refractive index of a gas.^ The
results agree fairly well with those deduced from the kinetic theory. In illustration,
J. H. Jeans, The Dynamical Theory of Gases (Cambridge, 1916), obtained the values
indicated in Table III for the molecular diameters estimated by four different
methods, and expressed in centimetres. M. Knudsen computes the . molecular
weight of a gas from the viscosity data.
THE KINETIC THEORY OF ATOMS AND MOLECULES 753
Table III.— J. H. Jeans' Estimates of Molecular Diameters.
Coefficient
Conduction
CoeflBcient
Deviations
Average of
I to 111.
Gas.
of viscosity
of lieat
II.
2-68x10-8
of diffusion
III.
2-68 X 10-8
from Boyle's
law.
2-52 X 10-8
Hydrogen
2-68x10-8
2-68x10-8
Helium .
2-86
2-28
—
1-96
2-22
Steam
4-54
_-
. — .
^—
4-54
Carbon monoxide
3-78
3-80
3-72
—
3-76
Ethylene
5-52
6-52
5-48
>—
5-50
Nitrogen .
3-76
3-82
3-82
3-54
3-80
Air .
3-82
3-82
3-82
3-30
3-72
Nitric oxide
3-82
2-84
. — .
. — .
3-72
Oxygen .
3-62
3-60
3-62
—
2-62
Argon
3-64
3-60
—
. —
3-62
Carbon dioxide
4-54
4-84
4-30
3-40
4-56
Nitrous oxide .
4-6
4-62
4-54
. —
4-58
Ethyl chloride .
4-12
—
. —
■
5-12
Chlorine .
5-36
. —
. —
5-36
Benzene .
7-44
•~-
— '
7-44
The number of molecules in unit volume of a gas. — If all gases obeyed
the laws of Boyle and Charles, and Avogadro's hypothesis were valid, all gases
would have the same number of molecules per unit volume under the same con-
ditions of temperature and pressure. Assuming that the molecules of a gas are
spherical, then, the volume of each sphere will be Irrd^, where d denotes the diameter
of the molecule. If the gas contains N molecules per gram-molecular weight of
gas at 0° and 760 mm., their aggregate volume will be ^rrNd^. Again, let v denote
the apparent or total volume of a gas, and let h denote the space occupied by the
molecules ; the volume not occupied by molecules will be v—b ; and if v=h, the
molecules will be in contact provided b does not vary with pressure. J. D. van
der Waals' estimate lo of the value of b from H. V. Eegnault's observation is, for
air, 6=0'00198 ; and J. Rose-Innes' estimate from H. Callendar's observations,
6=0"00209. The mean is 0'002035. The values of b expressed in c.c. per gram-
molecule of the gas, and calculated from the deviations of the gases from Boyle's
and Charles' laws, are
16-28
31-56
39-50
32-22
H2O
30-52
CO2
42-83
Hg
35-67
According to J. T>. van der Waals, the actual volume of the molecules is one-fourth
the value of b ; consequently, Jfe=^7^r^F<?3,
b=^7rNd^
Consequently, ^3:^0-00097 c.c. Since ^2^3306 sq. cm., by eliminating d, it
follows that (iVf|2)3/(AV.3)2=iV=4-92xlOi9. This result, for air, is rather lower
than the value obtained for other gases. The errors of observation are considerably
magnified in the calculation — those of b are doubled, and those of rj are trebled.
Again, when evaluated by electrical methods N is approximately 4x10^9 per c.c.
The numerical value of the constant N has been determined by nearly a dozen
independent methods, and the most reliable determinations approximate ^"=6-062
X 1023 molecules per gram-molecule of the gas, and this number is called Avogadro's
constant. The number n per c.c. of the gas is iY/22412, or 2-7048x10^9 molecules
per c.c. of the gas at 0° and 760 mm. The approximate agreement of the numbers
is so close that R. A. Millikan 11 could say :
To-day wp are counting the number of atoms in a given ma-ss of matter with as much
certainty and precision as we can attain in counting the inhabitants of a city. No census
VOb. I. 3 C
754 INORGANIC AND THEORETICAL CHEMISTRY
is correct to more than one or two parts in a thousand. . . . There is httle probability
that the number of molecules in a cubic centimetre of gas under standard conditions
(0° and 760 mm.) differs by more than this amount from 2-70 X 10^^.
The letter R used for the gas constant appears to have been taken by Isaac Newton in
his Philosophice natiiralis principia 7nathetnatica {Itondon, 302, 1713) from the term resistance.
In his study of the inner resistance of a gas, he showed that if two parallel plates, at a
distance r apart, move with a difference of velocity v, the inner resistance for unit surface
is proportional to the increase of velocity v/r so that R = — 7){v/r), where tj is the coefficient
of friction. E. Clapeyron (1834) employed R to represent the constant in the gas equation
pv=R{a-\-6), where a is constant, now represented by 273. E. Clapeyron used 267.
According to the kinetic theory of gases, Newton's rj is equal to ^nmvl, so th&t R =lnmvH/r.^^
References.
^ J. J. Waterston, Phil. Trans., 183. A, 1, 1892 — ^posthumous publication.
2 A. Neumann, Ber., 2. 690, 1868 ; 3. 862, 1869 ; 4. 270, 1870 ; J. Thomsen, ib., 3. 828, 1869 ;
L Meyer, t6., 4. 25, 1870; F. Mohr,i6., 4. 78, 1870; R. A. Mees, t6.,4. 198, 1870; C. del Lungo,
Atti Accad. Lincei, (5), 25. ii, 322, 1916.
3 G. J. Stoney, Phil Mag., (4), 36. 132, 1868 ; Lord Kelvin (W. Thomson), Nature, 28. 203,
250, 274, 1883 ; Proc. Roy. Inst., 10. 185, 1883 ; 11. 483, 1887 ; A. W. Riicker and A. W. Reinold,
Phil. Trans., 177. 627, 1886.
* J. W. Mellor, Higher Mathematics for Students of Chemistry and Physics, London, 506, 1913.
« R. Clausius, Pogg. Ann., 105. 239, 1858 ; Phil. Mag. (4), 17. 81, 1859.
« O. E. Meyer, Pogg. Ann., 125. 177, 1865 ; 143. 14, 1871 ; Wied. Ann., 32. 642, 1887 ;
J. C.Maxwell, Phil. Mag., (4), 19 31, 1860; (4), 35. 209, 1868; Phil. Trans., 156. 249, 1866;
L. Boltzmann, Sitzber. Akad. Wien, 66. 324, 1872 ; 84. 41, 1881 ; 86. 8, 1230, 1881 ; 96. 895,
1887; J. Stefan, ib., 65. 363, 1872; P. G. Tait, Trans. Roy. Soc. Edin., 33. 259, 1887;
R. Clausius, Pogg. Ann., 105. 239, 1858 ; Phil. Mag., (4), 17. 81, 1859 ; (4), 19. 434, 1860 ;
W. Sutherland, ib., (5), 36. 507, 1893.
"> P. Lenard, Ann. Physik, (4), 12. 714, 1904 ; J. Robinson, Phys. Zeit., 11. 11, 1910 ;
J. Franck and G. Hertz, Zeit. deut. phys. Ges., 14. 596, 1912 ; 15. 373, 1913.
8 R. Clausius, Pogg. Ann., 105. 239, 1858 ; Phil. Mag., (4), 17. 81, 1859 ; J. C. Maxwell,
ib., (4), 46. 453, 1873 ; W. Sutherland, ib., (5), 36. 507, 1893 ; J. H. Jeans, ib., (6), 8. 692, 700,
1904 ; The Dynamical Theory of Gases, Cambridge, 236, 1904 ; S. Chapman, Phil. Trans., 211.
433, 1911 ; O. Loschmidt, Sitzber. Akad. Wien, 52. 395, 1865.
» A. W. Reinold, B. A. Rep., 986, 1885 ; Lord Kelvin (W. Thomson), Proc. Roy. Inst., 10.
185, 1883 ; M. Knudsen, Ann. Physik, (4), 44. 625, 1914.
i<* J. D. van der Waals, The Continuity of the Liquid and Gaseous States, London, 400, 1891 ;
J. Rose-Innes, Phil. Mag., (6), 5. 48, 1903 ; J. H. Jeans, ib., (6), 8. 692, 1904.
" R. A. Millikan, Phys. Rev., (2), 2. 109, 1913.
« E. Hoppe, Zeit. Elektrochem., 25. 216, 324, 1920; E. Clapeyron, Journ. iJcole Polyt., 14.
170, 1834 ; D. Bernoulli, Hydrodynamik, Argentorati, 1738.
§ 4. Attempts to Obtain a More Exact Gas Equation
All nature widens upwards. Evermore
The simpler essence lower lies.
More complex is more perfect, . wning more
Discourse, more widely wise. — Tennyson.
The general equation of state pv=RT does not exactly describe the behaviour
of real gases with respect to changes in volume with variations of temperature and
pressure. The same gas does not behave in the same way at high and at low pres-
sures. The laws of Boyle and Charles are fairly exact for some gases — e.g. hydrogen,
oxygen, etc. — at temperatures and pressures not far removed from normal atmo-
spheric conditions ; and it is often convenient to neglect small deviations with
other gases — e.g. carbon dioxide, ethylene, etc. This means that gas calculations
with pv=RT are made upon imaginary gases sometimes styled ideal or perfect gases.
When the pressure upon the gas is very great, the error becomes quite appreciable,
and it is necessary to revise the simple gas law : j)v=RT. This was emphasized
by H. V. Regnault in his Relation des experiences entreprises pour determiner les
principales lois physiques et les donnees nmneriques qui entrent dans le calcul des
machines a vapour (Paris, 1847). He said :
THE KINETIC THEORY OF ATOMS AND MOLECULES 755
The law does not express the actual relations between the same quantities of gas and
the pressure which they support, and it is desirable to find if it is possible to represent
these relations by a new law. . . . Unfortunately this relation is evidently too complex
to hope to find it by purely experimental methods. It is to be hoped that mathematicians
will try to find the form of this function by developing certain hypotheses on the nature
of molecular forces ; the necessary data for calculating the constants can be readily obtained
from observed measurements, and the formulae themselves subjected to un criterium
rigoureux.
The effect of the size o! molecules. — The reduction in the volume
which occurs when a gas is highly concentrated is smaller than corresponds with
Boyle's relation, and this is now explained in the following way : Under great
pressures the volume of the molecule becomes comparable in magnitude with the
space through which the molecule can move. The volume of the space in which
the molecules move is alone reduced by pressure, and therefore only part of the
total volume occupied by the gas can be reduced by pressure. Hence, at high
pressures the apparent volume and the product pv appear to be greater than is
described by Boyle's law. With hydrogen, for instance, when the pressure is
doubled, the volume is not quite halved. The same remark applies to other gases,
e.g. carbon dioxide, at great pressures. This is illustrated by the upward course
of the curves, Fig. 3, Cap. IV.
Let h denote the space occupied by the molecule as it moves to and fro between
the boundary walls AB, Fig. 2. If this distance be halved, AC, while the volume
of the molecule remains constant, the molecule will have less than half its former
distance to pass from one side to the other. For instance, suppose that AC repre-
sents one unit, and AB two units, and the diameter of the molecule is y^^th unit ;
L-@ J L_@ J
Fig. 2. — The Effect of the Size of the Molecules on the Volume of Gases.
the molecule oscillating between AB then moves through 1*9 units of space in one
journey ; and between AC through 0'90 unit, not 0*95 unit. Hence, the molecule
will strike the walls more frequently than before per unit time, and the outward
pressure due to molecular bombardment will increase more rapidly with decreasing
volume, than is described by Boyle's law. Boyle's law refers to the whole volume
of the gas, but rather should it refer to the space in which the molecules move.
Consequently, v—b should be substituted in place of the v of Boyle's law, and the
result is : p{v^h)=RT, where b is called the molecular CO-VOlume. An apphcation
of the theory of probability to the kinetic theory has led to the view that b is
very nearly four times the volume actually occupied by the whole of the mole-
cules contained in unit volume of the gas, so that a molecule is represented to be
a complex vibratory system with a material nucleus J6/« in size, which requires
b/n volumes of space in which to perform its oscillatory movements — n represents
the number of molecules in unit volume of the gas — hence b is also called the
vibratory volume of the molecule. The fundamental assumptions have not
been so firmly established that there is no room for doubt, and some consider that
b represents the real volume of the molecules ; others believe that b is much greater
than four times the size of the material nucleus of the molecules. In any case,
virtually all are agreed that b is not quite constant, but varies with the temperature,
and possibly also with pressure changes. J. D. van der Waals, in his celebrated
Over de Continuiteit van den Gasen Vloeistoftoestand,^ pubhshed at Leiden in 1873,
worked on the assumption that b is constant.
The effect of molecular attraction. — If the molecules of a gas have appreci-
able cohesion or attraction for one another, they must be swerved from their
rectihuear paths when they come within the sphere of one another's influence,
756
INORGANIC AND THEORETICAL CHEMISTRY
and they must then move in curved, not in straight paths. Doubling the number
of particles per unit volume will not then give exactly twice the number of impacts
on the boundary walls. When the molecular attraction is marked, the product
pv must be less than corresponds with Boyle's law. Molecular attraction deflects
some of the molecules from the straight path so that they do not strike the walls
of the vessel under conditions where they otherwise would, ahd the pressure is
accordingly diminished. This appears to be the case with carbon dioxide, and most
gases with a smaller apparent volume v, or a smaller value of jrv, that is, a
greater concentration, than corresponds with an increase of pressure as described
by Boyle's law. This is illustrated by the downward slope of the pv curves, Fig. 3,
for carbon dioxide below 150 atmospheres pressure. The closer the proximity of
the molecules to one another, the greater will be the effect of the attractive forces
between the molecules. This attractive force, which may be denoted by F, will
tend to make the gas occupy a smaller volume. The effect is much the same as if
the gas were subjected to the action of a greater external pressure p-\-F than the
observed or apparent pressure p of the gas. With these ideas before him, G. A.
Hirn (1868) - proposed to use an equation of the form (p-\-F){v—b)=RT in place
of the regular equation pv=RT. The assumption that the attractive force F
between the molecules varies inversely as the fourth power of the distance between
the molecules, leads to a/v^ as the magnitude of the molecular attraction, F, where
a is a constant which varies with the nature of the gas, and v denotes the observed
volume of the gas. Granting the assumption, a/v- must be added to the observed
pressure of the gas in order to indicate the total pressure tending to compress the
gas. The magnitude of the internal or cohesive pressure a/v^ for liquid water is
nearly 11,000 atm. — a surprisingly large value which is in approximate agreement
with results obtained by other independent methods of calculation.
On correcting the equation pv—RT for the volume and cohesion of the mole-
cules, J. D. van der Waals (1873) obtained the so-called J. D. van der Waals'
equation :
{vA-^)=nT
(12)
This amended equation agrees fairly well with a number of observations of gases
under large pressures, and of gases near their points of liquefaction — e.g. ethylene,
carbon dioxide, etc. It also describes many of the properties of liquids, and of
the continuous passage of a gas to the liquid condition. The constants a and h
can be evaluated from observations. The numerical values of J. D. van der Waals'
constants a and h for some substances are indicated in Table IV.
J. D. van der Waals (1888) found that for carbon dioxide, yi'-:l-00646 ; h
=0*0023 ; and a~0"00874, when the unit of pressure is one atmosphere, and the unit
THE KINETIC THEORY OF ATOMS AND MOLECULES 757
of volume is the volume of one gram at 0° and one atmosphere pressure. With
these numbers, J. D. van der Waals' equation for carbon dioxide assumes the form
This equation may be employed for comparing the observed values for a gas which
is known to deviate rather considerably from R. Boyle's simple relation. Table V
has been computed from the numbers obtained by E. H. Amagat (1893) ^ at 20°.
Table V. — Comparison of Boyle's and Van der Waals' Equations for pv Observa-
tions WITH Carbon Dioxide.
p
pv.
(atmospheres).
Observed.
Calculated.
(J. H. van dec Waals' law).
Calculated.
(Boyle's law).
1
50
75
100
200
500
1000
0-680
0-180
0-228
0-419
0-938
1-000
0-678
0-179
0-226 .
0-411
0-936
1-000
1-000
1-000
1-000
1-000
1-000
The agreement between theory (J. D. van der Waals) and fact (observed data)
is quite good. It will be seen that if the gas behaved according to the Boyle's
equation, the product pv would have had the same constant value for all pressures.
As a matter of fact, the value of pv first decreases and then increases for all gases
except hydrogen and helium. The two corrections act in opposite ways. At first
the value of pv is decreased by the molecular attraction, but increased to a greater
extent by the finite dimensions of the molecule ; the two corrections balance one
another at ordinary pressures ; and at low pressures, the correction for molecular
attraction preponderates over that required for the volume of the molecule. The
correction for the volume of the molecule is relatively large when the volume of the
gas is compressed very small by a large pressure. If the numerical value of the
term a/v^ could exceed that of RT/{v—b), negative pressures would appear as
indicated by the dotted line in Fig. 3. This is unreal and therefore unsatisfactory,
and does not spppear with a second approximation to the gas equation proposed by
C. Dieterici * in 1899. The experiments of M. Berthelot (1850) and of A. M.
Worthington (1892) are sometimes quoted to demonstrate the existence of negative
pressures. Here, sealed tubes, quite full of liquid, were cooled slowly, and in some
cases the thick-walled tube collapsed owing to enormous tension. It may mean
that the molecules of the fluid exert an attractive influence on the walls of the
vessel, whereas J. D. van der Waals' equation assumes that there is no such attrac-
tion. The alleged negative external pressure is probably a myth.
The numerical values of a and h of van der Waals' equation are, in reality, not
constant at different temperatures ; thus, F. B. MacDougall (1916) calculated
from E. H. Amagat's results for carbon dioxide :
20°
40°
60°
80°
100°
137°
a .
.
.
0-00983
0-00919
0-00852
0-00797
0-00749
0-00708
b .
.
.
0-00202
0-00221
0-00227
0-00228
0-00226
0-00227
showing that a decreases with increasing temperatures even above the critical
temperature, while h, if it increases at all, increases very slowly ; and above the
critical temperature is virtually constant and independent of temperature. There
are also indications that for low pressures, h is not affected, but diminishes when
758 INORGANIC AND THEORETICAL CHEMISTRY
the pressures are very large. There is a wide divergence in the values of the con-
stant a by different methods, thus, for carbon disulphide, numbers ranging from
the 1683 (15°) of H. Davies to the 3363 (0°) of J. D. van der Waals, have been
reported.^
There are various methods for evaluating the constant a of J. D. van der Waals' equa-
tion. These methods have been examined by A. P. Mathews. They are: (1) from the
surface tension ; (2) from R. Eotvos rule or T. Young's rule ; (3) from J. D. van der Waals'
equation at the critical temperature; (4) from the latent heat of vaporization ; and (5) from
A. P. Mathews' formula a = 1*249 X lO^^MZv, where M denotes the molecular weight, and
2v, the number of valencies per molecule.
There has been a great deal of tinkering with J. D. van der Waals' equation.
Over thirty attempts have been made to modify still further the gas equation
to make it better describe the behaviour of gases under wide variations of pressure
and temperature. C. Dieterici (1899) has made one of the best attempts. In most
cases, other terms involving special constants which have to be evaluated from the
experimental numbers, have been introduced. Such equations are therefore of
limited application.
C. Dieterici's gas equation. — Instead of assuming that the observed pressure
of a gas should be reduced by the subtraction of a term ajv^, as was done by
J. D. van der Waals, to give his equation the form
t+>-*)=«2';or^=^^^-^« . . . (13)
C. Dieterici (1899) assumed that the term RT/{v—b) should be reduced by multi-
plication with a function, e-a/t'^^j always less than unity. Consequently, his
equation assumed the form
— _J?_ nm «
j)(v-h)=RTe ^^^; or, 7?= .e ^^2' . . . (14)
V — 0
where e represents the base of the natural logarithms.
C. Dieterici argued that in the interior of the liquid the attractive forces between the
molecules are balanced ; at the surface, the molecules are subjected to an unbalanced
force directed inwards. Consequently, the density of the layers near the surface will
decrease from the interior outwards. Only those molecules with a velocity exceeding a
certain value will be able to penetrate the surface layer and exert a pressure on the con-
taining wall : molecules moving towards the interior are assisted by the force directed
inwards. If the fraction of the total number of molecules which has a velocity greater
than this limiting value can be estimated, it follows that the observed pressure p will be
that fraction of the interior pressure RT/{v—b). From the theory of probability, if a
represents the most probable speed of the moleciiles ; and S, the speed the molecules must
possess to be able just to penetrate the surface, the required fraction will be e"*^'*' ;
and if V denotes the mean velocity, more exactly, the square root of the mean of the squares
of the speeds, V^ = ^a^, or o^^fF^. [f M represents the mass of a molecule and n the
number of molecules, e-S'-ja- becomes e ^^^V', and from (2), \nMV^=-RT -, and \MnS'^
will represent the work W done by a molecule penetrating the surface layer of molecules
against the molecular forces. The previous expression thus becomes e~^/-^^. C. Dieterici
then assumed that the work W is proportional to the density of the gas or W =alv, where
a is a constant. Whence follows equation (14) above.
The curves obtained by plotting J. D. van der Waals' and C. Dieterici's equations
have the same general form. Fig. 3. Both equations reproduce the critical state
very well, but Dieterici's equation agrees better with the general results of observa-
tion particularly at high pressures, where J. D. van der Waals' equation usually
breaks down. At low pressures, where v is large in comparison with h, both equations
give equally good results. Since, under these conditions, C. Dieterici's equation
reduces to that of J. D. van der Waals. This can be shown by expanding Dieterici's
THE KINETIC THEORY OF ATOMS AND MOLECULES
759
equation and omitting the higher powers of v, because when v is large, these terms
are negligibly small. In that case,
^ RTr a \RT__^ Rl
^ v—b\ vRfy v—b v{v—h)~v—
RT
b
(15)
van
smce, when h is small in comparison with v, v{v—b) approximates to v^. J. D.
der Waals' equation is thus a special case of C. Dieterici's equation applicable to
low pressures, just as the equation j)v=RT is a special case of the same equation
applicable to gases, for which a and b
are negligibly small in comparison with
p and V.
If the simple equation pv=RT be
regarded as a first approximation to a
true gas equation, J. D. van der Waals'
equation can be regarded as a second, and
C. Dieterici's as a third approximation
towards a complete law. Most, if not all,
the formulsB of physics and chemistry are
in the earlier stages of such a process of
evolution. As I have said elsewhere,
m-p
1
i\
\
\\
f,
?^j
'-)'("-
*)
-RT
—
\ '
s
V
V
\
80
l -1^
^ji
6^
\
[^
TT
-^
■^
■~-~
^
^^
-.
r^
::^
-^
-_
H
M/
u
—
^
^
K
»?-
M4^
0-!l2-
003
—
0)4
""^o^i;
-Graphs of J. D. van der Waals'
Equation.
There is a prevaiHng impression that once v'
a mathematical formula has been theoretically Fig. 3.-
deduced, the law embodied in the formula has
been sufficiently demonstrated provided the
differences between the calculated and the observed results fall within the limits of experi-
mental error. With improved instruments, and better methods of measurement, more
accurate data are from time to time available. The errors of observation being thus reduced
the approximate nature of the original formula becomes more and more apparent.
Ultimatelj'^ the discrepancy between theory and fact becomes too great to be ignored. It
is then necessary to "go over the fundamentals." New formiilas must be obtained em-
bodying less of hypothesis, more of fact. Thus, from the first primitive guess, succeeding
generations progress step by step towards a comprehensive and a complete formulation of
the several laws of Nature.
References.
^ J. D. van der Waals, Die Continuitdt des gasformigen mid fliissigen Zustandes, Leipzig,
1899 ; Physical Memoirs, 1. 333, 1891 ; J. P. Kuenen, Die Zustandsgleichung der Gase und
Flussigkeiten und die Kontinuitdtstheorie, Braunschweig, 1907 ; H. K. Onnes and W. H. Keesom,
Die Zustandsgleichung, Leipzig, 1912 ; S. Young, Stoichiometry, London, 1918 ; L. Graetz,
Kritischer Zustand der Flussigkeiten und Ddmpfe, Leipzig, 1906 ; W. V. Metcalf, Journ. Phys.
Chem., 19. 705, 1915 ; 20. 177, 1916 ; M. N. Shaha and S. N. Basu, Phil. Mag., (6), 36. 199, 1918,
2 G. A. Hirn, Theorie mecanigue de la chaleur, Paris, 2. 215, 1864 ; Ann. Ghim. Phys., (4),
11. 47, 1867 ; Phil. Mag., (4), 35. 461, 1868.
3 E. H. Amagat, Compt. Rend., 115. 919, 1893 ; P. A. Guye and L. Friderich, Arch. Phys. Nat.
Geneve, (4), 9. 505, 1900 ; L. Friderich, Journ. Chim. Phys., 4. 123, 1906.
* C. Dieterici, Wied. Ann., 69. 685, 1895 ; A7in. Physik, (4), 5. 51, 1901 ; F. H. MacDougaU,
Journ. Amer. Chem. Soc, 38. 528, 1916 ; A. M. Worthington, Phil. Trans., 183. A, 355, 1892;
M. Berthelot, Ann. Chim, Phys., (3), 30. 232, 1850.
5 H. Davies, Phil. Mag.,HS), 24. 422, 1912 ; W. C. McC. Lewis, ib., (6), 25. 61, 1912 ; Trans.
Faraday Soc, 7. 94, 1911 ; P. Walden, Zeit. phys. Chem., 66. 385, 1909 ; C. Winther, ib., 60.
590, 1907; I. Traubc, ib., 68. 289, 1909; A. P. Mathews, Journ. Phys. Chem., 17. 154, 180,
605, 1913.
§ 5. J. D. van der Waals' Theory of Corresponding States
J. D. van der Waals' equation assumes the form of an equation of the third
degree in v when it is multipHed out :
^ f,. RT^^. a ab ^
i;3— ( b-\ 1^24- -V =0
V. p J p p
(16)
760
INORGANIC AND THEORETICAL CHEMISTRY
CBA
In algebra, we are taught that such an equation must have three roots real or
imaginary, equal or unequal ; and of the real roots, there may be one or three
equal or unequal — imaginary roots have no physical meaning. Otherwise expressed,
there may be one or three different volumes corresponding with certain assigned
values of p and T. If there is only one real root, A, Fig. 4, the equation furnishes
only one value of v for every assigned value of p. The graph for carbon dioxide
above the critical temperature is an example. If the equation has three unequal
roots, C, Fig. 4, there ought to be three different values of v at the given pressure
and temperature, but only two of these have been realized, since the middle portion
of the curve is physically unstable. The line of
constant pressure cuts the theoretical curve in
three places as indicated at a, b, and c, Fig. 4, but,
instead of the pressure increasing to a maximum
a (Fig. 4), falling to a minimum j3, and then
increasing indefinitely, as the volume is diminished,
the pressure increases to a certain value, and then
remains constant until the gas has completely
condensed to a liquid. The curve C'a has been
prolonged a very short distance towards j3 by
undercooling ; and the curve Co prolonged a little
towards a by supersaturation. The suggestion
that the dotted line. Fig. 4, be substituted for
the horizontal part of Andrew's curves was made
by J. Thomson in 1862. For three unequal roots,
the line of constant pressure cuts the theoretical
curve in three places, as indicated at a, b, c.
Fig. 4 ; but, when there are three equal roots — j8. Fig. 4 — there is only one numerical
value for v for the assigned values of j) and T. This occurs at the critical tem-
perature K.
Let Vc, pci and Tg respectively denote the critical volume, critical pressure,
and critical temperature ; Vc will be the root of van der Waals' equation at the
critical temperature, and (v—Vc)^=^0. Expand this equation, and the result is an
identity with (16) above. Equate coefficients of like powers of v, and it follows that
Cl
r^
S\
4t^^
n >■
4^^^^
-.L^-^^\^
^li: 15^55-^,
''^i ^-B
r :i:-c
Temperatures
Fig. 4. — Graphs showing the
Roots of J, D. van der Waals'
Equation.
Vc=36; p.
2762
8a
21bn
(17)
These results enable the values of the constants a, b, and U to be calculated when
the critical volume, pressure, and temperature are known, on the assumption that
a, b, and R are constant. Corresponding values for C. Dieterici's equations are
or * m <*
ibR
. (18)
Again, let V=plPc ', V=v/Vc ; and T=TITc, then, by the substitution of these
values in J. D. van der Waals' equation (18), there remains
(p+ai
V+^,pv-l)=ST
(19)
The magnitude v is called the reduced volume ; p the reduced pressure ; T the reduced
temperature ; and (19), the reduced equation of state. The operation of reduction
seems to have freed van der Waals' equation of specific constants peculiar to in-
dividual substances, and substituted numbers of universal application in their
place. The result means that if the initial assumption be granted, different sub-
stances can exist in such states or conditions that their volume, pressure, and tem-
perature are respectively the same fractions of their critical values. Hence, states
characterized by the same values of v, p, and T, were called corresponding
THE KINETIC THEORY OF ATOMS AND MOLECULES 7(U
states — uhereinstimmende Ziistande — by J. D. van der Waals. This extraordinary
conclusion means that at the critical point the relation between the pressure,
volume, and temperature is the same for all substances ; no matter what
the substance, no matter what be the diameters of the particles, the range
and magnitude of the molecular forces, or the potential energy of the particles,
the same relation holds ; any two of these variables being given, the third can be
calculated.
J. D. van der Waals' theory of corresponding states was first regarded as being
derived from the molecular theory, but it is now treated as being based on a purely
empirical equation of state, like those of R. Clausius ^ and D. Berthelot, both of
which were empirically devised to represent the facts more nearly than the state-
equation of J. D. van der Waals. G. Meslin has shown that the theory of corre-
sponding states follows directly from any equation of state with not more than
three constants. R. Clausius' equation of state has four constants — namely a, b, c, R
— and it gives a reduced equation with one constant A :
)(.-6)=i2r;(p+ . ^,)(V-1)=T
V ' T(v+c)2/^ ' ' N ' T(v+A)
If the constant c be removed, D. Berthelot' s equation of state with three constants,
and a reduced equation with no constants :
(?'+/,2>-6)=-B2' ; (p+xy^'-i)^
are obtained. P. Curie has also shown that at the critical point (dpldv)T=Oy
{d'^p/dv-)T=0, and any critical point so defined will serve for setting up reduced
equations. For example, let Pq, Vq, and Tq be critical points, then, from J. D. van
der Waals' equation, h—Bv^ ; a=ApQVQ^ ; and R^CpqVq/Tq, where A, B, and C
represent pure numbers. Then p/'Po=V ', v/vq^=y ; and T/Tq=T, so that
(p_^)(v_B)=CT
a reduced equation containing only numerical constants ; J. D. van der Waals'
form of the reduced equation is a special case of this. D. Berthelot has set up three
reduced equations with special properties — in one, the unit of reference in the
critical point is defined by dp/dT=^0, where p is a, maximum. As G. von Kaufmann
has shown, the theory of corresponding states is quite independent of any critical
point, for by assigning specific or special units for p, v, and T, for each substance,
three specific constants can be eliminated from any equation of state.
S. Young 2 tested the law of corresponding states with a few substances ; it
seems to hold fairly well for a few groups of related compounds — hydrocarbons,
esters, ketones, ethers, etc. — where the results are not disturbed by molecular
association. It does not agree closely with water, the alcohols, and the fatty
acids. The results with the monatomic gases argon and helium agree amongst
themselves, but not with those of other groups. A few examples are indicated
in Table VII. In S. Young's method the values of v^, pc, and Tc are measured and
the different functions compared with equal values of p and T. E. H. Amagat
(1896) recommended a method in which no knowledge of the critical values is
needed, because curves with the variables log p, log v, log T, etc. are plotted, and
from the theory of corresponding states, the curves for different substances should
be of identical shape and superposable by a parallel shifting of the axes. C. Raveau
(1897) applied E. H. Amagat's method to ethylene and carbon dioxide. K. Meyer
and D. Berthelot found that a fairly accurate correspondence of states exists if the
reduced variables p/pc, {'v—v,)/{vc—Vm), and {T—Tc)/{Tc—T,),) be chosen, where
Tn and % have specific values for each substance. G. von Kaufmann sums up the
position :
762
INORGANIC AND THEORETICAL CHEMISTRY
Although in many cases a fairly approximate correspondence of states has been found
to exist, the theory in its entirety has been proved without doubt inexact ; it has not
been foimd completely true for even a single pair of substances. There is therefore no
general jt), v, T-equation of state with only three specific constants. Nevertheless, a theory
which is so far-reaching and fundamental as this, and which over a whole range of phenomena
gives a good first approximation to the facts, will not be lightly discarded, and in the present
position, attempts are being made to modify it in such a way as to bring it more into agree-
ment with the truth.
The critical density.— P. A. Guye (1890) and S. Young (1892)3 showed
interesting consequences of combining the three equations (18). S. Young obtained
Vepc=iRTc. Let v denote the volume occupied by the substance in the gaseous
Table VI.- — Iixustrations of the Law of Corresponding States.
Gases and vapours.
'-1
^ To
Carbon tetrachloride
0-725
0-408
27-5
Stannic chloride .
0-736
0-403
28-1
Ether .
0-738
0-403
28-3
Benzene
0-728
0-407
28-3
Fluorobenzene
0-733
0-407
28-4
Ethyl alcohol
0-735
0-400
29-6
Acetic acid .
0-762
0-410
25-4
state, and assume that it behaves like an ideal gas at the critical state when V'Pc=^RTc-
By division, Vc=^^v. If dc denotes the observed density of the gas in the critical
state, and Dc the theoretical density required for an ideal gas, then, remembering
that the density is the reciprocal of the volume, (?c/-^c=i==2*67, so that if J. D.
van der Waals' equation accurately describes the behaviour of the gas in the critical
state, the observed critical density of all gases ought to be 2 "67 times the theoretical
density of an ideal gas at the critical temperature and pressure. This is not the
case. The actual results are larger, being somewhere near 3*67 for the hydro-
carbons, esters, ketones, and ethers.
CC14
SnCl4
CO 2
SO2
CH4
02
3-65
3-76
3-61
3-62
3-67
3-49
3-53
Abnormally high results are obtained with associated substances — e.g. the fatty
acids and alcohols have values approximating 4 or 5. Argon has the value 2" 71
(D. Berthelot, 1901), and hydrogen 2*69 (J. J. van Laar, 1904) ; and these gases
alone approximate with any reasonable accuracy to the value required by van der
Waals' hypothesis. According to C. Dieterici's values for the critical data, the
critical density is 3*6945 times greater than the density of an ideal gas at the critical
temperature and pressure. This is a much closer approximation to the actual
results with normal substances than is obtained with the equation of J. D. van der
Waals, for, according to S. Young (1892), the value of dcjDc is nearly 3' 7 for all
substances which can attain the critical state without chemical change.
Method of determining molecular weights from the critical constants. —
— If the pressure be expressed in atmospheres and the unit of volume be the volume
occupied by a gram-molecule of the gas under normal conditions, J. D. van der
Waals * has shown that the equation
so that the relative molecular volumes of the different gases at 0° and 1 atm. pressure,
THE KINETIC THEORY OF ATOMS AND MOLECULES
763
or the volumes of the different gases at 0° and 1 atm. which contain the same number
of molecules, are proportional to
1 1
(1 +«)(!-&)' (l+a')(l-hr ' "
and the volume in litres occupied by a gram-molecule of the gas under
normal conditions, by Avogadro's rule, will be the same for all gases, so that
(l-\-a){l—h)MIW=a. constant, where W denotes the weight of a litre of a gas
under normal conditions :
M-.
22412Tf
(l+«)(l-fe)
since P. A. Guye (1905) has shown Avogadro's constant to possess the same normal
value 22' 41 2 for all gases. The constants a and b vary with temperature, and
Guye considers that the value of the critical constants at the temperature T can
be represented by the equations :
.m<)S .=<i+fc'x-^p
The numerical value of the constant j8=0'0032229 has been deduced from the critical
constants, density, and molecular weight of carbon dioxide, consequently the mole-
cular weight of the gas is given by the expression :
22-412Tf
J.TJ.<
JlK!\iU.l.a,L wcigxiv
(l+ao)(l+W
The following
examples illustrate the application of the rule :
Carbon
Nitrous
Sulphur
Hydrogen
* Acetylene.
dioxide.
oxide.
dioxide.
chloride.
W .
1-9768
1-9774
2-9266
1-6407
1-1707
Tc .
303-98
311-8
428-4
325
308-25
Pc '
72-93
77-8
78-9
83
61-03
ax 105
721
719
1345
726
879
6x105
. 191
185
251
180
231
ttoXlOS
847
878
2644
943
1055
6oXl05
161
156
255
153
207
M .
44-003
44 000
64-065
36-484
26-018
With gases which have a low critical temperature and which do not liquefy
very readily, the correction of the constants a and b for temperature is not necessary,
and
Molecular weight
(22-412 -mTc)F
Tl+a)(r-6r
7.^ (22-412
0-000q623rc)TF
is sufficiently accurate, where the value of the constant m has been fixed with
respect to oxygen =16, at m=0- 0000623. For example.
W .
T
axl05
6x105
M .
Oxygen.
1-4290
154-2
266
139
32
Hydrogen.
0-089873
32
28-8
73-7
2-0153
Nitrogen.
1-2507
128
275
174
28-013
Argon.
1-7802
152
260
138
39-866
Carbon monoxide.
1-2504
133-6
284
172
28-003
A. Leduc's^ method of molecular volumes for determining the molecular
or atomic weight of gases. — A. Leduc's method is related to D. Berthelot's
method of limiting densities, but as experimental data it requires a knowledge
764 INORGANIC AND THEORETICAL CHEMISTRY
of the densities and the critical constants. It is based on the theory of
corresponding states. The variations of pressure f, molecular volume 7, and
absolute temperature T of a gas described by Boyle's and Charles' laws, are related
by pV=KT, where Kis a. constant the same for all gases. All known gases deviate
from this rule, and the relation is then represented by ^F'=^T, where K^ is related
with K so that K'/K=V'/V. Let F'/F, the ratio of the molecular volume F' of
a gas to the molecular volume F of an ideal gas at the same temperature, be repre-
sented by <^-=F7F. Consequently, it follows that K'IK=^, and j)V'=KT<l>y
where 0 is variable quantity. If M denotes the molecular weight, and v the specific
volume, it follows, since Mv=V\
Mpv=KT(l>
For oxygen gas M =32 ; v=Vi ; and <^=<^i ; accordingly, Z2'pvi=KT<j)i ; and by
division, remembering that if D and D^ respectively denote the densities of a
given gas and oxygen, D—l/v and Di^ljvi — all at the same temperature and
pressure :
M_cl>D
32~^iZ)i ^^^
If 1 —E and 1 —Ei be respectively substituted for </> and ^j, this equation is identical
with that used in D. Berthelot's method of limiting densities.
In order to evaluate cj) and </>i, it follows that the relation for zero pressure
becomes MpQVQ—KT(f)Q, and by division pv/pQVQ=(l)/<f>Q. A. Leduc then assumes
that <f)Q may be regarded as unity at a common temperature, and at an indefinitely
small pressure ; this means that under these conditions all gases have the same
molecular volume, consequently
A. Leduc further assumes that the compressibility of a gas over a range of a few
atmospheres, can be represented by the expression E=l—pvj'PQVQ, or
1 — — =mv4-np^
where m and n are small constants to be determined for each gas. Accordingly,
<j>=\—mp—np'^y and therefore
^=l-n^pl^)-npM'f .... (2)
A. Leduc measures p in cm. of mercury, and pc in atm., and he denotes the ratio
pip bye.
To evaluate the constants m and n, A. Leduc assumes that the molecular volumes
of gases are equal at the same reduced temperature T/Tc, and reduced pressure,
pIPc- At the same reduced pressure, different gases at the same reduced tem-
perature give the same values of <j). Consequently, the coefficients mpc and npc^
must each be functions of the reduced temperature only. A. Leduc finds that
within the limits of experimental error, these functions are respectively :
T/T3 _T3 T \ TJ^iTn \
mpc=18-85y2^j; -V2^, +2V2 jT -IjxlO-^ ; «p.2=3.5^;(^ -"_1 )xl0-4
The results represent his experiments on compressibility very closely, excepting
with ammonia, phosphine, hydrogen sulphide, and methyl ether.
THE KINETIC THEOKY OF ATOMS AND MOLECULES 765
In applying this method, (i) the density B of the gas at the temperature T and
pressure f ; and (ii), the critical pressure fc and temperature Tc, are supposed to
be known. This enables the constants m and n to be evaluated. It is then necessary
to calculate <j> from equation (2). The value of <j>i for oxygen is also supposed to
be known ; and the required molecular weight is then calculated from (1). In
practice r=273 and 2>=1. A comparison of some results by these different physical
methods gives :
Method of
Hydrogen.
Nitrogen.
Carbon.
Chlorine.
Limiting densities
10075
14-008
12-009
35-461
Critical constants
1-0075
14-010
12-003
35-436
Molecular volumes .
10075
14-006
12 005
35-450
References.
1 R. Clausius, Wied. Ann., 9. 337, 1880 ; 14. 279, 692, 1881 ; D. Berthelot, Arch. NeerU
(2), 5. 417, 1900 ; G. Meslin, Compt, Bend., 116. 135, 1893 ; P. Curie, Arch. Sciences Genkve, 26.
13, 1893 ; K. Onnes, Proc. Acad. Amsterdam, (2), 16. 241, 1881 ; Arch. Neerl, 30. 101, 1897 ;
H. Happel, Phya. ZeiL, 6. 389, 1905.
2 S. Young, Phil. Mag., (5), 30. 423, 1890 ; (5), 34. 506, 1892 ; (5), 33. 153, 1892 ; (5), 37.
1, 1894 ; Jowrw. Chem. Soc, 59. 125, 1891 ; 63. 1254, 1893 ; E. H. Amagat, Compt. Bend., 123.
30, 83, 1896 ; 156. 271, 843, 1913 ; Journ. Phys., (3), 6. 1, 1897 ; C. Raveau, ih., (3), 6. 432,
1897; Compt. Bend., 123. 109, 1896; G. von Kaufraann, Phil. Mag., (6), 30. 146, 1915;
K. Meyer, Zeit. phys. Chem., Z2. 1, 1900; D. Berthelot, Journ. Phys., (4), 2. 186, 1903;
H. K. Onnes, Arch. Neel., 30. 101, 1897 ; Proc. Akad. Amsterdam, (2), 16. 45, 1881 ; W. Natanson,
Compt. Bend., 109. 855, 890, 1889 ; H. Happel, Phys. Zeit., 6. 389, 1905.
» P. A. Guye, Compt. Bend., 110. 141, 1890 ; S. Young, Phil. Mag., (5), 33. 153, 1892.
* J. D. van der Waals, Die Continuitat des gasformigen und fliissigen Ztistandes, Leipzig, 85,
1899 ; P. A. Guye and L. Friderich, Arch. Phys. Nat. Geneve, (4), 9. 505, 1900 ; D. Berthelot,
Zeit. Elektrochem., 10. 62, 1904 ; P. A. Guye, Journ. Chim. Phys., 3. 321, 1905.
6 A. Ledue, Ann. Chim. Phys., (7), 15. 5, 1898 ; (8), 19. 441, 1910 ; H. F. V. Little, Science
Progress, 7. 504, 1913.
§ 6. Summary of the Kinetic Theory of Molecules
The phenomena are our data, and behind them we cannot go except in imagination. —
A. Schopenhauer.
The fundamental assumptions of the kinetic theory in its simplest form, can
now be summarized — the term " kinetic," by the way, is derived from the Greek
Kiviiti, I move. (1) Matter is composed of a finite number of molecules. In
gases, the volume of the molecules is very small compared with the space not
occupied by the molecules. At great pressures, however, the relative sizes of the
molecules must be taken into consideration. (2) The molecules of a gas are in a
state of rapid perpetual motion in straight lines. The molecules are continually
colliding against the walls of the boundary vessel and against one another. (3) The
molecules are perfectly elastic and rebound after a collision without any loss of
momentum. (4) The molecules of gases do not always move quite independently
of one another, since some molecules have a slight attractive force one for the other.
This becomes appreciable with increasing concentrations. (5) Two gases are in
thermal equilibrium when the average kinetic energies of the molecules of both
gases are the same.
A. D. Risteen in his Molecules and the Molecular Theory (Boston, 1895), has
compared the results of observation with the deductions from tlie kinetic theory
in double columns. The following is modified from his scheme :
766 INORGANIC AND THEORETICAL CHEMISTRY
Table VII. — Comparison of the Kinetic Theory with Facts.
Sesults of theory.
1. The molecules of a particular gas are
all alike. There are special cases of
dissociation and polymerization.
2. Molecules are at relatively great distances
apart, and in constant motion in
straight lines.
3. In a given mass of molecules, the product
pv is proportional to the average ki-
netic energy per molecule.
4. The average kinetic energy is constant for
every set of molecules in a mixture of
esses .
5. If two sets of molecules have the same
kinetic energy, and the same pressure,
they contain the same number of mole-
cules per unit volume.
6. Diffusion.
Hesults of observation.
1 . Gases are homogeneous and show no signs
of settling, nor can the molecules of any
particular gas in general be separated,
by diffusion, into different molecules.
2. The compressibility, permeability, and
diffusivity of gases is great. The in-
compressibility of gases at high pressures
is supposed to be due to the abnormal
crowding of the molecules.
3. In a given mass of gas the product pv is
proportional to the absolute tempera-
ture, etc. This includes the laws of
Boyle, Dalton, and Charles.
4. So far as we can tell, the temperature of
each constituent of a mixture of gases is
the same.
5. Avogadro's hypothesis, and hence also
Gay Lussac's law. This is not a result
of observation, but it has been inferred
independently from purely chemical
phenomena.
6 Graham's law.
Molecular magnitudes. — The following Table VIII summarizes the results of
some preceding calculations for a few common gases :
Table VIII. — Magnitudes in the Molecular World.
Gas.
Mole-
cular
weight,
M.
Mean velo-
city F, at
0°, in cm.
per sec.
Number of
collisions
per sec.
Average free
path, L, at
0° and 106
bars cm.
16-00 X 10-«
Molecular
diameter,
d cm.
Mass of the
molecules,
m grm.
Hydrogen, Hg .
■
2-016
1-838x105
10-6x108
2-403 X 10-8
13-33x10-24
Oxygen, Og
32-00
0-461
4-7
9-05
2-975
52-78
Nitrogen, Ng
28-02
0-493
5-3
8-50
3146
46-53
Argon, A .
39-88
0-413
4-2
8-98
2-876
65-79
Water, HgO
18-02
0-615
6-3
8-31
2-900
29-73
Carbon dioxide,C02
4400
0-393
6-5
5-56
3-335
72-59
The following i also represent some constants which occur in calculations :
Volimie of ideal gas per gram molecule at 0°, 760 mm. . 22,412 c.c.
Number of molecules per gram molecule at 0°, 760 mm. 6'062 x 10 ^^
Nimiber of molecules per c.c. at 0° and 760 mm. . . 2*705 x lO^'
Kinetic energy of a molecule at 0° and 10« dynes . . 5-621 x lO-^* erg
Gas constant, R, 1-987 cals. per degree, or . . . 83-15 x 10* ergs per degree
Boltzmann's gas (
5oiistant
R/N=k
.
I'C
172 X 10-i« er^
;s per degree
It is sometimes convenient to use a millionth of a metre, i.e. a thousandth of a milli-
metre, as a unit of smallness, and to represent this unit by the symbol /* ; this unit is called
a micron. In illustration, 0-001 mm. or 10"^ mm. is /x ; and thus O'Ol mm. or 10"^ mm.
will be lO/iA, and 0-0001 mm. or 10"* mm. will be O'l/x. Similarly, the double ^ or milli fi,
symbolized /x/*, and called a millimicron, represents a thousandth part or 0*001/* = uu
=0-000001 mm.
In every explanation of natural phenomena, said H. von Helmholtz, we are
compelled to leave the sphere of sense perceptions and to pass to things which are
not the objects of sense, and are defined only by abstract conceptions. It is almost
THE KINETIC THEORY OF ATOMS AND MOLECULES 767
the same with magnitudes in the molecular world. Most of the numbers repre-
senting the motions and magnitudes of molecules convey no meaning to the mind
because they are utterly beyond the range of our comprehension, and they might
almost as well be abstract conceptions. The following considerations will serve
to emphasize our inability to form a clear concept of the scale of magnitudes in the
world of molecules. They have been employed by several writers. 2
First y A normal human eye, at a distance of 10 inches, can see objects 2^x1*^ ^^^ ^^
diameter ; with a good microscope objects not much smaller than ^^y^^nr ^^^^ ^ diameter can
be clearly seen, but this is nearly 5000 times the magnitude of the molecule of an element.
It would take about 40,000,000 molecules, touching one another, to make a row an inch
long. Second, If all the molecules in a cubic inch of a gas were laid in a row, touching one
another, although they are so inconceivably small, yet they are so very numerous that
they would form a line about 35,000,000 miles long, and this line would extend more than
1000 times round the earth, and this in spite of the fact that only about one- three- thousandth
of the volume is actually occupied by matter, the remainder being vacuous space. Third,
If the gas were magnified on such a scale that a molecule was an inch in diameter, each
cubic foot would contain about one molecule, and a molecule would then travel about 100
feet before it collided with another. Fourth, It would take about 53 years, counting at the
rate of three per second, 24 hours a day, to count the number of collisions — 5,000,000,000 —
made by a molecule with its fellows every second. Fifth, A molecule travels at the rate of
nearly a quarter of a mile per second.
If the molecules occupy only a fractional part of the space taken up by a mass
of matter, it is natural to inquire : Is there absolutely nothing in the intermolecular
spaces ? Students of light, heat, electricity, and magnetism say that the inter-
molecular space, where no ponderable matter exists, is full of " an entity of a highly
rarefied nature called sether." This hypothetical medium is continually crossing
the path of the student of chemical theory.
Early history o! the kinetic theory. — The first inkling of the idea that many
of the observed properties of matter may be explained by the motion of its con-
stituent particles without the introduction of separate adventitious hypotheses as to
the nature of matter, has been traced back to the so-called atomic theory
of Leucippus (c. 450 B.C.), Democritus (c. 420 B.C.), Epicurus (c. 300 B.C.), and
Lucretius (c. 80 B.C.). Francis Bacon, in his De principiis atque originibus (London,
1612), said :
Almost all the ancients- — Empedocles, Anaxagoras, Anaximenes, Heraclitus, and
Democritus — though they differed in other respects about the first matter, agreed in this,
that they set down matter as active, as having some form, as dispensing with that form,
and as having the principle of motion in itself. Nor can any one think otherwise imless he
plainly deserts experience. ^
The theory that the properties of matter are dependent on the mode of motion of
the constituent particles was taken up by P. Gassend, in his Syntagma philosophicuin
(Lugduni Batavorum, 1658), where he explained the three states of matter by
postulating absolutely rigid atoms moving in all directions in empty space. Robert
Boyle also, in his Considerations and Ex'periynents touching the Origin of Qualities
and Forms (London, 1664), assumed the existence of a continued motion of the
primitive atoms. As previously indicated, Robert Hooke (1678) attributed the
pressure of gases to the impact of similar particles ; and the same notion occurred
independently to D. Bernoulli (1738), and to T. Herapath (1821). J. P. Joule
(1848) also applied the principle to calculate the average speed of the particles of
a mass of hydrogen. J. P. Joule's calculation is independent of the number of
particles, as well as of their direction of motion and of their mutual collisions.
These ideas did not develop into a satisfactory hypothesis until R. Clausius, in
1857 and subsequent years, took into consideration the mutual impacts, and the
internal rotations and vibrations which the molecules communicate to one another.
R. Clausius also discussed the bearing of the internal motions of the molecules on
specific heat, and he explained the comparative slowness of the process of diffusion
of one gas into another in spite of the swiftness of the motions of the molecules.
768 INORGANIC AND THEORETICAL CHEMISTRY
In 1860, J. C. Maxwell applied the statistical method or method of averages to the
distribution of velocities among the molecules of a gas, and he made the first
numerical estimate of the average length of the free path of a moving molecule
between two collisions. The work of A. Kronig (1856), R. Clausius (1857), J. C.
Maxwell (1859), L. Boltzmann (1868), and of others played an important part in
the subsequent development of the hypothesis.
The kinetic theory and the corresponding molecular theory of liquids and gases
have been of great service in helping chemists to form mental pictures of many
processes which would be otherwise too difficult to conceive clearly. No one
pretends that the picture corresponds with reality, but it has been of great assistance
in applying the method of deduction and verification. The theory has its faults ;
at present, it throws no light on many of the properties of gases, while the applica-
tions to liquids and solids have scarcely been touched. A great deal of work
remains to be done, but most of the outstanding difficulties relate to the nature
of the atoms and molecules, and do not affect the main outline of the theory. A
short time back there was a school of chemists which repudiated the kinetic theory
as an exhausted moribund hypothesis — for instance, C. L. Speyers rather prematurely
said in his Textbook of Physical Chemistry (New York, 1898) : " The kinetic theory
is a troublesome thing, and is becoming an object of ridicule." As a matter of fact,
the kinetic theory of molecules may be a troublesome thing, but it still promises
to live long when our mathematicians get strong enough to wrestle with its many
difficulties ; few have any doubts as to the validity of the essential features of the
doctrine.
References.
1 R. A. Millikan, Proc. Nat. Acad. Sciences, 3. 314, 1917.
•2 G. J. Stoney, Proc. Boy. Dublin Soc, (2), 8. 351, 1895; A. D. Risteen, Molecules and the
Molecular Theory, Boston, 149, 1895; J. Becqueral, Scient. Amer. SuppL, 87- 260, 1919.
§ 7. Ultramicroscopic Particles — Ultramicroscopy
The sun discovers atoms and makes them dance naked in his beams.- — D. Culverwell.
In practice, a good microscope will not clearly resolve particles much smaller
than 0*00025 mm. (0"25fA/x, or 2*5x10-5 cm.) in diameter by direct illumination ;
with oblique illumination, using a naphthalene monobromide immersion lens, and
violet light,' particles 0'000012 mm. in diameter have been noted. The term
ultramicroscopic particles is applied to granules smaller than the limits of a
good microscope. The ultramicroscopic particles cannot be seen with a powerful
microscope illuminated in the ordinary manner, because the light waves bend
round the minute particles and enter the eye just as if the particle did not exist.
If the particles be illuminated with a lateral beam of light, their very smallness
enables them to scatter the light, so that their presence can be inferred from the
fact that each particle is surrounded by visible diffraction rings, or halos of light,
just as surely as the presence of smoke indicates fire. The motes dancing in a
beam of sunlight would be invisible but for this phenomena. The diffraction
rings which surround the particles in the track of a beam of sunlight make them
appear as if they were self-luminous, and they are more clearly seen against a dark
background.
Clear solutions with particles too small to be resolved by the most powerful
microscope, appear more or less opalescent when a beam of converging ' light is
focussed into the solution. A solution free from these particles would not produce
the opalescence, and such a solution is said to be optically empty. This is
the so-called Tyndall's optical test.i Air and gases generally, if quite free
from suspended particles, are said to be optically empty because the track of a
THE KINETIC THEORY OF ATOMS AND MOLECULES 769
beam of light therein is invisible. Air can be made optically empty by allowing
it to stand overnight in a glass box whose sides are smeared with, say, glycerol.
The difficulty in removing fine particles suspended in a gas is well shown by trying
to remove sulphur trioxide from air by passing air charged with this compound
through a number of wash bottles charged with water for which sulphur trioxide
has a great affinity. The sulphur trioxide will persist after passing slowly through
half a dozen ordinary washing bottles. It is also exceedingly difficult to get a
liquid optically empty so that the track of a converging beam of light in the liquid
is not visible in a darkened room. Distillation and filtration will not do. If
colloidal silicic acid be suspended in the water, and then precipitated by passing
an electric current through the solution, or if colloidal ferric or aluminium hydroxides
be precipitated in aqueous solution, the precipitate in settling catches and drags
the suspended particles down ; the supernatant liquid is then optically empty.
Precipitates which settle in a crystalline form do not clarify the liquid in the same
way as colloidal precipitates. Ordinary air must be excluded or it will again charge
the liquid with suspended particles. The work of R. C. Tolman, and others, shows
that for the range of particles in actual smoke (5 X 10""^ to 10~* cm.) and for particles
in suspension 10~* cm. upwards, the Tyndall beam becomes more intense at a
given concentration the greater the subdivision. W. Mecklenburg found that
with particles less than 10 ~* cm., the Tyndall beam increased in intensity with an
increase in the size of particles — concentration constant.
The sensitiveness of J. Tyndall's optical test (1868) has been greatly developed
by the use of the so-called ultramicroscope of H. Siedentopf and R. Zsigmondy.2
In this microscope, an intense beam of light— arc light, or, better, a beam of bright
sunlight — is focussed into the liquid under examination, so that the light enters
the liquid at right angles to the direction in which it is viewed under the microscope.
The positions of the particles then become visible as points of light against a dark
background. If transmitted light be used, the eye is dazzled by the profusion of
light, and it cannot distinguish the slight differences of brilliancy caused by the
diffraction of light by the small particles ; just as it is impossible to see the stars by
daylight. The presence of particles about 6x10"^ or 7x10-5 mm. in diameter
can be demonstrated by this mode of illumination. Although the position of the
particles can be seen, their form or shape cannot be distinguished. The relative
sizes of the particles can be roughly estimated from their relative brightness. The
ultramicroscope is therefore a microscope in which particles are illuminated against
a dark background with the strongest possible light. It does not give an image
of the object in the microscopic sense of the term, but it does give proof of the
existence of small particles with a refractive index different from the surrounding
medium. The stronger the illumination, the smaller the particles which can be
perceived, but the efficiency of the ultramicroscope is limited by the decrease in the
brightness of the particles at a rate which is proportional to the sixth power of their
diameter.
While the opalescence produced by Tyndall's optical test merely shows that a
solution contains a number of distinct individual particles in suspension, the ultra-
microscope enables the individual particles to be detected under conditions where
the most powerful microscope would fail to reveal any sign of non-homogeneity.
When viewed in the ultramicroscope, the ultramicroscopic particles appear as
glittering discs of light with a dim or dark background. A solution may thus
appear perfectly homogeneous when viewed under the most powerful microscope,
and yet appear distinctly heterogeneous when viewed under the microscope with
Tyndall's illumination.
W. Ostwald suggested that in the two-phase system, the one phase which is
finely subdivided and discontinuous is called the disperse phase ; the other phase
which is usually continuous is called the dispersion medium ; the disperse phase
may also extend tlirougli tlie dispersion medium as a kind of reticulum or network.
The degree of dispersion or tlie dispersity of a colloid refers to the state of
VOL. I. 3d
770 INORGANIC AND THEORETICAL CHEMISTRY
subdivision to which the disperse phase has been carried ; the dispersion medium
may be I. A gas, and the disperse phase (a) a liquid [e.g. cloud or mist), or (6) a solid
{e.g, dust or smoke) ; II. A liquid and the disperse phase (a) a gas {e.g. foam), (6) a
liquid {e.g. emulsion), or (c) a solid {e.g. suspensions) ; III. A solid, and the disperse
phase is then {a) a gas {e.g. solid foam and scoriae), (6) a liquid {e.g. certain liquid
inclusions and gels), or (c) a solid {e.g. certain solid mixtures). When the dispersion
medium and the disperse phase are both liquids, emulsions are formed if the
degree of dispersion is not high, and emulsoids if the degree of dispersion is large ;
while if the dispersion medium is a liquid and the disperse phase is solid, suspensions
are formed if the dispersity is not high, and suspensoids if the degree of dispersion
is large. The general term dispersoids covers both emulsoids and suspensoids.
H. Siedentopf and R. Zsigmondy proposed to call dispersoids which are visible
under the microscope microns, and those which can be seen only by the application
of ultramicroscopic methods ultramicrons or submicrons ; ultramicroscopic particles
which cannot be seen by ultramicroscopic methods are called amicrons. The
limiting sizes are as follows :
(Visible under microscope—Mtcrons ..... 0'25/u, or 2*5 x 10~6 cm.
rSuhmicTOfift /Electric arc light . 1 5/x/x, or 15 x lO"' cm.
Ultramicroscopic particles/ *^^^^* \ Strong sunlight . . 1-0/i^, or lO"' cm.
\ Amicrons ..... <1-0ju/li, or < 10~' cm.
Early in the nineteenth century, the products obtained by reducing solutions
of the salts were generally regarded as solutions although several investigators
believed them to be suspensions of the metals, and not solutions at all, and those
formed by cathodic reduction were considered to be hydrides, although R. L.
Ruhland (1815) 3 and J. C. PoggendorfE (1848) believed them to be metals in a very
fine state of subdivision. J. J. Berzelius (1844) said that the arsenious sulphide
obtained by the action of hydrogen sulphide is for the present to be regarded rather
as a suspension of transparent particles than a solution, for arsenious sulphide
gradually separates. out as a precipitate ; similarly, H. W. F. Wackenroder (1846)
found that the reaction between solutions of sulphur dioxide and hydrogen sulphide
furnishes a liquid from which the suspended sulphur can be separated by thawing
and freezing ; but it immediately separates out in large flocks if a neutral salt of
an alkah Uke sodium chloride be added to the acid liquid. A. Sobrero and F. Selmi
(1850) also gave an elaborate account of the same suspension of sulphur, and stated
that sulphur belongs to a class of substances which possess the power of dispersing
and dividing themselves in a liquid without completely dissolving therein — e.g.
soap, starch, and Prussian blue, and which F. Selmi (1844) classed together under
the name pseudo-solutions. J. L. Gay Lussac (1810) and W. Crum (1853) noted
the formation of a suspension during the hydrolysis of aluminium acetate ; L. Pean
de St. Gilles (1854), the hydrolysis of ferric acetate ; J. J. Berzelius (1833), E. Fremy
(1853), and H. Kiihn (1853), the formation of a solution of silicic acid by the
hydrolysis of silicon sulphide by water, and the coagulation of the solution by
alkalies. Then followed T. Graham's researches on dialysis in 1861-64.
T. Graham found that substances like potassium hydroxide, potassium sulphate,
sugar, and alcohol diffuse much more rapidly in aqueous solution than hydrated
silicic acid, dextrin, tannin, gelatin, and albumin. He found also that the former
diffuse much naore rapidly than the latter through a parchment membrane. Since
the slow diffusing substances are apt to occur in the gelatinous or non-crystalline
form. T. Graham suggested calling them colloids— from KoXXa, glue ; on the
other hand, since the crystalline salts are typical of those substances which
diffuse rapidly, T. Graham called them crystalloids. In illustration, potassium
chloride, cane sugar, magnesium sulphate, hydrochloric acid, sodium chloride, and
barium chloride are crystalloids ; while albumin, gums, starch, gelatinous aluminium
hydroxide, gelatinous ferric hydroxide, and gelatinous silicic acids are colloids.
It must not be supposed that the colloids do not pass through the parchment at all.
THE KINETIC THEORY OF ATOMS AND MOLECULES 771
T. Graham found that when the time of diffusion of hydrochloric acid — HCl — was
taken as unity, the rate of diffusion of an equal quantity of sodium chloride was
2'3, cane sugar 7, egg albumen 49, and caramel 98. On account of these great
differences, T. Graham proposed the useful method of separating substances in the
colloidal and crystalloid states. The crystalloid is removed by diffusion through
a membrane of parchment, bladder, or some similar substance. The process is
called dialysis — from the Greek 8ta, through ; A.vw, I loosen. The operation
will be understood from the following description :
A piece of parchment or bladder is bound across one end of a glass or guttapercha hoop
so as to form a kind of shallow dish, Fig. 5, narrower at the
base than the open top. A mixed solution of albumin (the
white of an egg) and potassium chloride in water is poured
into the dish. This vessel is placed in another dish, B^
containing distilled water. The water in the outer vessel is
renewed every few hours. The dish containing the mixed
solution is covered by a clock-glass to protect it from dust. In ji ^ ^Dialvzer
about three days, practically all the potassium chloride will
have passed through the membrane into the outer vessel,
while the egg albumin will remain in the inner compartment. The whole apparatus is called
a dialyzer.
T. Graham held the opinion that the distinction between a crystalloid and
colloid was due to a difference in molecular conditions, for, he said :
Crystalloids and colloids . . . appear like different worlds of matter and give occasion
to a corresponding division of chemical science. The distinction between these kinds of
matter is that subsisting between the material of a mineral, and the material of an organized
mass.
It is now believed that crystalloids and colloids are not different hinds of matter,
but rather different states of matter. W. Ostwald emphasized the idea that we
should speak rather of a colloidal state than of a colloidal substance, and that if
the phase is sufficiently subdivided it is to be regarded as colloidal phase or
colloidal state ; and he defines colloidal chemistry not as the study of colloid
materials but as that of the colloidal state of materials. T. Graham's classifica-
tion of substances into colloids and crystalloids left the wrong impression that
a colloidal substance must be amorphous, whereas it is now believed that a
phase is colloidal when it is in a sufficiently fine state of subdivision whether it
be crystalline or amorphous, and that every substance may appear under different
circumstances either in the colloidal or in the crystalline nature of the substance
concerned. P. P. von Weimarn has shown that a mere change in the concentration
of the components of a reacting system suffices to precipitate a substance in either
the crystalline or the colloidal states. W. D. Bancroft accordingly says that
colloidal chemistry differs from ordinary chemistry through variations resulting
from the increasing dispersity of one or more phases. T. Graham also distinguished
between colloids in solutions and colloids in the gelatinous form, and he applied
the term sol to the colloids when the system appeared to be liquid, and gel
when the colloids assigned a jelly-like condition. If the one component was water,
he employed the terms hydrosol and hydrogel ; if alcohol, alcosol and alcogel ; etc.
The terms sol and gel are in fairly common use. J. Perrin (1905)'* proposed
the term lyophile — ^from Xv^tv, to loosen ; <^tAos, loving — ^for those systems in
which there is a marked affinity between the two phases of a colloidal solution,
and lyophobe — <f)6f:io<;, fear — to the others ; if the colloidal dispersion medium
is water, hydrophile and hydrophobe— vSwp, water— are used. If the degree
of dispersion of a dispersoid can be increased or decreased by reversing the conditions
which brought about the change, the dispersoid is said to be a reversible COlloid,
and if this cannot be done, an irreversible colloid. These terms were intro-
duced by W. B. Hardy in 1900.
The particles which can be perceived in the ultramicroscope are more or less
approximately the same order of magnitude as the molecules themselves. For
772 INORGANIC AND THEORETICAL CHEMISTRY
instance, ultramicroscopic particles of colloidal gold, I'lfMfjL, have been measured,
and, according to C. A. Lobry de Bruyn, the estimated size of a molecule of soluble
starch in solution is 5/Lt/tx ; a molecule of chloroform is roughly 0*8/x/x, according to
G. Jager; a molecule of carbon dioxide approximates to 0'285/Lt/x; a molecule of water
vapour, O'US/x/x ; and a molecule of hydrogen gas is between 0067 and 0'159/x/x,
according to an estimate of 0. E. Meyer.^ Hence particles smaller than the com-
plex molecule of soluble starch have been perceived.
The definition o£ solutions. — Solutions are usually defined as " homogeneous
mixtures which cannot be separated into their constituent parts by filtration."
This definition forces us back to the distinction between homogeneous and hetero-
geneous mixtures ; and this, in turn, upon the sensitiveness of the tests for
homogeneity.
S. E. Linder and H. Picton ^ filtered arsenious sulphide suspensions through porous
earthenware ; and foiuid that while particles over a certain size were arrested, others
passed through unchanged. C. Barus tried to estimate the size of the disperse phase in
a silver suspension from the pore size of the plate which just permitted filtration, and
H. Bechhold elaborated the principle by preparing a graduated series of filters by impreg-
nating filter paper, wire gauze, or fabric with a solution of collodion in acetic acid or gelatin
in water, and subsequently hardening the solid. The filters for ultrafiltration, as it is called,
were then standardized with a solution of haemoglobin.
A solution may appear clear and homogeneous ; the particles in solution may
not be separable by the ordinary methods of filtration ; and the substance in the
solution may remain suspended an indefinite time ; and yet when Tyndall's optical
test is applied, an opalescence will prove that minute particles are in suspension ;
and the ultramicroscope will enable the particles to be recognized as distinct in-
dividuals. Perfect solutions, said T. 0. Bergmann (1779), should be transparent,
but there are all possible gradations between liquids carrying rapidly settling par-
ticles in suspension, and liquids which carry particles in suspension an indefinite
time without settling, and in which the particles are so small that they can only
just be perceived by the ultramicroscope. Consequently, if the above definition
of a solution be accepted, every time the sensitiveness of the method for detecting
non-homogeneity is increased, a certain number of solutions previously classed
as homogeneous will probably appear heterogeneous or colloidal. This difficulty
can be partially overcome by restricting the term solution, by an arbitrary con-
vention : Solutions are mixtures which appear clear and homogeneous with
Tyndall's illumination, and which cannot be separated into their constituent
parts by ordinary mechanical processes of filtration through paper or settling.
J. W. Gibbs' concept of phase and component helps to clarify and generalize
the distinction between chemical and physical action and between colloidal and
true solutions. The idea underlying J, W. Gibbs' concept of a phase involves a
distinction between molecular magnitudes and matter en 7nasse, and it certainly
holds good for matter which is homogeneous so long as it is not reduced in bulk
to molecular dimensions. J. W. Gibbs' phase-concept serves as a criterion for a
sharp classification of systems until the colloidal solution is reached, and then
difficulties occur. In order to emphasize his opinion that the study of colloidal
solutions renders it impracticable to draw even an arbitrary line between molecular
and molar magnitudes, W. Ostwald advocates the use of the term dispersoid
system in place of solution, and he appHes the general term dispersoids in the
following manner :
Size of particles. Examples.
iOver O'lfx .... Suspension, emulsions, etc.
Between O'lfi and 1/^^ . . Colloidal solutions
About 1 /x,/x or less . . . Molecular (and ionic) solutions
because he also believes that suspensions, colloidal and true solutions, represent
varying degrees of dispersion of the solute. Although colloidal solutions are usually
considered to be two-phase systems, and ordinary solutions one-phase systems,
THE KINETIC THEORY OF ATOMS AND MOLECULES 773
there appears to be an unbroken continuity between the heterogeneity of suspensions
and the homogeneity of true solutions. Hence W. Ostwald tried to emphasize
this by calling dispersoids with a degree of dispersion greater than 6 X 10' rnolecular
dispersoids ; and molecules may ionize — ionically disperse — to form ionic dis-
persoids. The evidence is therefore pointing to the inference that the distinction
between colloidal two-phase solutions and ordinary one-phase solutions turns
on the relations between the dispersoids or solute and the solvent rather than
on the size of the particles.
Colloids have but a slight influence on the vapour pressure, freezing point, and
boiling point of the dispersion medium. Indeed, when the colloid has been purified
to a high degree, the colloid has no influence on these properties at all. For example,
E. Paterno found that tannic and gallic acids form colloidal solutions with water
and do not appreciably influence the freezing point of the solvent, but in glacial
acetic acid, a true solution is formed, and the freezing point is depressed in the
regular manner. The molecular weights of colloidal solutions have quite a different
meaning from the results obtained with the so-called true solutions. Apart
altogether from disturbances due to absorbed impurities, and possible hydro-
lytic changes, the different results obtained with systems having the same com-
position but varying degrees of dispersion, show that a series of progressively varying
molecular weights can be obtained for one and the same substance which, in the
systems with a low degree of dispersion, would approach infinity. In some cases
where a change has been observed, we are by no means sure that some compounds
other than the assumed colloid are not producing the change. For instance, it is
highly probable that hydrolysis in aqueous solutions of sodium palmitate, or oleate,
results in the formation of some sodium hydroxide. It must also be remembered
that it is exceedingly difiicult to get rid of impurities which are associated with
colloids during their preparation. Again, said W. D. Bancroft, if the suspended
colloidal particles are insoluble in the dispersion medium, there will be no change
of surface tension, no osmotic pressure, and no diffusion except what is due to
the Brownian movements ; and conversely, if the particles are slightly soluble,
there will be a slight change of surface tension, some osmotic pressure, and some
diffusion other than that due to the Brownian movements.
Many attempts have been made to calculate the molecular weight of colloids
from the osmotic pressure calculated in this and analogous methods, and numbers
like these have been reported. Table IX, for example.
Table IX. — Osmotic Pressure and Molecular Weights or Colloids.
Egg albumin
Gelatin .
Starch iodide
Dextrin
Gum arabic
Concentration
grams per litre.
12-5
12-5
30-0
100
10-0
Osmotic pressure
mm. of mercury.
20
6
15
165
72
Molecular weight.
11,000
36,000
34,000
1,000
2,400
Some confusion is here prevalent because the term molecular weight is imparted
to two totally different concepts. Molecular weight means something quite different
in the case of gas or a solute, and in the case of a suspended particle. The molecular
weight of a lump of, say, metal is not generally supposed to be high because the
metal does not diffuse and is visible ; why then should it be permitted to assign
a high' molecular weight to the finely divided metal ? The mere fact of grinding
cannot increase its molecular weight. This has been emphasized by W. B. Bancroft 7 :
The molecular weight of a suspended particle, if it means anything at all, means some-
thing entirely different from the molecular weight of a solute or gas. We can determine
the molecular weight of benzene approximately from the measurement of the surface
774 INORGANIC AND THEORETICAL CHEMISTRY
tension, but it is absurd to say that suspending fine drops of benzene in water causes the
molecular weight of benzene to become . . . very large. We are talking about entirely
different things in the two cases. What we mean is that liquid benzene has a molecular
weight of 78, and that liquid benzene suspended in water behaves, or may behave, as if
it were a dissolved substance having a . . . very large . . . molecular weight. As a matter
of fact, the experiments bring out clearly the enormous difference between a solution and
a suspension.
Many colloids show a small osmotic pressure, and measurements of the ostnotic
pressure of colloidal solutions have been made by W. Pfeffer,^ H. Picton and S. E.
Linder, C. E. Linebarger, etc. The results show that the osmotic pressure of most
colloids is a complex function not only depending on the number of particles in
unit volume but also on the degree of dispersion, and the state of the system. The
theoretical investigations of A. Einstein and M. von Smoluchowsky show that the
osmotic pressures — Pj and P2 — of two equally concentrated but differently dis-
persed phases are inversely proportional to the cubes of the radii — ri and r^—oi
their particles ; or P^ : P2=^2^ • ^i^- Observations on the vapour pressure of
colloidal solutions have been made by A. Smits,^ F. Guthrie, C. Liideking, etc. ;
on the boiling points of colloidal solutions by F. Guthrie,!^ C. Liideking, etc. ; and
on the depression of the freezing points of colloidal solutions by H. F. Brown and
C. H. Morris,ii F. Krafft, etc.
The compressibilities of colloidal solutions have been determined by G. de Metz,i2
H. Gilbaut, etc. The compressibility of a colloidal solution is in general smaller
than that of the dispersion medium, and it decreases as the concentration increases.
The specific gravities and the specific volumes of colloidal solutions do not follow the
mixture rule. The relations are somewhat complicated, but approach the mixture law
the less the degree of dispersion of the disperse phase. W. Ostwald i^ showed that
the capillary pressure causes the specific gravity of water to increase the smaller the
size of the globulets — for the water in droplets 3 in diameter has a density 0*005
per cent, greater, and the droplets O'OS/i in diameter, 0*5 per cent, greater than
water en masse. Observations have been made by G. Rose, J. P. Cholodny, etc.
The internal friction or viscosities of colloidal solutions is very marked. The con-
centrations of inorganic colloidal solutions of the metal and sulphides are small
and the viscosities are but little different from that of water ; the case is different
with organic colloids like gelatin and albumin, and inorganic colloids like silicic
acid and hydrated alumina, where the viscosity becomes relatively large such that
T. Graham called the viscosimeter a colloidoscope. The viscosities have been
measured by J. Friedlander,!^ H. W. Woudstra, etc., and the effects of changes in
concentration, temperature, and age have been investigated. The effect of the addition
of other substances in the solution has also been determined. A. Einstein deduced
the viscosity formula tj =rj q{1-\-2'6v), where v is the volume of colloidal matter per
unit volume of medium. It is assumed that the volume v is not too large, that is,
the concentration is not too great ; the radius of the particles assumed to be rigid
spheres is also supposed to be large compared with the range of molecular attraction.
Values for the constant ranging from lb to 4*75 have been obtained. The surface
tensions of colloidal solutions have been measured by Lord Rayleigh,!^ G. Quincke,
etc. The diffusibility of colloidal solutions was measured by T. Graham, ^^
T. Voightlander, etc. The dialysis of colloidal solutions has been examined by
T. Graham, 17 G. Malfitano, R. P. van Calcar, etc. The coefficients of thermal expan-
sion of colloidal solutions have been investigated by H. Rodenwald.i^
The settling 0! particles suspend^ in water. — It is well known that
the finer the grain-size of a given substance the slower the settHng in still
water. G. G. Stokes 1^ has shown that in a medium of specific gravity D\ and
viscosity rj, the rate of settling F of a spherical grain of measurable size and of radius
r and specific gravity D,
Rate of settling, F=?/-M^=^
9 77
THE KINETIC THEORY OF ATOMS AND MOLECULES 775
where g is the gravitation constant. This formula assumes that the resistance a fluid
opposes to the descent is proportional to the radius of the sphere and to the coefficient
of viscosity of the fluid. E. Cunningham showed that with microscopic or ultra-
microscopic particles, the resistance no longer depends exclusively on the coefficient
of viscosity, and it is necessary to take into consideration resistance elements
borrowed from the kinetic theory of gases ; he uses the same factor as G. G.
Stokes multiplied by 1 -\-Allr, where I is the average length of path of the molecules
of gas, and r is the radius of the spherule. Hence, the Stokes-Cunningham
formula is :
Rate of settling, F=:-/^ ^( 1 +
9 7] \ r /
The coefficient A depends on the nature of the shocks between gaseous molecules
and the surface of the sphere, and it has to be evaluated experimentally, it has
values ranging between 0*815 and 1*63. This formula has been verified by
M. Knudsen and S. Weber, R. A. Millikan, J. Roux, M. Keehan, A. Schidlof, 0. W.
Silvey, etc.
A particle of radius ^ settles in still water at the rate of 2'4: mm. per minute,
and a particle of radius 10/xjLt and specific gravity 3, settles at the rate of one mm.
per month. The formula does not apply to angular grains, nor to excessively
fine grains. P. Ebell found that ultramarine particles, when reduced to a very
fine state of subdivision by grinding, remain suspended in water for months ;
0. Miihlhauser found that similar remarks apply to carborundum powder ; and
some clays freed from soluble salts will remain in suspension an indefinitely long
time. There is then some disturbing factor which neutralizes the action of gravity ;
the clue to this was given by some observations by the botanist R. Brown in 1827.
He said :
While examining the form of the pollen grains from Clarckia pulchilla suspended in
water, ... I observed many of them very evidently in motion. . . . Their motions were
such as to satisfy me, after frequently repeated observation, that they arose neither from
currents in the fluid nor from its gradual evaporation, but belonged to the particle itself . . .
smallest moving particles observed and which I have termed Active Molecules, appear to be
spherical, or nearly so, and to be between ^■^^^y^fth and ^troxnrth of an inch in diameter (about
0*001 mm.). The molecules were not limited to organic bodies. ... In every m'ineral
which I could reduce to a powder, sufficiently fine to be temporarily suspended in water,
I found these molecules more or less copiously, and in some cases, more particularly in
siliceous crystals, the whole body submitted to examination appeared to be composed of
them. . . . By reducing a drop of water to microscopic minuteness ... by shaking or
triturating water with almond oil . . . the motion of the smaller particles (of water) takes
place with undiminished activity.
The bearing of R. Brown's observations on the phenomenon was not appreciated
for many years. The phenomenon is named the Brownian movement after its
discoverer R. Brown. 20 If water in which a little lycopodium — that is, the spores of
the club moss — is suspended be examined under the microscope, the small particles
appear to be incessantly vibrating with a slow trembling motion.
The phenomenon is demonstrated as follows : Rub a fragment of gamboge for a moment
on an ordinary 3x1 glass slip, and place a couple of drops of water on the slip where the
gamboge has been rubbed. Gently push a cover-glass up to the edge of the gamboge.
The brisk motion of the particles can now be readily seen through a ^" objective and a
dark -ground illumination ; it is easy to seal up the liquid in little capillary tubules about half
an inch long. These tubules can be mounted in Canada balsam under a cover-glass in the
usual manner, and they are then available for showing the phenomenon at any time.
There are references to the " dancing particles of rudimentar}^ animalcules "
in the writings of some of the earlier naturalists, which show that the phenomenon
had been noticed before Brown's time, but without a clear idea of the nature of
the particles. Fig. 6 gives a sketch from a photograph of the Brownian movement 21
776 INORGANIC AND THEORETICAL CHEMISTRY
prepared by moving a photographic film at a uniform rate across the field of the
microscope from left to right. Experiment has shown that the motion cannot be
due to convection currents set up by small differences of temperature or pressure,
to light effects, to the electrical state of the particles or of the liquid, nor indeed to any
known influence outside the liquid. The cause of the motion must be sought in the
liquid itself.
The kinetic theory of molecular motion seems to furnish the only admissible
explanation of the phenomenon. It is supposed that, owing to the perpetual move-
ments of the molecules of the fluid, the moving molecules are continually striking
the particles, thus driving them irregularly, to and fro, up and down, in the liquid.
As might be expected, the incessant movements become more and more vigorous
the smaller the particles. There is a big contrast between the apparently sluggish
movements of lycopodium and the vivacious motions of the ultramicroscopic par-
ticles of, say, M. Faraday's gold.22 ^q much so that R. Zsigmondy once thought the
two were really different phenomena. The movements of the particles of gold are
described by R. Zsigmondy somewhat as follows : "The particles move with astonish-
ing rapidity. A swarm of gnats in a sunbeam will give an idea of the motion. The
particles hop, dance, and jump, and they dash together and fly away from one another
so that it is difficult to get one's bearings." To this must be added the fact that if
the composition of the liquid remains unchanged, the motion in the liquid seems
to continue an indefinite time without ceasing, for the Brownian movements have
been observed in the fluid in cavities of quartose rocks, showing that the motion has
in all probability been main-
tained for enormous periods
of time^ — ^ever since the fluid
was sealed up in the rocks.
Fig. 6.-Sketch from a Photograph of the Brownian % studying the move-
Movement (T. Svedberg, 1912). ments of the fine particles of
gamboge — that is, the dried
latex of the Garcinia morella — suspended in water ; and of extremely fine particles
of silver dust obtained by striking an electric arc between silver poles — suspended
in air ; it has been proved by direct observation 23 that the distribution of the
particles, their velocities, and the frequency of their collision is the same as the
kinetic theory assumes to be the case with the particles of a gas. F. M. Exner
found that particles larger than 4/x do not show the Brownian movement in water ;
particles smaller than 0'1/x show vigorous movements ; and particles with a diameter
10/x/x have trajectories up to about 20/x/x. The movement takes place in all
fluids, though more actively the less the viscosity. J. Perrin detected it with par-
ticles suspended in the film of a soap bubble. It is just perceptible in glycerol,
and very active in gases. All attempts to find an explanation of the Brownian
movement outside the fluid have failed. As C. Wiener 24 stated in 1863, the agitation
does not originate either in the particles themselves or in any cause external to the
fluid, but must be attributed to internal movement characteristic of the fluid state ;
and J. Perrin adds that the apparent repose of a fluid in equilibrium is an illusion
due to the imperfection of our senses ; in reality the constituent particles of a fluid
are in a state of spontaneous, unco-ordinated, and never-ceasing agitation. A par-
ticle of a solid suspended in a fluid is being continuously bombarded by the neigh-
bouring molecules. W. Ramsay (1876), G. Gouy (1888), H. Siedentopf (1900),
and J. Delsaulx and J. G. G. Carbonelle (1877) had a clear view of the mode in which
the molecular bombardment of the fine particles suspended in a fluid can produce
the Brownian movement :
The internal movements of the molecules which determine the heat content of a fluid,
is well able to explain the facts. ... In the case of large surfaces, the molecular impacts
which cause pressure produce no displacement of the suspended body because the resultant
tends to move the body in all directions at once ; but if the surface of the suspended body be
smaller than the area necessary to ensure that all the irregular motions will be compensated,
THE KINETIC THEORY OF ATOMS AND MOLECULES 777
the preasures from different sides will be unequal, and continually shifting from point to
point ; these pressures will not be equalized by the law of averages, and their resultant
being no longer zero, they will vary continuously in intensity and direction.
A. Einstein (1905) and M. von Smoluchowsky (1906) developed a quantitative
theory of the phenomenon. The former showed that an emulsion diffuses like a solu-
tion, so that if x^ denote the mean square of the projection of the displacement
along a horizontal axis, the quotient 'X^jt is constant when t denotes the time, so that
the mean displacement is doubled when the time is increased four-fold ; and in-
creased ten times when the time is increased a hundred-fold. The quotient x^jt
is called the activity of the Brownian movement of the granules. A. Einstein next
showed that the coefficient of diffusion, 8=^x^lt ; and further, if N denotes Avo-
gadro's number, the
RT , a;2 RT
Coefficient of diffusion, o= ;n^ ; or, Activity of movement, =
QNnrr]' t SNirrrj
but the mean kinetic energy K of the molecules is ^RT/N, and therefore x^/t
^IKIirrri. This means that the activity of the agitation (or the rate of diffusion) is
proportional to the molecular energy or the absolute temperature, and inversely
proportional to the viscosity and absolute temperature of the grains. This formula
has been verified by L. Bancelin for sugar solutions, and for emulsions of gamboge
in solutions of glycerol. V. Henri has discussed the different methods available
for determining the grain size of the colloidal particles.
T. Svedberg showed that the amplitude A or the distance between the extreme
and mean positions of the particle, and the period of oscillation t or the time taken
by the particle to make a complete oscillation or excursion from one extreme
position and back again, is related with the velocity V, by the equation V—^Ajt.
T. Svedberg found that as the amplitude increases the period of oscillation increases
in such a way that ^Ajt is nearly constant,
Acetone. Ethyl acetate. Amyl acetate. Water, n-propyl alcohol,
A 3-1 2-0 1-5 1-1 0-7/x
t 0-032 0-028 0026 0-013 0-009 sec.
4^/« 390 290 230 340 310
F. M. Exner showed that with particles 0"4:/x in diameter, the velocity was 3'8/x
per sec. ; with particles l"3/x in diameter, the velocity was 27/x per sec. ; with
particles 3/x in diameter the velocity was just perceptible, and with particles 4jLt
in diameter no movement was discernible ; and K. Zsigmondy showed that the
amplitude depends on the size of the particles ; with particles 6, 10, and 35/>t/x
in diameter, the amplitudes were over lO/u,, 3 to 4ju,, and 1 to 7/x respectively.
T. Svedberg further showed that if the sizes are constant the viscosity of the medium
is alone of importance ; and that neither the nature of the particles nor the other
properties of the medium are of importance. For any given size of particles, the
amplitude is approximately inversely proportional to the viscosity of the dispersion
medium.
A . . . .
7yxl03
J. Perrin25 experimented on the arrangement of the particles of a colloid suspended
in water under the action of gravity and its own osmotic pressure. If n and n^
respectively denote the number of particles in unit volume at heights 0 and h ;
IV, the mass ; D, the density of the granules ; D' , that of the liquid ; g, the gravitation
constant ; and P, the osmotic pressure of a single particle in unit volume, then,
according to J. Perrin,
Acetone.
Ethjl acetate.
Amyl acetate.
Water.
n-propyl alcohol.
3-1
2-0
1-5
1-1
0-7/x
3-2
4-6
5-9
10-2
22-6
9-9
9-2
8-9
11-2
15-8
, n tvqh/\ Z)'\
778 INORGANIC AND THEORETICAL CHEMISTRY
The agreement of the calculated with the observed results led J. Perrin to conclude
that the concentration of the granules in a uniform emulsion decreases in an exponential
manner as a function of the height in the same way as the barometric pressure does as
a function of the height. Once this relation is established, the same equation
affords a means of evaluating Avogadro's constant, N — for example, the number
of molecules per gram-molecul^ of gas at 0° and 760 mm. Assuming that the
pressure exerted by the particle is analogous with that as a gas obeying Boyle's
law, P=RTjN, numbers ranging from iV=5xl023 to 8x10"^^ were obtained
— the generally accepted value determined by other independent methods is
iV=6xl023.
The perpetual agitation which is illustrated by the Brownian movement proceeds
ceaselessly without external cause ; it has therefore been asked if this motion
contradicts the principle of the conservation of energy. The principle would be
satisfied if every increment in velocity acquired by a granule is accompanied by the
cooling of the liquid in its immediate vicinity, and every diminution in velocity is
accompanied by local heating. J. Perrin tried to get round the difficulty by pointing
out that J. C. Maxwell, W. Gibbs, and L. Boltzmann have robbed Carnot's principle
of its claim to rank as an absolute truth, and reduced it to the mere expression of a
very high probability. He adds that on the scale of magnitudes that are of practical
interest to us, perpetual motion of the second kind is in general so insignificant
that it would be foolish to take it into consideration. The impression left by this
argument is that the principle is valid on the scale of magnitudes that are practical
to us, but not in the realm of molecular magnitudes.
The Brownian movement is an expression of the molecular movements usually
attributed to the molecules of matter. The molecular bombardment of the particles
suspended in an emulsion tends to give a uniform distribution. Now, said J. Perrin,
in his Les preuves de la realite moUculaire^
Let us suppose that it is possible to obtain an emulsion, with the granules all identical,
an emul^on which I shall call, for shortness, uniform. It appeared to me at first intuitively,
that the granules of such an emulsion should distribute themselves as a function of the
height in the same manner as the molecules of a gas under the influence of gravity. Just
as the air is more dense at sea-level than on a mountain-top, so the granules of an emulsion,
whatever may be their initial distribution, will attain a permanent state where the concen-
tration will go on diminishing as a function of the height from the lower layers, and the law
of rarefaction will be the same as for the air.
J. Perrin confirmed this hypothesis with suspensions of gamboge and mastic in water ;
with the former, particles 0"3/Lt in diameter, a rise of 30/x sufficed to lower the
concentration to one-tenth its value. J. Perrin found that unless the mixtures
are aseptic, they may be invaded by elongated and very active protozoa, which, by
stirring up the emulsion like fishes agitating the mud of a pond, much diminish the
inequality of distribution between the upper and lower Wers. H. Zangger 26 likewise
confirmed the theory for drops of mercury ; L. M. Brillouin, for emulsions of gamboge
in glycerol ; and B. Hjin, for emulsions of gamboge in water or castor oil, and an
aqueous solution of gum arable. B. Hjin also found that the movements of a given
particle are quite independent of the movements of other particles.
A. Einstein and M. von Smoluchowsky worked out a mathematical kinetic
theory of the process, and as a result of the close agreement between theory and
observations, J. Perrin was led to say in his Les atomes (Paris, 1913), and in his
Les preuves de la realite moleculaire (Paris, 1909) :
The objective reality of the molecules becomes difficult to deny. At the same time,
molecular movement has not been made visible. . . . The Brownian movement offers
us on a different scale the faithful picture of the movement possessed, for example, by the
molecules dissolved in the water of a lake which, encountering one another only rarely,
change their direction and speed by virtue of the impacts with the molecules of the solvent.
. . . The Brownian movement is a faithful reflection of molecular movement, better, it is a
molecular movement in itself, in the same sense that the infra-red vibration is still light. From
THE KINETIC THEORY OF ATOMS AND MOLECULES 779
the point of view Of agitation, there is no distinction between nitrogen moleciiles and the
visible molecules realized in the grains of an emulsion, which have a gramme molecule of the
order of 100,000 tons. Thus, as we might have supposed, an emulsion is actually a miniature
ponderable atmosphere ; or, rather, it is an atmosphere of colossal molecules, which are
actually visible. The rarefaction of this atmosphere varies with enormous rapidity, but it
may nevertheless be perceived. In a world with such an atmosphere, Alpine heights might
be represented by a few microns, in which case individual atmospheric molecules would be
as high as hills.
In fine, the experimental facts go very near towards establishing the validity
and essential reality of the molecular kinetic theory as an explanation of the
properties of matter. In the words of E. E. Fournier d'Albe :
We are face to face with this extraordinary situation : the molecule has ceased to be a
theoretical abstraction- — ^it has become a visible and tangible reality ; for we can not only
see it, but also " manipulate " it- — not, indeed, with our hands, but by means of heat, and
electricity, and the air pump.
Fine particles^ — say less than 0'5/x — will remain in suspension an indefinitely
long time, presumably because the Brownian movements tend to distribute the-
particles through the liquid against the action of gravity. If, however, enough
particles agglomerate or coalesce so as to form large aggregates, settling may ensue.
The aggregation, flocculation, or clotting of the fine particles and the converse dis-
persion, or deflocculation of aggregates is of great importance in many chemical
processes, and will be discussed later — see purple of Cassius, and colloidal gold.
References.
1 J. TyndaU, Proc. Roy. Soc, 17. 223, 1869 ; Proc. Roy. Inst., 6. 365, 1871 ; Phil. Mag.,
(4), 37. 384, 1869 ; R. C. Tolman and E. B. VHet, Journ. Amer. Chem. Soc, 41. 297, 1919 ;
R. C. Tolman, L. H. Reyerson, E. B. Vliet, R. H. Gerke, and A. P. Brooks, ib., 41. 300, 1919 ;
R. C. Tolman, R. H. Gerke, A. P. Brooks, A. G. Herman, R. S. Mulhken, and H. de W. Smyth,
ib., 41. 575, 1919 ; W. Mecklenburg, Zeit. Roll., 16. 97, 1915.
2 H. Siedentopf and R. Zsigmondy, Ann. Physik, (4), 10. 1, 1903 ; R. Zsigmondy, Zur
Erkenntnis der Kolloide, Jena, 1905 ; New York, 1909 ; A. Cotton and H. Mouton, Les uUrO'
microscopes : Les objets uUramicroscopiques, Paris, 1906.
3 T. Graham, Phil. Trans., 151. 183, 1861 ; Journ. Chem. Soc, 15. 216, 1862 ; 17. 318, 1864 ;
W. Ostwald, Grundriss der Kolloidchemie, Dresden, 79, 1909 ; J. J. Berzelius, Lehrbuch der
Chemie, Dresden, 2. 122, 1833; H. W. F. Wackenroder, Brandes' Archiv., 47. 272, 1846;
48. 140, 272, 1846 ; Liebig's Ann., 60. 189, 1846 ; Liebig's Ann., 60. 189, 1846 ; A. Sobrero and
F. Selmi, Ann. Chim. Phys., (3), 28. 210, 1850 ; E. Fremy, ib., (3), 38. 312, 1853 ; J. L. Gay
Lussac, ib., (1), 74. 193, 1810 ; H. Kiihn, Journ. prakt. Chem., (1), 59. 1, 1853 ; W. Crum, Journ.
Chem. Soc, 4. 216, 1853 ; L. Pean de St. Gilles, Compt. Rend., 40. 568, 1843, 1854 ; R. L. Riihland,
Schweigger's Journ., 15. 411, 1815; J. C. Poggendorf, Sitzber. Akad. Berlin, 249, 1848;
F. Selmi, Ann. Fis. Quim., 13. 157, 233, 1844.
* J. Perrin, Ann. Chim. Phys., (8), 3. 84, 1905 ; W. B. Hardy, Zeit. phys. Chem., 33. 326,
385, 1900 ; W. Biltz, Ber., 37. 1096, 1904 ; P. P. von Weimam, Zur Lehre von den Zustdnden
der Materie, Dresden, 1914 ; W. D. Bancroft, Journ. Phys. Chem., 18. 549, 1914.
« C. A. Lobry de Bruyh, Rec Trav. Chim. Pays- Bas, 19. 251, 1900; G. Jager, Sitzber. Akad.
Wien, 100. 1233, 1891 ; O. E. Meyer, Die kinetische Theorie der Case, Breslau, 1899; London,
1899; W. Mecklenburg, Die experimentelle Grundlegung der Atomislik, Jena, 1910.
6 S. E. Linder and H. Picton, Journ. Chem. Soc, 61. 148, 1892 ; A. Schop, Bull. Soc Chim.
Belgique, 24. 354, 1910 ; C. Barus, Amer. Journ. Science, (4), 48. 51, 1895 ; H. Bechhold, Zeit.
phys. Chem., 60. 257, 1907 ; 64. 328, 1908 ; G. Bredig, Anorganische Fermente, Leipzig, 27,
1901 ; J. Duclaux, Zeit. Roll, 3. 134, 1908 ; C. J. Martin, Journ. Physiol, 20. 364, 1896 ; E. W.
Reid, ib., 27. 161, 1903 ; A. Craw, Proc Roy. Soc, 77. 172, 311, 1899.
' W. D. Bancroft, Journ. Phys. Chem., 22. 273, 1918.
8 W. D. Bancroft, Journ. Phys. Chem., 18. 549, 1914 ; W. Pfeffer, Osmotische Untersuchungen,
Leipzig, 1877 ; H. Picton and S. E. Linder, Journ. Chem. Soc, 63. 148, 1892 ; C. E. Linebarger,
Amer. Journ. Science, (3), 43. 218, 1416, 1892 ; E. H. Starling, Journ. Physiol, 19. 312, 1896 ;
24. 317, 1899 : C. J. Martin, ib., 20. 3(54, 189() ; A. Lottermoser, Ueber anorganische Colloide,
Stuttgart, 1901 ; ZeiL phys. Chem., 60. 451, J 907: W. Biltz and A. von Vegesack, ib., 68.
357, 1909; 73. 481, 1910; B. Moore and W. H. Parker, Amer. Journ. Physiol, 7. 261, 1902;
H. E. Roaf, Jourti. Physiol, 39. 438, 1909; Quart. Journ. Physiol, 3. 75, 171, 1910;
Biochem. Journ., 3. 422, 1908 ; B. Moore and H. E. Roaf, ib., 2. 34, 1907 ; 3. 55, 1908 ; B. Moore
and D. Bigland, ib., 5. 32, 1909 ; E. W. Reid, Journ. Physiol, 31, 439, 1904 ; 33. 12, 1905 ;
780 INORGANIC AND THEORETICAL CHEMISTRY
J. Duclaux, Jouni. Chim. Phys., 5. 40, 1907 ; 7. 407, 1909 ; Koll. Zeit, 3. 126, 134, 1908 ; Compt.
Rend., 140. 14158, 1544, 1905 ; G. Maltitano, ib.y 142. 1418, 190« ; R. S. Lillie, Amer. Journ.
Physiol, 20. 127, 1907 ; W. M. Bayliss, Proc. Roy. Soc, 81. 209, 1909 ; Koll. ZciL, 6. 23, 1910 ;
J. Duclaux, Compt. Rend., 148. 714, 259, 1909 ; M. von Smoluchowski, Boltzmann's Festschrift^
G26, 1904; T. Svedberg, Bemmeleii's Gcdenkboek, 131, 1910; W. Spring, Koll. Zeit., 7. 22,
1910 ; S. Postermak, Ann. Imt. Pasteur, 15. 85. 1909 ; W. Pauli, Hofmeister's Beitr., 5. 27, 1900 ;
R. Hober, ib., 11. 35, 1907 ; W. Ostwald, Pjlivgers Arch., 108. 563. 1905 ; H. Handovsky, KolU
Zeit., 7. 192, 1910.
• A. Sraits, Zeit. phys. Chem., 45. 608, 1903 ; F. Guthrie, Phil. Mag., (5), 2. 219, 1876 ;
C. Ludeking, Wied. Ann., 35. 552, 1888 ; G. Tammann, Mem. Acad. St. Petersburg, (7), 35. 1,
1887 ; G. Bruni and N. Pappada, Rend. Accad. Lincei, (5), 9. i, 354, 1901 ; Gazz. Chim. Ital.y
31. 244, 1901 ; J. M. van Bemmelen, Die Absorption, Dresden, 1910.
10 F. Guthrie, Phil. Mag., (5), 2. 211, 1876; C. Liideking, Wied. Ann., 35. 552, 1888;
A. Sabanejeflf, Journ. Russian Phys. Chem. Ges., 22. 102, 1890 ; S. E. Linder and H. Pieton,
Journ. Chem. Soc, 61. 114, 1892; A. Lotterraoser, Ueber anorganische Colloide, Stuttgart, 74,
1901 ; F. Kraflft, Ber., 29. 1328, 1896.
11 H. T. Brown and C. H. Morris, Journ. Chem. Soc, 53. 1610, 1888 ; S. E. Linder and
H. Pieton, ib., 61. 114, 1892; J. Gladstone and W. Hibbert, Phil. Mag., (5), 28. 38, 1889;
A. Sabanejeff, Journ. Russian Phys. Chem. Soc, 21. 515, 1889 ; 22. 162, 1890 ; Ber., 23. 87, 1890 ;
24. 558, 1891; F. Krafft, ib., 32. 1614, 1899; E. Paterno, Zeit. phys. Chem., 4. 475, 1889;
G. Tammann, ib., 20. 180, 1896 ; N. Ljubavin, Journ. Russian Phys. Chem. Soc, 21. 397, 1889 ;
A. Sabanejeff and N. Alexandroff, ib., 23. 7, 1891 ; C. E. Lincbarger, Amer. Journ. Science^
(3), 43. 416, 1892 ; W. Meyer, Zur Kenntnis einiger anorganische Kolloidsubstanzen, Halberstadt,
1897 ; St. Bugarskyand L. Lieberman. Pfliiger's Arch., 72. 51, 1898 ; St. Bugarsky, L. Lieber-
man, and F. Tangl, ib., 72. 531, 1898; Z. Gatin-Gruszewska, ib., 102. 569, 1904; F. Bottazzi
and G. d'Enrico, ib., 115. 359, 1906 ; H. Freidenthal, Physiol. Zentralbl, 12. 849, 1899 ;
A. Lottermoser, Ueber anorganische Colloide, Stuttgart, 74, 1901 ; N. Pappada, Gazz. Chim. Italy
32. ii, 22, 1902 ; G. Bruni and N. Pappada, ib., 31. i, 244, 1901 ; W. R. Whitney and J. Blake,
Journ. Amer. Chem. Soc, 26. 1339, 1904 ; G. Malfitano and L. Michel, Compt. Rend., 143. 1141,
1907 ; J. Duclaux, ib., 148. 295, 1909 ; T. B. Robertson and T. C. Burnett, Journ. Biol Chem.,
6. 105, 1909 ; G. Moruzzi. Bioch. Zeit., 22. 232, 1809.
12 G. de Metz, Wied. Ann,, 35. 497, 1888 ; G. Quincke, ib., 19. 401, 1883 : E. H. Amagat,
Ann. Chim. Phys., (5), 11. 535, 1877 ; C. Barus, Amer. Journ. Science, (4), 6. 2^5, 1898 ; (3), 41.
110, 1891 ; W. Rontgen and J. Schneider, Wied. Ann., 29. 165, 1886 ; H. Gibbaut, Zeit. phys.
Chem., 24. 385, 1897 ; H. Guinchant, Compt. Rend., 132. 469, 1901.
1' W. Ostwald, Grundriss der allgemeinen Chemie, Leipzig, 533, 1910 ; G. Rose, Pogg. Ann.,
73. 1, 1848 ; J. P. Cholodny, Koll 'ZeiL, 2. 19, 340, 1907 ; B. Loffler, Ann. Physik, (4), 32. 3,
1907 ; H. Quincke, Pflitger's Arch., 3. 332, 1870 ; W. Schmidt, Pogg. Ann., 114. 337, 1861 ;
C. Liideking, Wied. Ann., 35. 552, 1888 ; H. Rodewald, Zeit. phys. Chem., 24. 193, 1897 ; J. M.
van Bemmelen, Die Absorption, Dresden, 275, 458, 1910 ; S. E. Linder and H. Pieton, Journ.
Chem. Soc, 67. 71, 1895 ; G. Quincke, Ann. Physik, (4). 9. 800, 1902 ; (4), 10. 486, 809, 1903.
1* A. Einstein, Ann. Physik, (4), 19. 289, 1906 ; J. Friedlander, Zeit. phys. Chem., 38. 430,
1901 ; H. W. Woudstra, ib., 63. 619, 1908 ; Chem. Weekb., 5. 303, 1908 ; Bemmelen's Gedenkboek,
36, 1910 ; F. Bottarri and G. d'Enrico, Pflilger's Arch., 115. 359, 1906 ; L. Gatin-Gruszewska,
ib., 102. 569, 1904; W. Ostwald and A. Genthe, Zool Jahrb., 18. 1. 1913; E. Hatschek,
Koll Zeit., 7. 301, 1910; S. J. Levites, ib., 2. 210, 239, 1907; M. Gokun, ib., 3. 84, 1908;
W. Pauli, H. Hawdovskv, K. Schorr, R. Wagner, M. Samec, ib., 3. 2, 1908; 7. 241,
1910; Hofmeister's Beitr.," W. 415, 1908; Biochem. Zeit., 27. 296, 1910; G. Moruzzi, ib.,
22. 232, 1909; P. von Schroder, ZeiL phys. Chem., 45. 75, 1903; W. Fleming, ib., 41.
407, 1902 ; V. Henri, S, Lalou, A. Mayer, G. Stodel, Compt. Rend., Soc Biol, 55. 1668,
1903; H. Garrett, Phil Mag., (6), 6. 374. 1903; A. Miiller, Ber., 37. 11, 1904; E. Laqueur
and O. Sackur, Hofmeister's Beitr., 3. 193, 1903; W. B. Hardy, Journ. Physiol, 33. 251,
1905; Proc Roy. Soc, 79. B, 413, 1907; G. Fano and G. Rossi, Arch. Fisiol, 1. 492,
609, 1904; G. Rossi, ib., 2. 272, 500, 599, 1905; 3. 171, 1906; G. Rossi and O. Scarpa,
ib., 2. 246, 1905; E. Cavarrani, ib., 2. 513, 1905; J. Simon, ib., 4. 594, 1907; 5. 394,
402, 477, 479, 1908; S. Axelrod, Gummizeit., 19. 1053, 1905; 20. 105, 1905; 23. 810,
1909 ; C. O. Weber, Chemistry of India-rubber, London, 80, 1902 ; V. Albanesc, Arch. Ital
Biol, 50. 387, 1909 ; P. Schidrawitz and H. A. Goldsbrough, Journ. Soc Chem. Ind., 28, 3,
1909 ; H. Handovsky, Koll. ZeiL, 7. 183, 267, 1910 ; Biochem. Zeit., 25. 510, 1910 ; L. Michaelis
and B. Mostynski, ib., 25. 401, 1910 ; W. Biltz, A. von Vegcsack, H. Steiner, Zeit. phys. Chem.,
73. 500, 1910; F. Bottazzi and C. Victorou, Rend. Accad. Lincei, 19. 659, 1910; N. Sahlbom,
Koll Beihefte, 2. 79, 1910 ; W. E. Ringer, Bemmelen's Gedenkboek, 243, 1910 ; H. W. Woudstra,
Koll ZeiL, 5. 33, 1909 ; L. Loja, Koll ZeiL, 3. 247, 1908 ; H. Procte., ib., 3. 307, 1908 ; G. Gale-
otti and G. Giampalmo, ib., 3. 118, 1908; H. Freundlich and W. Neumann, ib., 3. 80, 1908;
W. Ostwald, ib., 6. 103, 1910; 7. 132, 1910; D. Molde, ib., 4. 270, 1908; E. Hatschek, ib.,
6. 254, 1910 ; 7. 1 1, 1910 ; T. B. Robertson, ih., 7. 7, 1910 ; S. U. Pickering, ib., 7. 11, 1910 ;
M. W. Beyerinch, ib., 7. 16, 1910 ; F. G. Donnan and H, E. Potts, ib., 7. 208, 1910 ; F.G. Donnan,
Zeit. phy.9. Chem., 31. 42, 1899 : V. Rothmund, ib., 63. 54, 1908 : M. Frankenheim, Journ. prakt.
Che.m., 54. 433, 1851 ; M. Rose, Phys. Zeit., 83. 47, 1907 ; ZeiL Elektrochem., 13. 499, 1907 ;
R. Schenck, Kristallinische Flussigkeiten und Flilssigekrystalle, Leipzig, 32, 1901 ; E. Eichwald,
THE KINETIC THEORY OF ATOMS AND MOLECULES 781
Neuere Untersuchungm aber die flUssigen Kristalle, Marberg, 1905; H. Freundlich and N. Ishizaka,
Trans. Faraday Soc, 9. 60, 1913 ; M. Bancelin, Koll Zeit., 9. 154, 1911 : E. Hatschek, Proc.
Phys. Soc, 28. 250, 1916; M. von Sraoluchowski, Koll Zeit., 18. 180, 1916; G. Baume and
H. Vigneron, Ann. Chim. Anal, 1. 379, 1919.
15 Lord Rayleigh, Proc. Roy. Soc, 47. 364, 1890 ; W. Ramsden, ib. ,72. 156, 1904 ; Enfjalmann'' a
Arch., 517, 1894 ; Zeit. phys. Chern., 47. 341, 1902 ; A. Pockela, Nature, 46. 418, 1892 ; Ann.
Physik, (4), 8. 854, 1902 ; G. Quincke, ib., (4), 7. 631, 1901 ; (4), 9. 969, 1902 ; (4), 10. 478, 673,
1903 ; (4), 11. 54, 1904; Wied. Ann., 35. 582, 1888 ; Ber., 38. 493, 858, 1901 ; H. Picton and
S. E. Linder, Journ. Chem. Soc, 87. 1924, 1905 ; L. Llobicky, Bull Acad. Cracovie, 488, 1906 ;
W. Frei, Zwr Theorie der Hdmolyse, Zurich, 1907 ; G. Buglia, Biochem. ZeiL, 11. 311, 1908;
P. Bottazzi and C. Victorou, Rend. Accad. Lincei, 19. 659, 1910 ; N. Sahlbom, KoU.
Beihejte, 2. 1910; H. Freundlich and W. Neumann, Koll ZeiL, 3. 80, 1908; G. N. Antonow,
Journ. Chim. Phys., 5. 372, 1907 ; W. C. McC. Lewis, Phil Mag., (6), 15. 506, 1908 ; ZeiL phys.
Chem., 74. 619, 1910 ; R. Schenck, Kristallinische Flussigheiten und Flussige Kristalle, Leipzig,
1901.
" T. Graham, Phil Trans., 140. 1, 805, 1850 ; 141. 483, 1851 ; Liebig's Ann., 77. 56, 129,
1851 ; 121. 5, 29, 1862 ; F. Voightlander, ZeiL phys. Chem., 3. 329, 1889 ; G. Hiifner, ib., 27.
227, 1898; G. Scheflfer, ib., 2. 390, 1888; J. Stefan, Sitzber. Akad. Wien, 77. 661, 1879;
R. 0. Herzog and H. Kasamowski, Koll ZeiL, 2. 1, 1907; 3. 83, 1908; Biochem. ZeiL, 11.
172, 1908 ; S. Arrhenius and T. Madsen, Immunochemie, Leipzig, 16, 1907 ; L. Vignon,
CompL Rend., 150. 690, 1910 ; T. Svedberg, ZeiL phys. Chem., 67. 107, 1909 ; S. E. Linder
and H. Picton, Journ. Chem. Soc, 61. 14, 137, 143, 1892 ; 67. 63, 1895 ; 71. 568, 1897 ;
87. 1906, 1905 ; W. Ostwald, Koll ZeiL, 1. 298, 1907 ; S. Exner, Sitzber. Akad. Wien, 56. 116,
1867 ; L. L. Oholm, ZeiL phys. Chem., 70. 378, 1910 ; E. von Regeczy, Pjlvger's Arch., 34. 431,
1884 ; W. R. Whitney and J. Blake, Journ. Amer. Chem. Soc, 26. 1339, 1904 ; W. von Wittich,
Mailer's Arch. Physiol, 286, 1856 ; M. Blom, Skand. Arch. Physiol, 20. 102, 1904 ; W. Pauli,
KoU. ZeiL, 3. 11, 1908.
17 N. Sahlbom, Koll Beihefte, 2. 79, 1910 ; H. Freundlich and W. Neumann, Koll ZeiL, 3.
80, 1908 ; R. Hober, ib., 3. 76, 1908 ; Biochem. ZeiL, 20. 80, 1909 ; L. Vignon, CompL Rend.,
150. 619, 1910 ; W. Biltz and F. Pfenning, Bemmelen's Gedenkhoek, 108, 1910 ; J. Amann, Koll
ZeiL, 7. 67, 235, 1910; F. Mylius and E. Groschuff, Ber., 39. 119, 1906; S. E. Linder and
H. Picton, Journ. Chem. Soc, 87. 240, 1905; T. Graham, Phil Trans., 140. 1, 805, 1850;
141. 483, 1851 ; Liebig's Ann., 77. 56, 129, 1851 ; 121. 5, 29, 1862 ; R. P. von Calcar, Dialyse,
Eiweisschemie, und Immunitdt, Leipzig, Leiden, 1908; G. Malfitano, CompL Rend., 139. 1221,
1904 ; R. Zsigmondy and R. Meyer, Zeit. anorg. Chem., 68. 916, 1910 ; F. Krafft and G. Preuner,
Ber., 32. 1620, 1899 ; O. Teague and B. H. Buxton, ZeiL phys. Chem., 60. 469, 1907.
18 H. Rodenwald, ZeiL phys. Chem., 24. 193, 1897.
1^ G. G. Stokes, Mathematical and Physical Papers, Cambridge, 1. 1, 1901 ; Trans. Cambridg
Phil Soc, 9. 8, 1850 ; P. Ebell, Ber., 16. 2429, 1883 ; 0. Muhlhauser, Zeit. anorg. Chem., 5. 117,
1894 ; W. D. Bancroft, Journ. Franklin InsL, 185. 29, 199, 373, 1918 ; E. Cunningham, Proc.
Roy. Soc, 83. A, 357, 1910; M. Knudsen and S. Weber, Ann. Physik, 36. 981, 1911; R. A.
Millikan, Phys. Rev., (1), 32. 349, 1911 ; (2), 2. 109, 1913 ; M. Keehan, ib., (1), 33. 1.53, 1911 ;
0. W. Silvey, ib., (2), 7. 87, 106, 1916 ; R. A. Millikan, W. H. Barber, and G. Ishida, ib., (2),
5. 334, 1915; J. Roux, CompL Rend., 152. 1168, 1911 ; 155. 1490, 1912; Ann. Chim. Phvs.,
(8), 29. 69, 1913 ; A. Schidlof and J. Morgvnowska, Archiv. Sciences Geneve, 40. 386, 486, 1915 ;
A. Schidlof and A. Karpowicz, ib., 41. 125,'l48, 1916 : CompL Rend., 158. 1882, 1914; 0. Postma,
Proc. Acad. Amsterdam, 21. 616, 1919.
20 R. Brown, Edin. New. Phil Journ., 5. 358, 1828; 8. 41, 1830; Phil Mag., (2), 4. 101,
1828; (2). 6. 161,1829.
21 M. Seddig, Ph%js. ZeiL, 9. 465, 1908 ; V. Henri, CompL Rend..U7. 62, 1908 ; H. Siedentopf,
ZeiL wiss. Mikrosk., 26. 407, 1909 ; T. Svedberg, Koll ZeiL, 7. 1, 1910.
22 R. Zsigmondy, Zur Erkenntnis der Kolloide, Jena, 1905 ; New York, 1909 ; M. Faraday,
Phil Mag., (4), 14. 401, 512, 1857.
23 J. Perrin, Les atomes, Paris, 1913 ; London, 1916 ; T. Svedberg, Die Existenz der Mole-
kule, Leipzig, 1912 ; J. Perrin, Ann. Chim. Phys., (8), 18. 5, 1909, Brownian Movement and Mole-
cular Reality, London, 1910; Koll Beihefte, 1. 221, 1910; E. F. Burton, The Physical Pro-
perties of Colloidal Sohitions, London, 1916 ; E. E. F. d'Albe, Contemporary Chemistry, London,
1911 ; J. Becquerel, Scient. Amer. Snppl., 88 260, 1919.
24 C. Wiener, Pogg. Ann., 118. 79, 1863 ; C. Fuchs, ReperL Physik, 25. 735, 1889 ; P. M.
Exner, Ann. Physik, (4), 2. 843, 1900 ; S. Exner, Sitzber. Akad. Wien, 56. 116, 1867 ; A. Gouy,
Journ. Phys., (2), 7. 561, 1888 ; CompL Rend., 109. 102, 1889 ; J. Delsaulx, Journ. Roy. Microscop.
Soc, 18. 17, 1877 ; Rev. Questions scienL, 1. 319, 1877 ; J. Thirion, ib., 4. 53, 1880 ; J. G. G.
Carbonelle, U aveuglement scientifque, 378, 1877-80 ; R. Zsigraondv, Zur Erkenntnis der Kolloide,
Jena, 117, 1910 ; T. Svedberg, ZeiL Elektrochem., 12. 853, 909, 1906 ; W. Ramsav, Phil Mag.,
(5), 1. 328, 1876 ; H. Siedentopf, ZeiL wiss. Mikrosk., 26. 407, J909 : A. Einstein, 'yl?m. Physik,
(4), 17. 549, 1905 ; (4), 19. 371, 1906 ; M. von Smoluchowsky, ib., {4),2i. 756", 1906 ; Bull Acad.
Cracow, 577, 1906 ; J. Perrin, Les preures de In realite moleculaire, Paris, 1909 ; London, 1910;
Les atomes, Paris, 1913 ; London, 1916 ; V. Henri, Koll ZeiL, 12. 24(5, 1913.
25 J. Perrin, Com,pL Rend., 146. 967, 1908 ; 147. 530, 1908 ; J. ThoVert, ib., 133. 1197, 1901 ;
134. 507, 1902; 135. 579, 1902; R. S. Lillie, Amer. Journ. Physiol, 20. 127, 1907 ; B. Moore
782 INORGANIC AND THEORETICAL CHEMISTRY
and H. E. Roaf, Biochem. Journ., 2. 34, 1906 ; H. Rodewald, Zeit. phys. Chem., 13. 633, 1900 ;
R. O. Herzog, Zeit. Elektrochem., 13. 533, 1907.
*• H. Zangger, Koll. Zeit., 9- 216, 1911 ; A. Einstein and M. von Smoluchowsky, vide supra;
E. E. F. d'Albe, Contemporary Chemistry, London, 1911; B. Iljin, Zeit. phys. Chem.y 83. 692,
1913; M. L. Brillouin, Perrin's Atoms, London, 131, 1916.
§ 8. The Kinetic Theory of Atoms
There are countless worlds in countless heavens each revolving about its sun.—
G. Bruno.
The curve described by a single atom is as fixed as the path of a planet, and between the
two cases no other difference exists save that resulting from our ignorance.' — L. Meyer.
Many philosophers— E. Kant, G. W. E. Hegel, T. S. Hunt, etc. — have laboured
in vain to demonstrate by abstract reasoning that chemical combination is an
interpenet ration of masses or a juxtaposition of molecules. According to G. W. E.
Hegel, the chemical process is either an identification of the different, or a differenti-
ation of the identical. The characteristic mark of a chemical species or individual
is homogeneity . Is this homogeneity merely relative ? Can it be truly said : Tola
in minimis existit natura ?
Do the atoms of the molecules of a compound retain their individuality ? —
It may be quite true that the properties of a compound are mainly determined by the
character of the constituent elements, yet, it is not to be supposed that there is
necessarily any resemblance between the properties of the elements and of their
compounds. It is not yet possible for the chemist to infer a priori, nor explain
a posteriori the properties of a compound from the properties of the constituent
elements. For instance, no one would have suspected the peculiar properties of cyano-
gen from the qualities of its constituent elements carbon and nitrogen, or of sodium
chloride from the constituent elements sodium and chlorine. Chemists generally
consider that the atoms preserve a kind of individuality throughout their existence,
and when associated with other atoms, change their habit but not their nature. The
change in the habits of an atom depends entirely on its associates. For instance, the
properties of a molecule of water are very different from the properties of either of
the constituents hydrogen or oxygen ; the atoms of oxygen are magnetic when
associated together in pairs or triplets, but they are non-magnetic when compounded
with many other elements ; iron too is intensely magnetic, some if its compounds
are also magnetic, yet there are others which are non-magnetic, and still others which
are diamagnetic — e.g. the iron carbonyls. Although the atoms of a compound
molecule do not enjoy a separate external existence, yet, within the molecule, the
atoms are probably distinct individuals, self-contained and self-existent. As
Lucretius would have expressed it, they are solida pollentia simplicitate, or strong
in their solid singleness ; but the individual properties of the atoms are not always
unrecognizable in the properties of the molecules of their compounds. J. Larmor
(1908) 1 has well said :
It becomes increasingly difficult to resist the simple view that chemical combination
involves an arrangement of the atoms alongside each other under steady cohesive affinity,
the properties of each atom being somewhat modified, though not essentially, by the attachment
of the others: and that the space formulae of chemistry have more than an analogical signifi-
cance. The many instances in which the physical properties of the compoimd molecule
can be calculated additively with tolerable approximation from those of the constituent
atoms, are difficult to explain otherwise.
Those qualities which depend upon the nature of the atom in the molecule are called
additive properties when each atom exerts its own specific influence whatever its
state of combination. The following is selected from the evidence which might
be cited to show that the atom retains its individuality in all its migrations no matter
THE KINETIC THEORY OF ATOMS AND MOLECULES 78.3
how many of its properties might be disguised by association with other atoms :
(i) The weight of an atom remains intact whatever be its associated partners,
(ii) The atom emits a peculiar type of Rontgen ray when stimulated in a suitable
manner, and this property can neither be changed nor disguised by associa-
tion with other atoms, (iii) The absorption of Rontgen and cathode rays is
an atomic property, for each atom has its own specific absorptive power which
is independent of the nature of the partners, with which it may be associated.
There are numerous other examples — specific heat, crystalline form, etc. — not
quite so decisive.
Are the atoms of a molecule at sensible distances apart ? — It is sometimes
asserted that the atoms are at insensible distances apart and that the atoms of a
molecule are accordingly very close together. These statements have given rise to
a misconception, for if the size of the atom be taken as a standard of reference
it is probable that in the molecule the distances of the atoms from one another are
comparatively great. Nothing is great or small unless it be considered in relation
to other things regarded as standards of comparison.
Are the atoms of a molecule at rest or in motion ? — The molecules are in motion,
and therefore the atoms which make up the molecules must also be in motion just as
a train waiting at the station is said to be at rest, although it is moving with the earth
about the sun, at a great velocity. Accordingly, the term rest here refers to the
position of the atom with respect to the molecule regarded as a standard of compari-
son. J. B. A. Dumas (1837),2 S. Brown (1843), and D. I. MendeleefE (1868), like
many previous philosophers — ^notably C. L. BerthoUet (1803) — picture a complex
molecule to be analogous with a kind of miniature solar system with the atoms
whirling rhythmically about one another at great speeds. Like the planets and their
satellites, the atoms are supposed to be " endowed with an everlasting motion.'*
The atoms are further supposed to be held in position, and to move in definite orbits
owing to their attraction for one another, just as the planets and satellites move in
definite orbits owing to the action of gravitational forces. D. I. Mendeleeff, like
S. Brown, was an enthusiast ; he said :
Chemically, the atoms may be likened to the heavenly bodies, the stars, sun, planets,
satellites, etc. The building up of the molecules from atoms, and of substances from mole-
cules is then conceived to resemble the building up of systems, such as the solar system,
or that of twin stars, or constellations from individual bodies. This is not a simple play of
words in modern chemistry, nor a mere analogy, but a reality which directs the course of all
chemical research, analysis, and synthesis.
Molecular models. — M. Berthelot (1875) ^ said that a complete representation of
chemical compounds must involve the notion of rotatory and vibratory movements
by which each particular atom, and each group of atoms in the molecule are animated.
Although many chemists have similarly expressed their belief in a kinetic theory of
atoms based upon a supposed analogy between atomic and planetary systems, they
would yet recoil from any attempt to represent the idea pictorially or by mechanical
models ; but what Lord Kelvin said of himself applies to most : "I never satisfy
myself until I can make a mechanical model of a thing. If I can make a mechanical
model, I can understand it." So long as we are not seduced by a prepossessing
analogy, there is no harm in constructing a model or diagrammatic picture because
the strength and weakness of the analogy may be then better apprehended. Follow-
ing up the analogy between planetary systems and the constitution of molecules,
a two-atom molecule of hydrogen, H2 ; iodine, I2 ; or oxygen, O2, can be regarded
as a binary star — that is, as a pair of stars — in which each atom in the molecule
rapidly revolves about the other in a regular orbit. A molecule of water, H2O,
would be represented by three atoms revolving in a similar manner ; ammonia,
NH3, said D. I. MendeleefE (1889), may be represented in the simplest manner by
supposing the sun, nitrogen, to be surrounded by three planets, hydrogen atoms ;
and a molecule of sulphuric acid, H2SO4, might be depicted as a complex system
784 INORGANIC AND THEORETICAL CHEMISTRY
with a central revolving sulphur atom around which the other atoms whirl in definite
orbits :
Each orb, the smallest in its motion sings.
First would come one sulphur or two oxygen atoms representing the nucleus, SO2,
outside these would encircle two oxygen atoms each with a revolving hydrogen atom
as satellite. The imaginary picture so obtained would be a kinetic model of the
formula (HO)2=S02. E. Frankland and F. R. Japp ^ explained the constitution of
acetic acid, CH3.CO.OH, in a similar manner in 1884. The chemist determines the
constitution of these tiny systems by a process which G. Martin has compared with
the plucking of, say, the earth and moon from the solar system, or by replacing one
planet by another and observing the disturbing effects of the transposition on the
whole system ; for, said C. Daubeny (1850), it is probable that any of the planets in
the solar system could be replaced by a ball of matter with totally different properties,
provided its gravitational mass were the same, without disturbing in the least the
conditions of mechanical equilibrium. A kind of orrery would therefore give a better
idea of the structure of a molecule than the crude plane formulae usually employed.
By this analogy, the planets Mercury and Venus represent single atoms, the
Earth, Jupiter, and Saturn with their moons represent radicles — each composed
of several distinct atoms so as to form a small sub-system complete in itself.
All these individuals and sub-systems are linked to one another so as to form a
balanced or stable molecular system, in some respects analogous with the solar
system.
Valency. — Supposing that the above 'speculations were to be established by
unassailable evidence, that would not alter the value of graphic or constitutional
formulae. E. Molinari (1893) considered that the constitution of compounds is
rather dependent upon the intramolecular movements of the atoms in relation to
each other, than on the relative positions of the atoms in space ; and that the so-
called valency bonds denote the nature of the motion or energy of the atoms with
regard to each other. Hence, so far as graphic formulae are concerned, it really
makes little difference whether the atoms are actually attached to one another, or
whether they are held in position by their mutual attractions while they are revolving
about a centre of stability. Indeed, some assume that the conditions of temperature,
light, or electricity necessary for the formation of a stable system determine whether
a given atom can form a stable system with 1, 2, 3 . . . other atoms ; otherwise
stated, the valency of an element is determined by the necessity for harmonizing
the peculiar motions of the combining atoms to form stable molecular systems.
When J. Dalton was asked why an atom of carbon would take up one or two atoms
of oxygen, but not three or four, he replied :
The reason I would assign is that in the state of COj there are two atoms of oxygen com-
bined with one of carbon, and a third or fourth oxygen atom, however it may be attracted
by the carbon, cannot join it without repelling one or more of the atoms of oxygen already
combined. The attraction of carbon is able to restrain the mutual repulsion of two atoms
of oxygen but not that of three or four.
S. Brown expressed the same idea in 1843 : " The conception can perhaps be made
still more lucid by the counter statement in astronomy that a sun cannot be over-
loaded with planets." S. Brown's view of valency shows that it is not necessary to
postulate a distinct force emanating from the atoms in order to explain how, say,
HCl forms a stable system, while HCI2 and H2CI do not form stable molecular
systems. If such systems were momentarily formed, the supernumerary atoms
would be immediately flung off. After trying motions and unions of every kind,
the atoms no doubt fall into those favourable arrangements which can persist as
stable configurations. There may, of course, be a number of different stable systems
corresponding with the different stable molecules of, sav, iron and chlorine, FeCU
and FeClg. L. Meyer (1884),5 E. Molinari (1893), F. P. Venable (1899), and others
THE KINETIC THEORY OF ATOMS AND MOLECULES 785
have advocated similar views. The plausibility of this hypothesis, of course, is not
a proof that it is true.
The energy of atoms. — Each elementary atom presumably has its own definite
charge of energy. The energy possibly exists in the form of rhythmical atomic
motions, so that when one atom unites with another atom, each atom possibly gives
up a part of its energy or absorbs energy from some external source, so that the
motions of the one atom may be co-mingled with the motions of the other atoms to
form a stable molecular system. The hypothesis thus suggests a plausible explana-
tion of selective affinity.
Selective chemical affinity. — ^Lucretius frequently affirmed that it was abso-
lutely decreed from the beginning what each thing can and cannot do ; and to-day
it is assumed that the molecules of matter are endowed with certain peculiar qualities,
for the physicist explains gravitational phenomena by investing all the molecules
of matter with a common property which he calls gravity, even though he may say
with Isaac Newton (1717) : " Gravity is not to be taken as an essential property of
bodies." The chemist too explains chemical action by endowing the atoms with a
selective power which he calls affinity. Gravitation is purely a physical relation
common to all molecules of all known kinds of matter, while ajB&nity is a very special-
ized chemical relation characteristic of specific types of matter. The kinetic
theory of selective affinity assumes with W. M. Wundt (1897) that all the qualitative
properties of matter are derived from the different modes of motion assumed by the
atoms ; the atoms themselves are completely devoid of quality. It is supposed
that when two molecules meet, they can react chemically only when the motions
of the atoms of the one molecule can be co-mingled with the motions of the atoms of
the other molecules, so that instead of *' shattering, confounding, and dispersing "
one another's motions, they move in cadence and form harmoniously working systems
called molecules. Two atoms moving in unison support and sustain one another's
attractions ; two atoms moving in discordant periods, despite their mutual influence,
cannot form a stable combination because they offer a certain resistance to conjuga-
tion. As it has been otherwise expressed, " Every atom according to its nature is
always striving to get into harmonious relations wdth other atoms." The idea
recalls Democritus' view that the atoms are attracted to one another on account
of their whirling motions ; paraphrasing Lucretius, " the atoms unite in all manner
of ways, and thoroughly test motions and combinations of every possible kind ; "
consequently, it is not at all strange that the atoms have at last formed arrangements
which can be maintained more or less permanently. These speculations may give
the impression that chemical phenomena will be ultimately referred to fundamental
mechanical laws ; but we have passed in imagination beyond the region of demon-
strated fact, and are dimly conscious of an illimitable expanse where hypothesis and
conjecture can but wander aimlessly and blindly. Here Newton paused : " The
whole frame of nature may have been wrought into various forms, at first by the
immediate hand of the Creator, and ever after by the power of nature."
What makes the atoms and molecules move ? — We do not know ! How can
matter of itself initiate motion, ^ and particularly motion in a harmoniously working
system ? Ignoraynus ! In the words of C. Kingsley, " Everywhere skin-deep
below our boasted science we are brought up short by mystery impalpable, and by
the adamantine gates of transcendental forces and incomprehensible laws." We are
profoundly ignorant of the cause of the specific activities of atoms, molecules, and
planets. Immortal Newton could get no further than this : " The motions which
the planets now have could not spring from any natural cause." It seems as if
Full many a secret in her sacred veil
Hath Nature folded. She vouchsafes to knowledge
Not every mystery, reserving much
For human veneration, not research.— Anon. (1851).
Consequently, the kinetic theories of the Brownian movements, of atoms, of mole-
cules, of the planetary systems, and indeed of the solar system itself, are all compelled
VOL. I. 3 E
786 INORGANIC AND THEORETICAL CHEMISTRY
to prescribe or postulate an initial state of motion which is self-sustained and self-
regulated. Guesses at the birth-history of these motions has been whispered only
by the poets. Deus mundum cedijwavit, said Cicero ; and, in the oft-quoted lines
of Virgil :
Know first, the heaven, the earth, the main,
The moon's pale orb, the starry train.
Are nourished by a soul,
A bright intelligence, whose flame
Glows in each member of the frame.
And stirs the mighty whole.
References.
1 J. Larraor, Mem. Manchester Lit. Phil. Soc, 52. ii, 1, 1908.
^ J. B. A. Dumas, Lecons sur la philosophie chimique, Paris, 232, 1837 ; S. Brown, Lectures
on the Atomic Theory^ Edinburgh, 1858; A. Wurtz, The Atomic Theory, London, 313, 1880;
D. J. Mendeleeff, The Principles of Chemistry, London, 1891 ; C. L. Berthollet, Essai de statique
chimique, Paris, 1803.
3 M. Berthelot, Bull. Soc. Chim., (2), 23. 338, 1875.
* E. Frankland and F. R. Japp, Inorganic Chemistry, London, 1884 ; C. Daubcny, An Lntro-
duction to the Atomic Theory, Oxford, 1850 ; G. Martin, Triumphs and Wonders of Modern Chem-
istry, London, 1911.
^ E. Molinari, Journ. prakt. Chem., (2), 48. 113, 1893 ; L. Meyer, Die modernen Theorien der
Ghemie und ihre Bedeutung fiir die chemische Mechanik, Brcslau, 1884 ; London, 1888 ; F. P.
Venable, Journ. Amer. Chem. Soc, 21. 192, 220, 1899.
« J. Croll, Phil. Mag., (4), 44. 1, 1872.
§ 9. The Two Specific Heats of Gases
The kinetic theory of molecules assumes that the temperature of a gas is propor-
tional to the average speed of translation of the moving molecules — an increase of
the speed is accompanied by a rise of temperature, and conversely.
It will be remembered that specific heat is a term employed to represent the amount of
heat required to raise the temperature of one gram of a substance 1°. A gas can be heated
by simple compression, its specific heat then appears to be zero ; but, in reality, a certain
amount of energy, equivalent to the specific heat, is needed for the work of compression.
Again, a gas, if it be expanded, is cooled ; if the cooling effect of expansion just coimter-
balances the heat added to the gas, the temperature remains constant ; and the specific
heat appears to be indefinitely large. Here work, equivalent to the heat supplied, is per-
formed by the expanding gas. These facts show that the condition of the gas must be stated
before it is possible to define its specific heat. It is conventionally agreed that if the gas be
allowed to expand during a change of temperature so that its pressure remains constant,
the amount of heat required to raise the temperature of one gram-molecule of the gas 1°
shall be called the specific heat under constant pressure, and symbolized by Cp. If the
pressure be increased so that the volume remains constant when the gas is heated, the
amount of heat required to raise the temperature 1° of a gram-molecule of the gas is likewise
called the specific heat under constant volume, and symbolized Cv.
The heat imparted to a gas is not spent merely in raising the temperature of the
gas ; that is, in speeding up of the motions of the molecules. Energy is spent in — (1)
Augmenting the speed of the moving molecules. The heat required actually to increase
the kinetic energy of the moving molecules so as to produce a rise of temperature
is the same for all gases. Let K denote this quantity. (2) Performing external
work. Heat energy is needed to overcome the pressure of the atmosphere when the
gas is allowed to expand. Call this quantity W. Since the coefficient of thermal
expansion of all gases is the same, this quantity is practically constant for
equal volumes or equimolecular weights. (3) Performing internal work. Heat
energy is required to produce changes within the molecule which may alter the
motions or orientation of the constituent atoms of the molecule, or raise the kinetic
energy of the atoms moving with the molecule. Let e denote the energy spent
THE KINETIC THEORY OF ATOMS AND MOLECULES 787
within the molecule per degree rise of temperature. A certain amount of energy
must also be spent in overcoming the effects of intermolecular attractions. This
can be neglected for the time being. Consequently, the ratio of the two specific
heats may now be written, after J. J. Waterston (1845) : i
The specific heat of a gas at constant volume.— We have seen, (1) that pv
=iMV^, where ilf denotes the mass, and V the average velocity of the molecules.
But the kinetic energy of a body of mass M moving with a velocity V is JMF^ ;
hence pv=^xiMV^ ; or the kinetic energy of the molecular motions is §pu,
since pv—RT, the kinetic energy of molecular motion is '^RT. If one gram-molecule
of gas be heated 1°, the kinetic energy becomes |-K(T+1). Hence if the gas be
heated 1° at constant volume, the thermal value of the increased kinetic energy is
"^RiT -^1) ~^RT ='^R cals. This result represents the specific heat of the gas at
constant volume per gram-molecule ; or, C^^'^R.
The specific heat at constant pressure. — Again, if a gram-molecule of gas
expands against atmospheric pressure when its temperature is raised 1°, the gas,
in consequence, does work by pressing back the atmosphere, so to speak. The
equivalent of this work must be supplied in the form of heat. This work is equivalent
to the product of the pressure against the change in volume. Let x denote the change
in volume when the gas is heated 1°, under a constant pressure ; then, p{v-\-x)
~R{T-{-l), and pv=RT. By subtraction px=R. This means that when a
gram of gas is heated 1°, the resulting expansion against atmospheric pressure does
work equivalent to R cals. ; or the gas constant R, is numerically equal to the ivork done
by a gas expanding against a constant atmospheric pressure, when the temperature is
raised 1°. Hence, R cals. must be added to the previous result to obtain the thermal
equivalent of the energy supplied to one gram of gas in the form of heat when its
temperature is raised 1°. Otherwise expressed, if one gram-molecule of gas be
heated 1°, at constant pressure, an amount of heat equivalent to ^R-\-R=^R
is required. This result represents the specific heat of the gas at constant pressure
per gram-molecule or Cp^R. In passing, it is interesting to note that the differ-
ence between the tivo specific heats of a gas is numerically equal to the ivork done by the
gas expanding against a constant atmospheric pressure when the temperature is raised
1°, or in symbols,
Cp — Cv=R
a relation sometimes called Mayer's equation,^ because it was used by him in
1842 to calculate the mechanical equivalent of heat, for the difference in the two
specific heats represents the external work done during the expansion of 1 c.c. of
air {i.e. 0-001293 grm.) against atmospheric pressure {i.e. 1,031,000 dynes per sq.
cm.) when heated 1°. In Mayer's equation, when the work R is measured in calorics,
R is approximately 2 cals. The same value of R can be obtained another way — one
gram of oxygen occupies 699*8 c.c. at standard pressure 1,013,200 dynes per sq. cm.
and at 273° K. Hence, for one gram of oxygen, R'^^pvjT =2' 6x10^ ergs;
and for 32 grams, or one gram-molecule, 22=2-6x106x32 =8-3x107 ergs=8-3
joules=2 cals. (nearly). Instead of representing gram-molecules of a gas, the
specific heats Cp and C^ can be referred to one gram. In that case, R. Mayer's
equation becomes Cp--Cv=RIM, where M denotes the molecular weight of the
gas, and, for oxygen, i2/M==2-f32 =0-063 cal.
Examples.— (1) The specific heats of oxygen at constant pressure and constant vohime
are respectively 0*217 and 0"155. The difference in the molecular specific heats is therefore
32(0-217-0-155)-=2 cals. nearly.
(2) Compute the mechanical equivalent of heat when for air Ct)=0'1685 and Cp
— 0'2375. The mechanical equivalent of heat J is equal to the ratio W/Q, where Q denotes
the amount of heat in calories required for performing W ergs of work. The work of
788
INORGANIC AND THEORETICAL CHEMISTRY
expansion when 1 c.c. of gas expands ^^g c.c. is ^ij-g x 1,013,000 — ly, and the heat Q is
equivalent to 0'001293(0-2375— 0'1685) cal. Hence J=41-(i x 10« ergs per calorie.
The ratio of the two specific heats of a gas.-
two specific heats, which is usually symbolized y,
W=R, or,
^^=0
-Returning to the ratio of the
we can now write K=^^R, and
(2
The magnitude of € will vary with different gases, for it will naturally be related
somehow with the complexity of the molecule. The greater the value of € the less
the value of the ratio of the two specific heats. For a monatomic gas, e probably
approaches zero, and the numerical value of the ratio becomes y=§, or 1'67. The
greater the complexity of the molecule, the greater the value of €, and the smaller
the value of the ratio of the two specific heats. This is illustrated by Table X.
Table X.— Ratio of the Two Specific Heats of Gases.
Atoms
Atoms
Gas.
Mole-
cule.
per
mole-
cule.
7
1-67
Gas.
Mole-
cule.
per
mole-
cule.
7
Mercury
Hg
1
Carbon dioxide .
CO2
3
1-31
Argon .
A
1
1-65
Nitrous oxide
N2O
3
1-31
Hydrogen
Ha
2
1-41
Hydrogen sulphide
H,S
3
1-31
Nitrogen
N2
2
1-41
Ammonia .
NH3
4
1-30
Oxygen .
O2
2
1-40
Methane
CH4
5
1-27
Carbon monoxide .
CO
2
1-40
Ethylene .
C,H,
6
1-24
Hydrogen chloride .
HCl
2
1-39
Ethane
C^He
8
1-18
Chlorine
Ch
2
1-32
Alcohol
C^HgOH
9
1-13
Bromine
Br,
2
1-29
Benzene
CeHe
12
1-09
Iodine .
I2
2
1-29
Ether
C4H,oO
15
1-06
Iodine chloride
ICl
2
1-31
Turpentine.
C10H16
26
1-03
The effect of variations of temperature and pressure on y. — ^The ratio of the
two specific heats of gases decreases as the temperature rises. Thus, A. Wiillner ^
found the ratio of the specific heats of the following gases to fall as the tempera-
ture rose from 0° to 100 :
7 for
Air
CO
CO 2
N.,0
NH3
C^H,
0°
. 1-4052
1-4032
1-3113
1-3106
1-3172 .
1-2455
100°
. 1-4051
1-3970
1-2843
1-2745
1-2791
1-1889
and E. H. Stevens found the value for air to fall from 1-4006 at 0°, to 1'3993 at 100°,
to 1-3400 at 950° ; and M. Trautz obtained for water vapour saturated at 100°,
y=l-3290, and saturated at 110°, 13301 ; at 120°, 13129 ; and at 130°, 1-3119.
Conversely, the ratio of the two specific heats of gases increases as the temperature
falls from about 20° to —180°. For instance, M. Trautz (1913) has shown :
Nitrogen,
Carbon mon-
Oxygen,
Hydrogen,
Helium
N2
oxide, CO
O2
H2
He
1-400
1-398
1-399
1-407
1-660
1-468
1-472
1-447
1-597
1-673
Y from 18° to 20° .
y from -180° to -181°
A. Witkowsky obtained analogous results with air, and he also found that the ratio
of the two specific heats increased when the pressure is augmented from 10 to
100 atm. both at 0° and at — 78'5°, and attains a maximum value at —120°, as
indicated in Table XI.
The ratio of the two specific heats and molecular weights. — The numbers in
Table XI mean that if the ratio of the two specific heats of a gas be about 1*6, the
THE KINETIC THEORY OF ATOMS AND MOLECULES
789
gas will usually have one atom per molecule ; with a ratio about 1*4, two atoms per
molecule ; and with a ratio of about 13, three atoms per molecule. The kinetic
theory would have no explanation to offer if the value of y were greater than If ;
but no cases are known. There are a number of discrepancies. This must be
expected owing to differences in molecular attraction, tendencies to polymerization,
Table XI. — The Effects of Temperature and Pressure on the Ratio of the
Two Specific Heats of Air.
Temperature.
10
30
60
100
- 0°
1-43
1-44
1'63
1-60
- 60°
1-42
1-49
1-58
1-72
-100°
1-44
1-53
1-71
210
-120"
1-45
1-56
1-79
-UO"*
1-38
1-46
1-54
1-80
dissociation, etc., which affect the value of c. The coloured gases — chlorine,
bromine, iodine, and iodine chloride, with two atoms per molecule — give lower
values than is usually obtained with the colourless diatomic molecules ; and gases
which are readily condensed to liquids give rather lower values than those less readily
liquefied. Hence, if the ratio of the two specific heats of a gas falls into one of
these groups — 1*6, 1*4, 1*3 — this fact may be taken as circumstantial evidence,
but not conclusive proof, that the molecule has a corresponding number of
atoms per molecule. There is, however, no unimpeachable relation connecting
the specific heat of a complex molecule with the nuivbtr of the constituent atoms
which is independent of their nature. The ratio of the two specific heats of argon
and the inert gases appears to be about 1'6, and hence it is supposed that the mole-
cules of these gases are monatomic, like mercury. This means that the density
(^=2), the molecular weight, and the atomic weight will probably have the same
numerical value. Hence, the determination of the ratio of the two specific heats
provides an independent method of ascertaining the number of atoms in the molecules
of a gas without reference to the compounds of the element. In the case of mercur)^
the monatomicity of the gas has been established altogether apart from this
reasoning.
This subject cannot be passed by without bringing a weak step in the above
reasoning into prominence. The low molecular heats of the inert gases are assumed
to prove that these gases have one-atom molecules. But it is easy to see that if Uttle
or no heat is expended in doing internal work when the temperature of a gas is
raised, a gas with polyatomic molecules might be reported to have monatomic
molecules. Unlike mercury, the inert gases do not form chemical compounds, and
hence the number of atoms in the molecule cannot be determined by the usual
methods based upon Avogadro's hypothesis. The inference that the molecules of
the inert gases are monatomic involves an assumption which is less readily granted
than is the case with mercury, cadmium, etc., because these elements form volatile
compounds which enable their atomic weights to be evaluated. Hydrogen at very
low temperatures behaves in this respect like a monatomic gas.
References.
i J. J. Waterston, Phil. Tram., 183. A, 1, 1892.
2 J. R. Mayer, Liebig's Ann., 42. 1, 1842 ; Phil. Mag., (4), 24. 371, 1803.
3 A. Wullner, Wied. Ann., 4. 321, 1878; M. Trautz, Phys. ZeiL, 14. 1170, 1013; Ber. dent,
phys. Ges., 15. 9(39, 1913 ; E. H. Stevens, Ann. Physik, (4), 7. 28i5, 1902 ; A. Witkowsky, Bidi.
Inter nat. Cracow, 138, 1899 ; Phil. Mag., (5), 42. 1, 189G.
790 INORGANIC AND THEORETICAL CHEMISTRY
§ 10. The Relation between the Two Specific Heats oi a Gas and the
Degree of Freedom of its Molecules
Invisible movements, invisible particles- — these and kindred assumptions have supplied
the window through which the human mind has sought to spy into the inner machinery
of phenomena. — T. Gomperz (1912).
R. Clausius i has shown that a vahie for the total heat energy contained in a gas
can be obtained by assuming that the gas has been brought into its present condition
by being warmed at constant volume from absolute zero to the temperature T.
The density D of the gas represents the mass of the gas, and, if Cv be the specific
heat at constant volume assumed for convenience not to vary with temperature, the
total heat energy of the gas will be CJ)T. From (5), § 2, the kinetic energy of the
branslatory motions of the molecules of a gas is |^, a magnitude which also increases
proportionally with the absolute temperature. The ratio of the total heat energy to
the total kinetic energy is therefore a constant independent of temperature — provided
Ct, is independent of the temperature. Both magnitudes are proportional to the
density D of the gas, and accordingly, the kinetic energy of the molecules of a perfect
gas stands in a constant ratio to the total energy of the gas. From J. R. Mayer's
equation, the kinetic energy of translatory motion is ^f=^^(Cp—C^)DTi and
accordingly, the ratio of the two forms of energy is :
Energy of translatory motion 3(y — 1) /ov
Total energy ~ "" 2 ' * • w
which shows that the ratio of these two forms of energy is determined by the two
specific heats.
The kinetic energy of the molecules of a gas is supposed to be divided between
the kinetic energy of the translatory motions, and of the various rotatory motions
of the molecules. The former is sometimes called the external, and the latter
the internal energy of the gas. During a collision there is probably a rapid exchange
of kinetic energy between the external and internal motions. There are probably
also elastic vibratory internal motions which are scarcely affected by molecular
collisions, although a gain or loss of this form of internal energy may become appreci-
able after countless collisions. The vibratory energy may be dissipated as radiant
heat, etc., and if the gas were not exposed to an external source of energy, it would be
cooled by the loss of radiant vibratory energy, since the kinetic energy of the trans-
latory and rotational motions would be gradually transformed into vibrational
energy. Conversely, if a gas be exposed to a source of radiant energy, the vibrating
motions are accelerated owing to the absorption of vibrational energy, and the
energy so gained is but slowly converted into kinetic energy of translatory motion
whereby the gas is warmed. The difficulty of heating gases by simple radiation
shows that radiant energy is not rapidly absorbed by the molecules of a gas.
From equation (3) it follows that if Q be the total energy of a gas, and K the
kinetic energy of translatory motion, Q—K represents the internal energy, or the
Internal energy __ 3/5— SyN _ 5— 3y - . .
Total energy 2\ 3 / 2 * * ' \ )
The energy of the translatory motions of the molecules has been shown to be equal
to f 12. Consequently, the ratio
Internal energy of the molecules 5 — 3y /_.
Energy of translatory motion 3(y — 1)
The constancy of the ratio y for certain gases thus shows that the ratio of the
internal or vibratory energy to the energy of translatory motion is likewise constant
for these gases at ordinary temperatures ; or the internal energy of a gas is propor-
tional to the kinetic energy of the gas. 2 The ratio of the vibratory to the translatory
1
THE KINETIC THEORY OF ATOMS AND MOLECULES
791
energy increases with the number of atoms in the molecule. With monatomic gases
this ratio (5) is zero, and Table XII shows the value of the ratios (3) and (5)
for a number of other gases as well as for ratio (5) divided by the number of atoms
in the molecule.
Table XII. — Moleculir and Atomic Energies.
Oxygen .
Nitrogen .
Hydrogen .
Carbon monoxide, 0
Carbon monoxide, 100"
Nitric oxide
Hydrogen chloride
Chlorine iodide, ClI
Chlorine .
Iodine
Carbon dioxide .
Nitrous oxide .
Water vapour, 103°
Ammonia
IMethane .
Ethylene .
Ethylchloride .
Ethylether
1-403
1-405
1-394
1-403
-397
-394
•392
-317
-323
•294
-300
-270
1-277
1-262
1-316
1-243
1-126
1^029
Kinetic energy | Internal energy Internal energy ,
Total energy I Kinetic enerev Kinetic eneray *^
0-604
0-607
0-591
0-605
0-595
0-591
0-586
0-475
0-485
0-441
0-449
0-405
0-415
0-393
0-474
0-364
0-189
0043
0-656
0-646
0-692
0-653
0-679
0-692
0-706
1-103
1-064
1-268
1-226
1-469
1-407
1-543
1-110
1-740
4-300
22-200
Kinetic energy
0-328
0-323
0-346
0-327
0-340
0-346
0-353
0-551
0-532
0-634
0-409
0-490
0-469
0-386
0-222
0-291
0-537
1-480
For moderate temperatures, the ratio of the molecular and atomic energies is
nearly constant for diatomic molecules, but varies with temperature and other
circumstances for more complex molecules. While the ratio of the two specific heats
usually decreases a little with a rise of temperature, the ratio between the internal
and kinetic energies, per atom, increases with a rise of temperature, and this increase
appears to be the greater, the larger the number of atoms per molecule. This has
been explained by assuming that some of the energy is consumed in work against the
chemical affinity which hold the atoms of the molecule together.
In his Ueher Molekularphysik (Konigsberg, 1888), F. Lindemann 3 assumes that
the development of heat during chemical action consists merely in the transfer
of the internal vibratory energy into translatory energy so that the products of the
action possess less internal energy than before. If this be true, the internal vibratory
energy of the initial products of the reaction between, say, hydrogen and chlorine
must be enormously greater than the translatory energy, because of the great develop-
ment of heat which occurs, and this is quite incompatible with the observed ratios
of the two specific heats of the gases. Attempts ^ to correlate the specific heat ratio
with the numbers of atoms in the molecule have not been very successful. J. C. Max-
well 5 sought a relation between the ratio of the two specific heats and the number
of ways a system of particles is movable — Beiveglichkeitsarten — or the so-called
number of degrees of freedom of the molecules.
Degrees of freedom. — At any instant, the position of a particle compelled to oscillate to
and fro on a given straight line is completely described by its distance from a fixed point in
that line ; if the particle moves in a plane, its position will be described by its distance from
two intersecting straight lines in that plane ; while if the particle moves in space, its position
can be described by its distance from three fixed intersecting planes. In the first case, the
particle is said to have one degree of freedom ; in the second, it is said to have two degrees of
freedom, because two relations are needed to define its position, and each of the two relations
can change independently of the other ; in the third case, the particle is said to have three
degrees of freedom, because three independent relations are needed to define its position.
The degree of freedom of an object is the number of facts which must be specified in order
to define completely its position or state; more precisely, a degree of freedom is an
793 INORGANIC AND THEORETICAL CHEMISTRY
independent mode in which the condition or state of a body can be altered. Consequently,
the number of degrees of freedom of a particle in sjoace cannot be less than three. If a
short straight indefinitely thin rod bo substituted for the particle, then, the position of the
rod in space can be defined by indicating the distances of the two ends of the rod from the
three intersecting planes. Hence, six relations are then used in defining the position of the
rod. The rod, it is to be remembered, has a fixed definite length, and one of the six relations
can be eliminated because it is related with the other five ; as a result, the position of the
rod can be completely defined by five independent relations. The rod is therefore said to
have five degrees of freedom. If a rigid body, ABC, be fixed in space, its position can be
defined by the relations of any three points, taken as the apices of a triangle drawn on
the body, to the three intersecting planes of reference. Each point is defined by three
relations, making nine in all ; but the relation of A to B, of B to C, and of C to A are fixed
distances, and hence three of the nine relations are not independent. A rigid body in space
has therefore six degrees of freedom. A pair of compasses has seven degrees of freedom,
but six if the joint is " ankylosed."
If a molecule has n atoms which have a definite and fixed relation with one another, the
number of degrees of freedom cannot exceed 3m. If a gas has but one atom per molecule,
then the molecule will have three degrees of freedom ; if two atoms at a fixed distance
apart — dumb-bell fashion — five degrees of freedom ; and if the atoms are so arranged such
that (i) the atoms are at the apices of an imaginary triangle, there will be 3m — 3 = 6 degrees
of freedom; (ii) if the three atoms are situated as if they were in one fixed line, there will
be five degrees of freedom ; and (iii) if the atoms are fixed so that two of the atoms can
oscillate about a central atom, there will be 3m — 2 = 7 degrees of freedom. For more
complex molecules, and for more complex movements, the number of degrees of freedom
will be greater. The six degrees of freedom of a rigid body, free to move in space, can be
resolved into three translatory movements parallel to three fixed and intersecting planes of
reference, and three rotations about the same axes.
The kinetic energy of the translatory motions of the molecules of a gas is
measured by the pressure, and is analogous with Helmholtz's free energy. The
molecule itself, however, may be a complex system of two or more atoms capable
of rotation about their centres of mass ; and possibly also each is capable of rotation
about its centre of mass ; and possibly also each is capable of rotatory motion.
The kinetic energy absorbed by the motions of this secondary sj^stem is usually
called the internal energy, and is analogous with Helmholtz's bound energy,
although a small proportion of the internal kinetic energy of gases is also vibratory
or oscillatory energy.
The ratio of the two specific heats of gases makes it probable that during the colli-
sions between the molecules, there is a constant exchange and re-distribution of the
energy between the translatory and the internal energy ; but for every gas, a con-
stant ratio is preserved between the two forms of energy, namely, that portion of the
energy which is manifested in the translatory motions of the molecules between their
encounters, and that portion which is concerned in the rotatory and other motions.
The energy distributed between each form of motion probably preserves a constant
ratio to the total energy. An important assumption can now be introduced :
During the fortuitous collisions of the molecules of a gas, the total kinetic energy
K is divided equally among the n degrees of freedom of the molecules, so that the
kinetic energy of translatory motion for each degree of freedom is SK/n ; and for
monatomic molecules, each with three degrees of freedom, n=3, the total kinetic
energy is k. This hypothesis is sometimes called Maxwell's distribution theorem,
because J. C. Maxwell (1859) applied the proposition to systems of rigid particles.
L. Boltzmann (1861) extended the principle to particles which were not rigid, but
which were regarded as complex systems having great numbers of degrees of freedom,
and hence the hypothesis is also called Maxwell-Boltzmann's distribution theorem.
L. Boltzmann showed that this is the most probable distribution of the energy in
an aggregate of a large number of molecules, provided that all the values for the
coordinates and corresponding momenta of a single molecule are equally probable.
This hypothesis has been the subject of much discussion,*^ and even though the
hypothesis is not now considered to be of general application, it has been an impor-
tant stimulus to investigation.
Since translatory motion involves only three degrees of freedom, it follows from
THE KINETIC THEORY OF ATOMS AND MOLECULES 793
Maxwell-Boltzmann's theorem, that if a molecule has n degrees of freedom, the
translatory energy of the molecules of the gas with mean velocity 7, will be equal to
SK/n, where K is the total kmetic energy |miVF2. Hence, the kinetic energy of the
translatory motion can be written 2,Kln=^lmNV'^, and since the kinetic energy,
j)v=lmNV^, it follows, l{\mNV^), or pv=2Kln. The work which a gas can do
when it expands adiabatically— that is, without receiving or giving out heat — is
equal to its total kinetic energy, which in turn is equal to pv^y—l), where y repre-
sents the ratio of the two specific heats of the gas. Since K{y—l)=pv and
pv=2K/n, it follows that by substituting for pvy
y-'^l
(6)
This expression enables (i) the ratio of the two specific heats of a gas to be computed
when the degree of freedom of the molecules is known ; and conversely, (ii) the
degree of freedom when the ratio of the two specific heats is known. The simplest
possible case of a gas with monatomic molecules, each with three degrees of
freedom, furnishes a gas with y=lf . The value found for mercury by A. Kundt and
E. Warburg (1876) 7 is 1-66 ; and the constant has the same value for members
of the argon family. If the three degrees of freedom concerned in the translatory
movements of a molecule between each encounter, be deducted from n, the total
number of degrees of freedom, then n—3 will represent the number of degrees of
freedom of the internal motions. The following table shows the result with a
few gases :
Table XIII.- — Relation between Specific Heats and Degrees of Freedom of
THE Molecules of some Gases.
Gas,
Observed y
n
n-3
Mercury ....
1-67
3-00
0 00
Helium
1-65
3-07
0-07
Oxygen
1-40
4-98
1-98
Carbon monoxide .
1-41
4-94
1-94
Hydrogen chloride
1-39
506
2-06
Carbon dioxide
1-30
6-67
3-67
Carbon disulphide
1-24
8-40
5-40
Ethane
M8
11-10
8-10
Carbon tetrachloride
MS
15-40
12-40
With diatomic molecules, n=5 and y=l*4. This corresponds with a molecule
whose shape and structure are symmetrical about one axis. Such a molecule would
be formed by the union of two spherical atoms, or of two atoms not necessarily
spherical, but each spherical about one axis and both axes corresponding with the
axes of the molecule. This conclusion agrees with the observed values of y for ox}^-
gen=l-4, as well as for the diatomic gases, nitrogen, hydrogen, carbon monoxide,
nitric oxide, and hydrogen chloride. For superheated steam, the observed value of
y is 1-3, which virtually corresponds with a molecule with six degrees of freedom.
There are some objections against a rigid application of the hypothesis :
(a) Intermolecular attraction has been neglected in developing the theory.
With gases like steam, this attraction may reach some magnitude, and it would
reduce the specific heat ratio, because instead of n, it would be necessary to
substitute n-{-d in equation (6), where disa. small positive quantity.
(b) If some of the molecules of a gas are dissociated, a gas with diatomic mole-
cules would have a value of y ranging between that appropriate for a diatomic gas
with five, and that for a monatomic gas with three degrees of freedom. The gradual
rise of the specific heat ratio with temperature also agrees with the assumption that a
greater and greater proportion of the molecules decrease in complexity owing to
794
INORGANIC AND THEORETICAL CHEMISTRY
dissociation with rise of temperature. The converse of this applies to molecular
aggregation which would give rise to molecules with a more complex structure, and
thus increase the value of n and decrease that of y.
(c) If the bond which holds the atoms of a molecule together behaves as if the
atoms are held together neither with perfect rigidity nor with perfect freedom, the
degrees of freedom may not all have the same value, and moreover another degree
of freedom would have to be added for, say, a diatomic molecule in which the atoms
do not remain at an invariable distance apart. The comparison of the ratio of the
two specific heats for transparent and coloured diatomic gases — Table XIV — shows
that the former have values of y in the neighbourhood of 1"4 corresponding with five
degrees of freedom, while with the latter, the value of y is 1'3, corresponding with
seven degrees of freedom, and it is possible that molecular aggregation or the so-called
loose-jointing of the molecules accounts for the discrepancy.
Table XIV. — Comparison of the
AND
Ratio of the Specific Heats of
CoLOUBED Gases.
Transparent
Transparent gases.
Coloured gases.
7
n
n-3
7
n
n-3
Hydrogen .
Hydrogen chloride
Hydrogen bromide
Hydrogen iodide .
1-40
1-39
1-42
1-40
5-0
51
4-8
5-0
21
2-1
1-8
2 0
Chlorine
Bromine
Iodine
Iodine monochloride
1-33
1-29
1-29
1-31
61
6-9
6-9
6-5
31
3-9
3-9
3-5
(d) The addition of heat may also set up motions other than those described by
the degrees of freedom of the molecules indicated above. Other kinds of motions may
be set up by, say, a reaction with the sether as indicated by the emission of radiant
energy. Even the atoms of monatomic gases may be capable of internal oscillatory
motions as is made probable by the complexity of their spectra ; but these motions
require so small an expenditure of energy that its amount may be neglected in com-
parison with the kinetic energy of translatory motion Further, adds 0. E. Meyer,8
It does not appear impossible that the ratio y = r67 should be found in the case of
chemically compound molecules also, if the connection of the atoms is so firm that internal
motions are excluded.
Attempts to connect the values of y or ^ with the number of atoms in the mole-
cule have not been successful. A. Naumann (1867) ^ suggested that n—3 is identical
with the number of atoms in the molecule ; and J. J. Thomson (1893) added that only
when the atoms are symmetrically arranged will n — 3 be proportional to the
number of atoms per molecule. J. W. Capstick (1894-5) found that the nature
as well as the number of atoms is of importance. Thus, with the chloro-derivatives
of methane, CH4, J. W. Capstick found
y . . . .
n .
This shows that the substitution of atoms of chlorine for atoms of hydrogen, step
by step, produces a perceptible rise in the value of n, although the number of atoms
remains unchanged ; and K. Strecker (1881-2) pointed out that while one halogen
can replace another in the hydrogen haloids without any change in value of n, the
substitution of a second halogen for the hydrogen raises the value of n.
CH4
CHjjCl
CH2CI2
CHCI3
1-164
CCI4
1-313
1-279
1-219
1-130
6-4
7-2
9-0
13-0
15-4
References.
I R. Clausius, Pogg. Ann., 100. 377, 1857 ; Phil. Mag., (4), 14. 108,
Kinetic Theory of Oases, London, IIG, 1899.
1857 ; 0. E. Meyer, The
THE KINETIC THEORY OF ATOMS AND MOLECULES 795
2 0. E. Meyer, Die kinetische Theorie der Gase, Breslau, 1899 ; London, 1899.
3 F. Lindemann, Nature, 38. 458, 578, 1888.
* C. H. D. Boedeker, Liebig's Ann., 104. 205, 1857 ; H. Buff, ib., 115. 306, 1864 ; A. Naumann,
ib., 142. 284, 1867 ; 0. Pilling, Ueber die Beziehungen der Wdrmecapacitdt der Gase zu der zwischen
Atomen wirkenden Krdjten, Jena, 1876 ; H. T. Eddy, Proc. Mech. Inst. Ohio, 42. 82, 1883 ;
F. Richarz, Wied. Ann., 48. 476, 1893 ; H. Staigmuller, ib., 65. 655, 1898.
5 J. C. Maxwell, Nature, 11. 357, 1875 ; 16. 242, 1877 ; H. W. Watson, A Treatise on the Kinetic
Theory of Gases, Oxford, 27, 1876 ; L. Boltzmann, Sitzber. Akad. Wien, 74. 553, 1877 ; Phil.
Mag., (5), 3. 320, 1877 ; A. Roite, Nuovo Cimento, (3), 2. 61, 1877 ; A. Violi, ib., (3), 14. 183, 1884 ;
Atti Accad. Lincei, (3), 7. 112, 1883 ; G. de Franchis, ib., (4), 1. 203, 331, 371, 1886 ; C. V. Burton,
Phil. Mag., (5), 24. 166, 1887.
« J. C. Maxwell, Phil. Mag., (4), 19. 19, 1860 ; (4), 35. 129, 185, 1868 ; L. Boltzmann, »6.,
(5), 23. 305, 1887 ; (5), 35. 153, 1893 ; Sitzber. Akad. Wien, 58. 517, 1868 ; 63. 397, 1871 ; 66.
275, 1872 ; 72. 427, 1875 ; 74. 503, 1876 ; 78. 7, 1878 ; 84. 136, 1881 ; 95. 153, 1887 ; 96. 89J,
1887 ; Wied. Ann., 8. 653, 1879 ; 11. 529, 1880 ; Vorlesungen ijther Gastheorie, Leipzig, 1836 ;
P. G. Tait, Phil. Mag., (5), 21. 343, 433, 1886 ; (5), 25. 38, 172, 1888 ; S. H. Burbury, ib., (5),
21. 481, 1886 ; (5), 25. 129, 1888 ; (5), 30. 298, 1890 ; (5), 37. 143, 1894 ; A Treatise on the Kinetic
Theory of Ga^es, Cambridge, 1899 ; Lord Rayleigh, ib., (5), 32. 424, 1891 ; (5), 49. 98, 1900 ; Lord
Kelvin, ib., (5), 33. 446, 1892 ; J. H, Jeans, The Dynamical Theory of Gases, Cambridge, 1916 ;
H. W. Watson, A Treatise on the Kinetic Theory of Gases, Oxford, 1893 ; A. C. Brown, Nature,
32. 352, 533, 1885 ; W. M. Hicks, B. A. Eep., 905, 1885.
' A. Kundt and E. Warburg, Pogg. Ann., 157. 353, 1876.
8 O. E. Meyer, The Kinetic Theory of Gases, London, 121, 1899.
» A. Naumann, Liebig's Ann., 143. 284, 1867 ; K. Streeker, Wied. Ann., 13. 20, 1881 ; 17.
85, 1882 ; J. W. Capstick, Proc. Roy. Soc, 57. 322, 1895 ; Phil. Trans., 185. 1, 1894 ; 186. 564,
1895 ; J. J. Thomson, Watts' Dictionary of Chemistry, London, 1. 89, 1893 ; J. H. Jeans, The
Dynamical Theory of Gases, Cambridge, 201, 1916.
§ 11. The Molecular Heats ot Gases
It follows that if n denotes the number of degrees of freedom per molecule of a
eas, and if K denotes the increase in the kinetic energy when the temperature is
raised from T° to T°-^l°, the total increase in the kinetic energy for N molecules
will be \nNK. If Cy denotes the specific heat of a gas, and M the molecular weight,
it follows that if the change of temperature be accompanied by no other than a
change in the kinetic energy of translatory motion, \nNK=NMCv, and hence,
\nK=MC^. From J. J. Wat erst on' s hypothesis, if two gases be raised through the
same range of temperature, the increase in the kinetic energy of each gas will be
the same, and remembering that Z is a constant, MCv={kJ^)n, which means that
the product of the specific heat and the molecular weight of a gas is proportional to
the number of degrees of freedom of its molecules, and equal to the product of the
number of degrees of freedom of the molecules of a gas and a constant. The constant
K is evaluated by substituting known values of C^, M, and n in the equation ; and
\K works out to be very nearly unity, as indicated in Table XV. Accordingly,
Molecular heat =Ot;ilf=w . . . • (7)
The degrees of freedom calculated from this equation agree with those obtained
for a number of gases in other ways. This relation makes the atomic heats of the
monatomic gases equal to 3.
The constancy of the molecular heats over a range of different gases implies
that their molecules have the same number of degrees of freedom ; and conversely,
the approximate number of degrees of freedom of the molecules of a gas is given
by the product of the specific heat and molecular weight. That is, the molecular
heat o! a gas is almost numerically equal to the degree of freedom of its molecules,
or about one calorie per degree of freedom. The observed value for all monatomic
gases excepting helium agrees with theory. The two-atom molecules with five
degrees of freedom seem to behave like a rigid dumb-bell when the gas is heated,
in that the possible vibratory motion of two atoms along the line joining them does
not appear to represent a degree of freedom, for the motions appear to be distributed
796
INORGANIC AND THEORETICAL CHEMISTRY
solely in two rotational and three translational movements. The results are con-
firmed by the molecular heats of the diatomic permanent gases. The molecular
heats of chlorine and bromine are about a calorie too high for agreement with the
hypothesis. This possibly corresponds with to-and-fro vibrations of the atoms
along an axis joining the two atoms.
Table XV. — Relation between Molecular Heats and Degrees of Freedom.
Argon
Mercury
Helium
Hydrogen
Nitrogen ,
Oxygen
Carbon monoxide
Hydrogen chloride
Nitric oxide, NO
Chlorine
Bromine .
Iodine
Ammonia
Carbon dioxide
Specific heat,
Cv
00731
00147
0-7465
2-4263
01721
01 544
0-1746
01392
0-1650"
0-0873
0-0428
0-0257
0-3803
0-1486
Molecular
weight, M
39-92
200
3-96
2
28
32
28
36-5
30
71
160
254
17
44
C^M
2-92
2-94
2-95
4-85
4-82
4-95
4-89
5-08
4-95
6-20
6-84
6-52
6-46
6-54
Degrees of
freedom, n
Constant,
0-97
1-00
0-98
0-97
0-96
0-99
0-97
0-95
1-01
1-03
1-14
1-09
108
0-94
L. Boltzmann ^ expresses these results another way, the thermal capacity of a
gas is equal to the product of its mass by its specific heat C-o ; consequently, if w
denotes the atomic weight, i.e. the molecular weight of a monatomic gas, the thermal
capacity or the atomic heat will be wC^. It has been shown that the mean kinetic
energy of a gram-molecule of a gas is equal to I'pv or ^RT, and if the gas at —273°
has no kinetic energy of translatory motion, the kinetic energy of a gram-molecule
of the gas at 0° will be 273 wCv=^RT ; but R=2 cals. nearly ; accordingly, the
atomic heat of a monatomic gas per 1° will be nearly 3. This agrees with the
results in Table X. In raising the temperature of a polyatomic gas from —273°
to 0°, intramolecular work may be done. Accordingly, if M denotes the molecular
weight of the gas, MC^ will be equal to ^i?(l+e), where e denotes the fractional part
of the kinetic energy which is spent in doing intramolecular work. For a diatomic
gas, €=f , and MC^—^ nearly, or the atomic heat of a diatomic gas is nearly 2*5 ;
similarly, L. Boltzmann has shown that for triatomic gases, €=1, or MCv=Q, or
ivCv=2 nearly.
The effect of temperature and pressure on molecular heats. — The specific heat
of a gas at constant pressure, Cp, increases with a rise of temperature such that the
mean specific heat Cp between 0° and 6° is Cp^a+hO ; or Cp=a-\-hd-]-cd^, where
a, h and c are constants. E. Wiedemann's ^ value for carbon dioxide between
0° and 200° is 0-1952 +0-000229^ ; and L. Holborn and F. Henning's value between
0° and 800° is 0-0208+0'0001384^+0-00000005^2. H. V. Kegnault thought that
pressure had no effect on the thermal capacity of hydrogen, carbon dioxide, etc.,
but J. Joly and S. Lussana ^ have shown that the specific heat increases with in-
creasing pressure when the temperature is constant. Thus, the value for carbon
dioxide increased 38 per cent, when the pressure rose from one atm. to 41 atm. The
observed results can be represented by analogous formulae, Cp=a-{-h(p—l) orCp—a
+?^(p-l)+c(^-l)2. Thus, for carbon dioxide, S. Lussana found that Cp=0-20130
+0-(X)19199(2?— 1). The specific heat at constant volume varies with temperature
and pressure somewhat similarly to the results for the specific heat at constant
pressure
W. Nernst's measurements (1911) show that the molecular heat of chlorine
rises from 5-85 at 0° to 7'0 as the temperature rises to 12(K)°, and still another calorie
THE KINETIC THEORY OF ATOMS AND MOLECULES
7U7
higher at about 2000°. This corresponds with the addition of two new degrees of
freedom. The gases with a molecular heat approaching 6 are easily dissociated, and
therefore the number of degrees of freedom of molecules whose atoms are joined
by strong chemical forces is less than when the atoms arc united by weaker forces.
The effect of temperature on the molecular heats of a few eases is further illustrated
by Table XVI.
Table XVI.-
-The Effect of Temperature on the Molecular Heats of Gases at
Constant Volume.
Theoretical
Temperature ("C).
Gaa.
value for
rigid mole-
cules.
0°
100°
300°
500°
1200°
2000°
Argon
3
2-98
2-98
2-98
2-98
2-98
2-98
Hydrogen .
5
4-75
4-78
502
5-20
5-80
6-50
Nitrogen .
5
4-90
4-93
5-17
5-35
6-00
6-70
Steam
6
5-93
5-97
6-45
6-95
8-62
1310
Carbon dioxide .
6
6-80
7-43
8-53
9-43
1100
11-50
Ammonia .
6
6-62
6-82
7-41
8-52
Similar values were obtained for oxygen, hydrogen chloride, and carbon monoxide
as those indicated in the table for nitrogen ; and sulphur dioxide furnished numbers
similar to those for carbon dioxide. According to A. Eucken (1912), * the molecular
heat of hydrogen falls from 5 to 3 calories as the temperature falls to —213°, and
diatomic gas then behaves as if it were monatomic. At still lower temperatures,
—238°, the molecular heat of hydrogen and helium at constant volume fall still
lower than 3. M. Trautz (1913) has shown that the molecular heat at constant
pressure decreases as the temperature diminishes from 20° to —180°. For
instance,
Nitrogen.
Carbon monoxide.
Oxygen.
Hydrogen.
Helium
18° to 20°.
6-983
7-006
6-980
6-860
4-993
-180° to -181°
7-162
7-244
7-300
6-330
4-934
The molecular heat of helium thus exhibits a positive temperature coefficient,
whereas, if helium were a monatomic gas, and nothing has been overlooked, its
molecular heat should not be affected by variations of temperature. The decrease
with hydrogen at low temperatures is explained by M. Trautz (1913) by assuming
that the hydrogen is converted into another modification with the properties of
a monatomic gas whereby it loses one or more degrees of freedom. A. H. Compton
(1915) made a similar assumption for solids.
References.
1 L. Boltzmann, Sitzher. Akad. Wien, 63. 731, 1871.
2 H. V. Regnault, Mem. Acad., 26. 58, 1862 ; Pogg. ^7W.,157. 1, 1876 ; Wicd. Ann., 2. 195,
1877 ; E. Wiedemann, L. Holborn, and F. Henning, Ann. Physik, (4), 18. 739, 1905 ; (4), 23.
809, 1907 ; L. Holborn and L. Austin, Sitzhcr. Akad. Berlin, 115, 1905.
3 J. Joly, Proc. Roy. Soc, 41. 352, 1886 ; 55. 290, 1894 ; Phil. Trans., 182. A, 73, 1892 ; 185.
A. 913, 1894 ; S. Lussana, Nuovo Cimento, (3), 36. 5, 70, 130, 1894 ; (4), 2. 327, 1895 ; (4), 3. 92,
1896; (4), 6. 81, 1897; (4), 7. 61, 365, 1898; E. Mallard and H. Je Chatelier, Compt. Rend., 93.
962, 1014, 1076, 1881 ; P. Vieille, ib., 96. 1358, 1883 ; M. Berthelotand P. Vieille, ib., 98. 545, 601,
770, 852, 1884 ; H. le Chatelier, ib., 104. 1780, 1887 ; M. Pier, Zeit. phys. Chem., 62. 385, 1908;
66. 759, 1909 ; Zeit. Elektrochem., 15. 536, 1909 ; 16. 897, 1910 ; N. Bjerrura, ib., 17. 731, 1911 ;
18. 101, 1912 ; Zeit. phys. Chem., 79. 513, 1912 ; 81. 281, 1912 ; W. Nemst, Phys. Zeit., 13.
1064, 1912.
* A. Eucken, Sitzbcr. Akad. Berlin, 141, 1912; M. Trautz, Phys. Zeit. A^.UHS, 1913; Ber.
dent. phys. Ges., 15. 969, 1913 ; A. H. Compton, Phys. Rev., (2). 6- 377. 1915.
798
INORGANIC AND THEORETICAL CHEMISTRY
§ 12. The Specific Heats of Elementary Solids— Dulong and Petit's Rule
Les atomes de tous les corps simples ont exactement la m^me capacity pour la chaleur.
— P. L. Dulong and A. T. Petit (1819).
In 1819, P. L. Dulong and A. T. Petit published an account of their Recherches
sur quelques points importanis de la theorie de la chaleur ; ^ they found that while
different substances have different capacities for heat, the atomic heats of nearly
all the elements then known were the same. Atomic heat is a convenient term for
the thermal capacity of a gram atom of an element ; it is the product of the
specific heat and the atomic weight of the element. P. L. Dulong and A. T. Petit
determined the specific heats of different solid elements, at constant pressure, and
obtained this remarkable result : The product of the atomic weight and the specific
hea.t of an element has nearly always the same numerical value — Dulong and
Petit's rule. This means that the atomic heats or the thermal capacity of the atoms
of the elements are approximately the same ; it requires the same amount of heat
to raise every atom, no matter of what kind, one degree of temperature. The
relation is usually expressed :
Atomic heat, Cw=^6 approximately
(8)
where w denotes the atomic weight of the element, and C its specific heat — at con-
stant pressure. The rule is remarkably near the truth for the solid elements at
ordinary temperatures, but it fails with the gaseous elements and a very small
number of solid elements. In illustration, a few examples selected at random from
a list containing nearly 50 elements for which data are available, are indicated in
Table XVII.
Table
X
VII.' — Atomic Heat of Eleme::ts.
Element.
Specific heat.
Atomic weight.
6-94
Atomic lieat.
Lithium ....
0-9408
6-53
Silver
0-0559
107-88
6-03
Gold
0-0304
197-2
6-25
Copper .
0-0923
63-57
5-88
Bismuth .
0-0305
208-0
6-34
Lead
0-0315
207-10
6-52
Aluminium
0-2143
271
5-81
Iron
0-1098
55-85
612
Uranium
0-0277 '
238-5
6-61
H. V. Kegnault extended and confirmed P. L. Dulong and A. T. Petit's observa-
tion. The average value of the constant is 6"36 or 6*4. The atomic weights range
from 6*94 to 238'5. and yet, when multiplied by the respective specific heats, the
products are nearly constant. Kigorous agreement cannot be expected. The di-
vergencies are too large to be accounted for by the inevitable errors of observation
involved in measuring the specific heats, but the very irregularity of the divergencies
leads to the view that Dulong and Petit's law approximates to a truth, and that
the observed differences are due to disturbing effects which are not functions of the
atomic weight. According to H. Kopp (1865) the atomic heats (determined
indirectly) of the following elements in the solid state are exceptionally low :
0
H
F
B
c
Si
S
P
4-0
2-3
50
2-7
1-8
3-8
5-4
5-4
Atomic heat
Beryllium also gives a low value. The atomic heats of the other forty elements
examined by H. Kopp were nearly all in the vicinity of 6. The elements with ex-
ceptionally low atomic heats usually form gaseous or readily volatile compounds,
THE KINETIC THEORY OF ATOMS AND MOLECULES ^ 799
and they have small atomic weights. Of these, those which exhibit the greatest
deviation usually have the smallest atomic weights, while those with the largest
atomic weights approximate most nearly to the normal value. Magnesium and
aluminium have small atomic weights, and though their atomic heats are considered
to be normal, they are on the low side. Sodium and lithium are exceptional in
possessing low atomic weights and normal atomic heats.
Attempts have been made to explain the divergencies (i) as a result of the fact
that the temperatures at which the specific heats have been determined stand in
a different relation to their fusion temperatures. Lead, for instance, at the tem-
perature of boiling water, is much nearer its fusion temperature (327°) than iron
is to its fusion temperature (1530°) ; but in the case of many elements, the specific
heat does not change very much with changes of temperature not far removed from
atmospheric. Another explanation assumes (ii) that the divergencies are due to
differences in the configuration or orientation of the molecules or atoms of the solid
elements. According to F. Richarz,^ the fact that solid bodies with low atomic
volumes — ^beryllium, boron, and carbon — ^have low values for the atomic heats
shows that, on a rising temperature, the displacement of the atoms, relatively to
their mutual distances apart, is large. This would make the thermal capacities
of these elements more sensitive to changes of temperature than when the dis-
placements are relatively small. This is also confirmed by the observation that the
allotropic forms of an element with the smaller atomic volume or the larger specific
heat have the lowest specific gravity.
The influence of pressure on atomic heats. — Measurements of the specific heats
of solids, Cp, are made at a constant pressure, and the results are probably not
very different from the specific heat, C^, at constant volume at very low temperatures,
but at higher temperatures the difference is significant. For instance, the observed
value of Cp for silver at 100° K. is 4*86, the computed value for C^, is 4*72 ; and at
589° K., Cp is likewise 6-64, and Ct, is 5-92. As in the case of gases, the difference
between the two depends upon the amount of work required to compress the heated
solid back to the volume it occupied before it was heated.
The internal energy U of compressed gases, or the heat developed during ex-
pansion without doing external work has an appreciable value, and is usually
negative, and when work W is done against an external pressure p, during an ex-
pansion from a volume Vi to V2, W={v2—Vi)p, and,regarding the volume as constant,
dW={v2—Vi)dj). Consequently, the heat Q developed during the change, follows
from Helmholtz's equsLtion, Q=W-^U=-(v.2—ii) T. dp jdT ; and the heat dQ
developed during an infinitesimal change, V2—Vi=^dv, will be dQ=T(dfldT)dv, or
(si=<ia («)
Again,3 by the partial differentiation of Q, p, v, and T,
dv
,df
These results obtain as a purely mathematical operation like nmltiplication or
division, and altogether apart from any meaning the symbols may have. Since
dQIdT denotes the specific heat at constant pressure or constant volume,
by substituting for dQ/dv from equation (9). E. H. Amagat applied this relation
to his data on highly compressed gases, and his results with carbon dioxide at 50°
are shown in Table XVIII.
800 -
INORGANIC AND THEORETICAL CHEMISTRY
Table XVIII.— Values of ^^, • .-^ fob Cakbon Dioxide.
dp
dv
dp . dv
Pressure.
df
df
dT dT
60
0-265
1
0-0001300
0-1353
60
0-370
0-0001250
01815
70
0-505
0-0001428
0-2831
80
0-689
0-0001687
0-4563
90
0-960
0-0002133
0-8040
100
1-460
0-0002020
1-1105
The numbers in the last column have been divided by the mechanical equivalent of
heat to convert them into gram- calories. On multiplying and dividing the right
member of (10) by VQdv^ it furnishes
%\
C"n C« — 1
n dvY
dT/
"i^dv
Vodp
(10)
where the coefficient of cubical expansion, 3a=dvlvQdT ; and the compressibility,
P=—dvlvQdp, and Vq is the volume in c.c. at 0°. These equations have been
employed by M. Margules (1888), E. H. Amagat (1896), E. Gruneiscu (1906), and
others. 4 The preceding discussion reduces to
C», — (7^:
VoT
(11)
which is a well-known relation in thermodynamics. All the magnitudes — specific
heat at constant pressure, Cp ; coefficient of linear expansion, a ; the volume at
0°, Vq ; and the coefficient of compressibility, jS — , can be determined experimentally.
If Cp and Cv represent the corresponding atomic heats, Vq will represent the
atomic volume ; and if the unit of pressure be megadynes per sq. cm., and the
unit volume is 1 c.c, the calculated values of Cp and C^ will be in mega-ergs
per degree. Since 41*82 x 10^ ergs, or 41-82 mega-ergs are equivalent to one calorie.
C« G|; =
(3^2
Tvq mega.-eTg& ; Cp^C^-
(3a) %r
41 -82^3
cals.
The data in Table XIX were compiled by G. N. Lewis, and they show that
within the limits of experimental error, the atomic heats of these elements at con-
stant volume and 20°, are the same for all the solid elements whose atomic weights
are greater than potassium. The mean of these values of the atomic heats at con-
stant pressure (Dulong and Petit's constant) is 6*2, and at constant volume 6"9.
The average deviation for the former is 0'26, and excluding the first four elements,
0*18 ; the average deviation for the latter is 0*15, or, excluding the first four elements,
0'09. F. Biirki found that the differences Cp—C^ for chemically related elements are
nearly the same ; and that for the halides of the elements of group I. in the periodic
table, the requirements of H. Kopp's rule are more nearly fulfilled by Cp than by C^
When the coefficients of compressibility and of cubical expansion at the
desired temperature are not known, the empirical expression of E. Griineisen,
Ci,=Cp—0'0214:Cp^T ITm, can be used when Tm denotes the melting point of the
solid. Here VqIP of (28) is assumed to be constant, and a^ to be proportional to the
atomic heat. Still further, the empirical expression of A. Magnus and F. A. Linde-
mann,^ C,,— C*^— aTS, gives results of sufficient accuracy for many purposes. The
constant a is evaluated by putting Cj,=3R at sufficiently high temperatures.
THE KINETIC THEORY OF ATOMS AND MOLECULES
801
The influence o! temperature on atomic heats.— Silicon, boron, beryllium, and
carbon at ordinary temperatures, have atomic heats represented respectively by
3-8, 27, 3-4, and 1'8. ' From the point of view of Dulong and Petit's rule, these
numbers are low. H. F. Weber (1874) found that the atomic heats of carbon,
Table XIX.— The Two Atomic Heats
OF THE
Elements
AT 20^
Element.
Atomic
weight.
Atomic
volumes, Vq
3oxlO«
pxlO«
Cp-cr.
Cp
C.
Sodiiim .
230
23-7
15-4
72
0-6
6-9
6-4
Magnesium
24-32
13-3
2-7
25
0-2
6-0
5-8
Aluminium
27-1
10-1
1-3
23
0-2
5-8
5-6
Potassium
39-1
45-5
31-5
83
0-6
7-1
6-6
Iron
55-84
7-1
0-40
10
01
6-0
5-9
Nickel .
68-68
6-7
0-27
13
0-2
6-1
5-9
Copper .
63-67
7-1
0-54
16
0-2
5-8
5-6
Zinc
65-37
9-5
1-5
29
0-3
6-0
5-7
Palladium
106-7
9-3
0-38
11
0-2
6-1
5-9
Silver .
107-88
10-3
0-84
19
0-3
61
5-8
Cadmium
112-40
13-0
1-9
28
0-3
6-2
5-9
Tin
118-7
16-2
1-7
22
0-3
6-4
6-1
Antimony
120-2
17-9
2-2
11
0-1
6-0
5-9
Iodine .
1 126-92
25-7
13-0
84
0-9
6-9
6-0
Platinum
1 195-2
9-1
0-21
9
0-2
6-1
5-9
Gold
j 197-2
10-2
0-47
14
0-3
6-2
5-9
Thallium
i 204-0
17-2
2-6
28
0-3
6-4
6-1
Lead .
207-2
18-2
2-2
29
0-4
6-3
5-9
Bismuth
208-0
21-2
2-8
13
0-1
6-3
6-2
boron, and silicon approximate closer and closer to the normal value, the higher the
temperature ; and A. G. Worthing has shown that the atomic heat of carbon rises
to wCp=5'35 at 1200° K., and to 6'05 at 2000° K. L. F. Nilson and 0. Pettersson
(1880), and T. S. Humpidge (1886) obtained a similar conclusionfor beryllium.<5 This
is illustrated by the diagram. Fig. 7, which represents the effect of a rising tempera-
ture on the atomic heats of the elements in question. The specific heats of these
elements increase with temperature until a point is
reached at which they are nearly constant. W. A.
Tilden's examination of the influence of tempera-
ture on atomic heats led him to conclude that there
is no one condition or set of conditions under ivhich
Dulong and Petit's rule is true for all the elements.
If carbon, boron, silicon, and beryllium be regarded
as exceptional, the mean specific heats between 0°
and 100° may be arbitrarily selected as a standard
for the best results. E. H. and E. Griffiths suggest
that the empirical relation Cw^'^^^^^'Bd^ fits the
specific heats of the metals at 0° better than Dulong
and Petit's rule ; and by extrapolation, they infer
that the atomic heats of the elements at absolute
zero will have the mean value 4*813, but this
induction does not fit the facts. There does not appear to be a true upper limiting
value in the sense indicated by Dulong and Petit's rule, for several substances
possess greater atomic heats even at 1000°; for example, A. G. Worthing has
shown that while the atomic heat of tungsten at constant volume is 5-95 at
moderate temperatures, the value is 6-25 at 1200° K., and 7*35 at 2400° K.
As a rule the temperatures at which the elements exhibit the same atomic heats
decrease as the atomic weights increase.
U. Behn 7 (1898) found that there is a decrease in the atomic heats with falling
VOL. I. 3 F
0" 200° 400' 600" 800° 1000*
Fig. 7.' — Atomic Heat Curves of
Beryllium, Boron, Carbon, and
Silicon.
INORGANIC AND THEORETICAL CHEMISTRY
temperatures, so that when the observed results are plotted, the curves appear as
if they would intersect at low temperatures, but T. W. Richards and F. G. Jackson
determined the specific heats ol many elements between the temperature of liquid
air and ordinary atmospheric temperatures, and found that with the exception of
the elements of low atomic weight, they conform to Dulong and Petit's rule. Further
observations 8 by W. Nernst and his co-workers, J. Dewar and H. K. Onnes, have
confirmed Behn's conclusion ; for while the atomic heats o£ the elements with
abnormally low values increase approximately to about the theoretical value
with a rise of temperature, the atomic heats of all the elements converge
towards zero as the temperature approaches absolute zero. For instance, with
silver :
Temperature
-238°
-228^
-196''
-173°
-73°
-58°
-30-6°
Atomic heat
1-68
2-47
4-07
4-86
5-78
6-01
6-64
The results with carbon (diamond), aluminium, and lead are illustrated by the
graphs in the diagram. Fig. 8. H. K. Onnes and G. Hoist found that the specific heat
of mercury is 0-00142 between 4-26° E. and 6-48° K., and 0'000534 between 293° K.
and 3 "97° K. J. Dewar's determinations of the specific heats of fifty -three elements
at temperatures between the boiling points of hydrogen and nitrogen when plotted
with the atomic weights, gave a periodic curve closely resembling the atomic volume
curve of L. Meyer; the
specific heats at ordinary
temperatures give a hyper-
bolic not a periodic curve —
Fig. 5, Cap. VI.
These results make it
clear that Dulong and Petit's
rule is a limiting rule to
which these elements ap-
proximate when the tern-
perature is high enough,
but, in virtue of the marked
relation between the thermal
capacity and temperature,
Dulong and Petit's constant
can be obtained for diverse
By the same procedure,
heat
He
-273°
Fig. 8.-
'173°
-73° 27° 127° 177°
•Atomic Heat Curves of Lead, Aluminium, an^ the
Diamond.
elements if arbitrary temperatures be employed.
R. Lammel^ suggests that it would be possible to use for the atomic
any number between 3 and 9-5 by arbitrarily varying the temperature,
suggests that the elements should be compared at their melting points and
shows that the atomic heats then lie between 9 and 10 calories as indicated
in Table XX; there are a good many exceptions. As a matter of fact, the
Table XX. — ^Atomic Heats of the Elements at their Melting Points.
Elements.
Atomic weight.
Melting point, Tm.
Specific heat at Tm.
Atomic heat.
Lithium
7
190°
1-3
9-45
Aluminium
27
700°
0-35
9-45
Sodium
23
100°
0-36
8-28
Sulphur
32
120°
0-25
8-00
Copper
64
1100°
0-145
9-28
Nickel
59
1600°
0-166
9-79
Zinc
65
420°
0-142
9-23
Bromine
80
7°
0-114
9-12
Silver
108
1040°
0-082
8-87
Cadmium
112
315°
0-066
7-39
Lead
207
330°
0-0413
8-55
THE KINETIC THEORY OF ATOMS AND MOLECULES 803
atomic heat indicates how rapidly heat, dQ/dT, is added per atom per degree rise
of temperature, and a high value for this constant means that the potential
energy of the molecules is being rapidly increased, possibly owing to the dis-
sociation or approaching dissociation of complex into simpler molecules. At low
temperatures where the specific heats are small, S. Pagliani (1915) has pointed
out that the atomic heats of the elements at a given temperature, tend to increase
with increasing atomic weight ; and this is not the case at temperatures near atmo-
spheric. The temperatures at which different elements show the same atomic
heats decrease as the atomic weights become larger, presumably because usually
the larger the atomic weight of the element the less is the strength of the union
between the atoms.
The influence of the state of aggregation on atomic heats.— According to the
kinetic theory, the heat required to raise the temperature of a body is spent (1) in
raising the kinetic energy of the molecules ; (2) in raising the kinetic energy
of the constituent atoms — e.g. in doing chemical work, etc. ; (3) in increasing
the volume of the body against atmospheric pressure ; (4) in overcoming
molecular attractions, etc. The coefficient of thermal expansion of solids is small,
and therefore also the work of expansion of solids against atmospheric pressure is
small, but the work done against molecular cohesive forces is probably large with
liquids and solids. This is illustrated by the comparative large difference in the
specific heats of solid, liquid, and gaseous elements. As a rule, the specific heat
of a substance in the gaseous state is less than it is in the liquid state. Thus, the
specific heat of liquid alcohol is 0'5475 and for the vapour 0*4534 : for ether, the
numbers are respectively 0*5290 and 0'4797. The atomic heat of solid iodine is
6'9, and of the gas, 3*3 ; liquid bromine, 8*56, and bromine gas, 4*7. There are
exceptions — ^for instance, the specific heat of ice is 0*500, liquid water, 1*000, and
steam, 0*477 ; the specific heat of solid mercury is 0*0314, of the liquid, 0*0333, and
of the vapour, 0*0147 ; the numbers for liquid and solid tin are 0*0637 and 0*0559,
and for liquid and solid lead, 0*0470 and 0*0314 respectively. H. Mache^® has
worked out a demonstration that the thermal capacity of a liquid ought to be
nearly double the true thermal capacity of its vapour, so that if C© is the true thermal
capacity of a vapour, the value for the corresponding liquid should be 20^,. This
applies to a number of liquids and gases. According to J. D. van der Waals' theory,
the specific heat at constant volume of any unassociated substance would be the same
in the liquid and gaseous states of aggregation, but the deduction has not been tested
experimentally.
The discontinuity in the atomic heat of metals at their melting point is usually
positive, but, according to I. litaka, the magnitude is generally smaller than
corresponds with the energy of one degree of freedom, i.e. 1 cal. In the case of
lead and tin, the change is negative. In the case of these metals, therefore, it
seems very probable that the molecules are not free to rotate during melting, or
at least that the rotation of the molecules in the liquid state does not increase
with the rise of temperature, i.e. these metals are to be considered as monatomic,
even in the liquid state.
The difference in the specific heats of liquid and solid tnay not he solely determined
hy differences in the states of aggregation, because during the transition, the molecules
7nay become more or less co7nplex. This is known to be very probably the case with
water. Variations in the complexity of the molecules of an element in one state
of aggregation may determine differences in the atomic heats. Thus H. V. Reg-
nault (1866) ^^ found the specific heat of amorphous carbon to be 0*2609 ; graphitic
carbon, 0*2000 ; and the diamond, 0*1470. R. Bunsen also found the specific heat of
ordinary tin to be 0*0559 and of allotropic tin, 0*0545. No difference was observed
between the specific heats of aragonite and calcite. According to F. Richarz (1893),
with carbon, silicon, boron, phosphorus, sulphur, arsenic, selenium, tellurium, and
tin, the allotropic modification of an element with the smaller specific gravity
has the larger specific heat — Richarz's rule ; otherwise expressed, the modification
804
INORGANIC AND THEORETICAL CHEMISTRY
with the smaller atomic volume has the smaller specific heat. Illustrations in
support of this rule are indicated in Table XXI.
Table XXI.- — Effect of Specific Gravity on Specific Heat.
Element.
I Carbon (diamond) .
Carbon (graphite)
Carbon (retort)
(Boron (crystalline)
"(Boron (amorphous)
J Phosphorus (red) .
"I Phosphorus (yellow)
J Arsenic (grey)
I Arsenic (black)
j Sulphur (rhombic)
1 Sulphur (monoclinic)
/ Sulphur (amorphous insoluble)
I Sulphur (amorphous soluble)
/ Selenium (crystalline)
|Seleniimi (amorphous)
r Tellurium (crystalline)
\ Tellurium (amorphous)
/Tin (white)
tTin (grey) .
Specific gravity.
3-52
2-25
1-89
2-49
2-45
3-30
1-83
6-87
4-78
2-06
1-9C
1-89
1-86
4-80
4-30
6-30
6-00
7-30
5-85
Specific iieat.
0-113
0-160
0-204
0-165
0-307
0-183
0-202
0-082
0-086
0-173
0-181
0-190
0-248
0-084
0-113
0-048
0-053
0-054
0-059
(lO'')
(10°)
(68°)
(21°)
(100°)
(51°)
(36°)
(100°)
(100°)
(54^^)
(52°)
(53°)
(50°)
(62°)
(57°)
(100°)
(100°)
(21°)
(18°)
0. Richter (1913) found F. Richarz's rule applicable to binary alloys of lead and
bismuth, and bismuth and tin, excepting for alloys in the vicinity of BiPb, where
it is assumed the formation of the compound BiPb renders the rule inapplicable.
Similar difierences have been noticed with the specific heats of metals in different
physical conditions — e.g. according to H. V. Regnault (1843), the specific heat of
hard tempered steel is 0*1175, and of soft tempered steel, 0-1165 ; while hard
bronze has a specific heat 0'0858, soft bronze has a specific heat 0'0862. Con-
sequently, it must be inferred that the heat does important work other than merely
raising the kinetic energy or temperature of the molecules ; and it therefore appears
strange that the relation pointed out by Dulong and Petit does not exhibit greater
divergencies.
The rectification of atomic weights by Dulong and Petit's rule : Cw=6*0 to
6*4, where w denotes the atomic weight of the element. It will be obvious that if
the specific heat of an element be known, it is possible to compute an approximate
value for the atomic weight. The number so obtained may be useful in deciding
between two numbers which are multiples of a common factor. The method is
obviously only applicable to elements whose specific heat can be determined. In
view of the variation of specific heats with temperature, the usual application of
this law to the rectification of atomic weights " is a rough empirical rule, which,
setting aside silicon, boron, beryllium, and carbon, is only available when the
specific heats have been determined at temperatures, usually and most con-
veniently, between 0° and 100°." The specific heat method of fixing the atomic
weights is not so much used as formerly, because so many other methods which are
more exact are available.
Examples.- — (1) What is the atomic weight of silver assuming that the specific heat is
0-0559 ? Here, 6-0-^-0-0559 = 108 nearly. This is close to the accepted value for the atomic
weight of this element.
(2) Platinum chloride, on analysis, furnished 35-5 grams of chlorine per 48*6 grams of
platinum. The specific heat of platinum is 0-0324, and the atomic weight is approximately
6-0-^0-0324 = 197 nearly. Hence, since 197-5-^486 = 4 nearly, it follows that if the atomic
weight of chlorine is 355, the atomic weight of platinum must be nearly 486 X 4 = 194.
(3) When indium was first discovered, the analysis of its chloride furnished indium,
37-8 ; chlorine, 355. The equivalent of indimn is therefore 37-8. The formula of the
THE KINETIC THEORY OF ATOMS AND MOLECULES 805
chlorine was thought to be InCla, and the atomic weight was accordingly represented 75*6.
The specific heat of the metal was found to be 0-057. Hence, 75-() x 0-057 =4*5. If 75-6
be the correct atomic weight, the product would approximate more closely to 6, and hence
it was inferred that 75-6 is not the correct atomic weight of indium ; rather does the atomic
weight approximate to 6-^0-057 = 112 nearly. If InClg be the formula of the chloride, the
atomic weight will be 37*8 X 3 = 1 13-4, which is the number usually adopted for the atomic
weight of this element.
Refeeences.
1 P. L. Dulong and A. T. Petit, Ann. Chim. Phys., (2), 10. 395, 1819 ; (2), 7, 144, 1817 ; H. V.
Regnault, ih., (2), 73. 5, 1840 ; (3), 1- 125, 1841 ; (3), 9. 322, 1843 ; (3), 26. 286, 1849 ; (3), 38.
129, 1853 ; (3), 46. 257, 1856 ; (3), 63. 1, 1861 ; (3), 67. 427, 1863.
2 H. Kopp, Liebig's Ann. Suppl, 3, 1, 290, 307, 1865 ; F. Richarz, Wied. Ann., 48. 708,
1893 ; 67. 704, 1899 ; Verh. deut. phys. Ges., 1. 47, 1899; J. PaschI, Sitzber. Akad. Wien, 112.
1230, 1903; L. Meyer, Moderne Theorien der Chemie, Breslau, 167, 1884; A. Wigand, Ann.
Physik, (4), 22. 64, 1907 ; A. Bettendorf and A. Wiillner, Pogg. Arm., 133. 293, 1868.
3 ,1. W. Mellor, Higher Mathematics, London, 81, 1913.
" E. Gruneisen, Ann. Physik, (4), 26. 401, 1908; W. Nernst, Zeit. Elektrochem., 17. 819, 191 1 ;
W. Nernst and P. A. Lindemann, ib., 17. 817, 1911 ; M. Margules, Sitzber. Akad. Wien, 97. 1385,
1888 ; E. H. Amagat, Journ. Phys., (3), 5. 114, 1896 ; F. Burki, Helv. Chim. Acta, 2. 27, 1919.
5 A. Magnus and F. A. Lindemann, Zeit. Elektrochem., 16 269, 1910 ; P. Duhem, Compt. Bend.,
143. 335, 371, 1906 ; G. N. Lewis, Journ. Amer. Chem. Soc, 29. 1165, 1516, 1907.
« H. F. Weber, Pogg. Ann., 154. 367, 553, 1875; Phil. Mag., (4), 49. 161, 276, 1875;
H. Moissanand H. Gautier, Ann. Chim. Phys., (7), 7. 568, 1896 ; S. T. Humpidge, Proc. Roy. Soc,
39. 1, 1886 ; A. G. Worthing, Journ. Franklin Inst., 185. 707, 1918 ; H. Lecher, Sitzber. Akad.
Wien, 117. Ill, 1908; 0. M. Corbino, Atti Accad. Lincei, 22. 430, 1913; L. F. Nilson and
0. Pettersson, Ber., 13. 1451, 1880 ; L. Meyer, ib., 13. 1780, 1880 ; W. A. Tilden, Journ. Chem.
Soc, 87. 551, 1905.
7 U. Behn, Wied. Ann., 66. 236, 1898 ; Ann. Physik, (4), 1. 257, 1900 ; W. A. TUden, P^tZ.
Tran^., 194. A, 233, 1900 ; 201, A, 37, 1903 ; 203. A, 139, 1904.
8 W. Nernst, Ann. Physik, (4), 36. 395, 1911 ; Sitzber. Akad. Berlin, 262, 1910 ; 306, 1911;
W. Nernst, F. Koref and F. A. Lindemann, ib., 247, 1910 ; W. Nernst and F. A. Lindemann,
ib., 817, 1911 ; F. Pollitzer, ib,, 5, 1911 ; H. Schimpff, Zeit. phys. Chem., 71. 257, 1910 ; T. W.
Richards and F. G. Jackson, ib., 70. 414, 1910 ; J. Dewar, Proc. Roy. Soc, 76. A, 325, 1905;
89. A, 158, 1914; W. Ew&ld, Ann. Physik, {4), 44, 1213, 1914; T. Estreicher and M. Staniewsky,
Acad. Science Cracow, 8. 834, 1912 ; A. Eucken, Verh. deut. phys. Ges., 15. 578, 1913 ; W. H.
Keeson and H. K. Onnes, Comm. Phys. Lab. Leiden, 143, 147a, 1915; 149a, 1916 ; H. K. Onnes
and G. Hoist, ib., 142c, 1914.
9 R. Lamrael, Anii. Physik, (4), 16. 551, 1905 ; S. Pagliani, Nuow Cimento, (6), 8. 157, 1914;
Gazz. Chim. Ital, 45. ii, 317, 1915; T. Titaka, Science Rep. Tohoku Imp. Univ., 8. 99, 1919.
1° H. Mache, Sitzber. Akad. Wien, 110. 176, 1901 ; A. Nadiejdine, Journ. Russian Phys.
Chem. Soc, 16. 222, 1884; I. litaka. Science Rep. Tohoku Imp. Univ., 8. 99, 1919.
" R. Bunsen, Pogg. Ann., 141. 1, 1870 ; 31. 1, 1887 ; F. Richarz, Wied. Ann.,^. 708, 1893 ;
67. 704, 1899 ; A. Wigand, Ann. Physik, (4), 22. 64, 1907 ; A. Richter, ib., (4), 42. 779, 1913 ;
A. Bettendorff and A. Wiillner, Pogg. Ann., 133. 293, 1868 ; H. V. Regnault, Ann. Chim. Phys.
(4), 7. 450, 1866 ; (3), 9. 322, 1843.
§ 13. Molecular Heats— Neumann's and Joule's Rules
Es verhalten sich bei chemisch ahnlich zusammengesetzten Stoffen die specifischen
Warmen ungekehrt, wie die stochiometrischen Quantitaten.— F. E. Neumann (1831).
The molecular heat, or thermal capacity of the molecules of a substance, is defined
as the product of its specific heat and its molecular weight. The molecular weight
of but few solids are known, and the simplest formula consistent with the valency
of the component elements is provisionally regarded as representing the molecule.
In the course of his Untersuchungen iiher die specijische Wdrme der Mineralien,^
in 1831, F. E. Neumann noticed that the product of the specific heats and the
molecular weights of compounds of similar composition is nearly constant —
Neumann's rule. This was confirmed by H. V. Regnault in 1841. Two illustra-
tions of Neumann's rule for molecular heats are indicated in Table XXII.
If R represents the symbol of the basic element, with the RO-oxides, the constant
806
INORGANIC AND THEORETICAL CHEMISTRY
is 11-0 ; with the ROg-oxides, 14*0 ; with the ROs-oxides, 18-8 ; with the R2O3-
oxides, 26-9 ; with the RS-sulphides, 11-9 ; and IS'l with the RSa-sulphides ; with
the RCl-chlorides, the constant is 12-75 ; and 18-7 with the RCi2-chlorides. With
the RCOa-carbonates, the constant is 21*4 ; and 29-1 with the R2C03-carbonate«.
With RS04-sulphates the constant is 25-4 ; with the R2S04-sulphates, 32-9 ; with
the RNOg-nitrates, 24-0 ; and with the R(N03)2-nitrates, 38-2.
Table XXII. ^ — ^MoLECUiJi.R Heats of {
Solids.
Carbonates.
Mol. wt.
Sp. heat.
Mol. heat.
Chloride.
Mol. wt.
Sp. heat.
Mol. heat.
CaCOs. .
SrCOa
BaCOa
PbCOa
100-09
147-62
197-37
277-02
0-206
0-145
0109
0-080
20-6
21-3
21-4
21-3
BaCla .
SrCla .
PbCla .
HgOU .
208-29
158-54
277-02
270-92
0-090
0-120
0-066
0-069
18-7
19-0
18-3
19-2
J. p. Joule, in 1844, brought forward some evidence indicating that the mole-
cular heat of a solid compound is approximately the sum 0! the atomic heats of
the constituent elements — Joule's rule. Consequently, if M and Cp respectively
denote the molecular weight and specific heat of the compound ; mi, m^, m^, ■ • .
the atomic weights of the constituent elements ; Ci, C2, C3, . . . the specific heats,
and Til, ^2j ^3 • • • the number of atoms of the respective elements.
Molecular heat, M Cp=:7limiCx+?^2*^2^2"l"%^^3^3 ~h • • •
This rule is sometimes called Woestyn's rule, after A. C. Woestyn (1848), and Kopp's
rule, after H. Kopp (1865), since they each expressed the same idea by saying that
each element has the same atomic heat in compounds as it has in the free state,
so that if Gi, a^, a^, . . . denote the atomic heats of the respective elements.
Molecular h6at=9i2ai-|-W2«2"i"^3^3'i~ • • •
or, if 6*4 be the atomic heat of each element, and the compound contains n atoms,
the molecular heat of the compound will be approximately 6 '4%. Otherwise ex-
pressed, the quotient obtained by dividing the molecular heat of a compound by
the number of elementary atoms in one molecule is approximately equal to 6-4.
This rule was also favoured by J. J. J. Garnier and S. Cannizzaro ; but H. Kopp
showed that it is not universally applicable. Hydrogen, nitrogen, oxygen, fluorine,
and chlorine give discordant results. There is obviously a difficulty with the
carbonates. This may possibly be connected with the difficulty previously found
for carbon. In the case of the lighter elements the atomic heats must be taken less
than 6 '4. Some atomic heats of the elements when in combination have been
previously indicated. Further, if the atomic heats of all but one of the elements
in a compound be known, the unknown atomic heat can be computed ; thus, the
atomic heat of chlorine in lead chloride is J(18-3-6'4)==5'95. A comparison of the
results of experiment with calculations based upon Neumann's and Joule's laws is
indicated in Table XXIII.
Table XXIII. — ^Molecui-ab Heats of Solids.
Compound.
rormula.
Sp. heat.
Mol. weight.
Molecular heat.
Observed.
Calculated.
Mercuric chloride
Mercuric iodide .
Mercurous chloride
Mercurous iodide
HgCl,
Hgia
HgCl
Hgl
0-0689
0-0420
0-0520
0-0385
270-92
253-84
235-46
326-92
18-67
19-06
12-25
12-91
18
18
12
12
THE KINETIC THEORY OF ATOMS AND MOLECULES 807
W. A. Tilden found the specific heats of solids show variations with temperature
analogous to those exhibited by the elements ; but in general, the molecular heat of
a compound or alloy did not difEer greatly from the sum of the atomic heats of the
component at the same temperature. S. Meyer 2 has drawn attention to the fact that
the additive law of mixtures is more nearly followed with molecular heats than it
is with molecular volumes. A. Winkelmann found the additive law applicable to
glasses ; H. V. Regnault, R. Durrer, W. Spring, and L. Schiiz applied it to alloys ; and
P. Bachmetjefi and M. Wascharoff, to amalgams. According to VV. Spring, the additive
rule fails with alloys of tin and lead. M. Trautz found that in the majority of cases
the additive rule with the internal atomic heats, C«— |i2, either holds good exactly,
or very approximately. J. J. J. Garni er confirmed the additive rule for hydrated
salts on the assumption that the combined water has nearly the same thermal
capacity as solid ice, 9*85 cals. per gram-molecule of H2O. There are many
discrepancies, as might be expected, and for the reasons stated in connection with
the atomic heats of the elements.
Examples.' — (1) Calculate the specific heat of solid oxygen given the specific heat of
potassium chlorate, KClOj, 0'194 ; and that of potassium chloride, 0'171. Here the
molecular heat of potassium chlorate (molecular weight x specific heat) is 25*7; and of
potassium chloride, 12 -8. The diSerence 25'7 — 12*8 = 12'9 represents the moleciilar heat
of O3, hence the atomic heat of oxygen will be ^ of 12'9=4*3. By definition, atomic
weight X specific heat of solid = atomic heat of solid = 4'3. Hence, the specific heat of solid
oxygen will be 4-3-^16 =0*27.
(2) The specific heat of silver chloride is 0'0911 and that of silver, 0'057 : assuming the
atomic weight of silver to be 107-9, what is the specific heat of solid chlorine ? The molecular
heat of silver chloride is 13*1, and the atomic heat of silver is 6*2. The difference 13*1
— 6*2 = 6'9 represents the atomic heat of solid chlorine. The specific heat of solid chlorine
is therefore 6*9^ 107-9 =0'064.
The molecular heats can be employed to rectify the atomic weights of elements
which do not form volatile compomids. Thus, the analyses of mercurous and
mercuric salts indicate that the atomic weight of mercury may be 100, 200, . . .
If the atomic weight be 100, the formula of mercurous chloride will be Hg2Cl, and
of mercuric chloride, HgCl ; while if the atomic weight be 200, the formulae will be
those indicated in the above table. There are some discrepancies as is illustrated
by the fact that E. Donath (1788) deduced the value 120 for the atomic weight of
uranium from the specific heat of uranoso-uranic oxide, where C. Zimmermann's
value (1881), from the observed specific heat of the metal, was twice E. Donath's
value — viz. 240.
Examples.— (1) The analysis of barium chloride fiimishes 35*5 parts of chlorine per
68*7 parts of barium. The specific heat of barium is 0-0465. What is the atomic weight of
barium, when the atomic weight of chlorine is 35-5 ? The formula of barium chloride may
be written Ba^Cl, where x is to be determined. The atomic weight of barium, by Dulong
and Petit's rule, will be of the order 6-4-^-0-0465 = 137. Taking 35-5 as the atomic weight
of chlorine, the fraction x must be of the order 68-7 -r- 137 =^. Hence the formula of barium
chloride is BajCl, that is, BaCla, or some multiple of this. Hence the atomic weight of
barium (chlorine, 35-5) must be 2 x 68*7 = 137-4.
(2) The percentage composition of platiniun chloride is : Platinum, 67*7 ; chlorine, 42-3.
The specific heat of platinum is 0-0324. What is the atomic weight of platinum ? Hint,
see (2) in the last but one set of examples. The ratio of the constituent elements is as
48-6 : 35-5 ; the atomic weight is of the order 197-5 ; the ratio x is nearly J ; and hence the
formula of the chloride is PtCl4 or some multiple of this. Hence, assuming the atomic
weight of chlorine is 35'5, the atomic weight of platinum will be 4 X 48-6 = 194-4.
References.
1 F. E. Neumann, Pogg. Ann., 23. 32, 1831 ; H. V. Regnault, Ann. Chim. Phys.,{S), 1. 129,
1841 ; J. P. Joule, Phil. Mag., (3), 25. 334, 1844 ; A. C. Woestyn, Ann. Chim. Phys., (3), 23. 296,
1848; H. Kopp, Liebig's Ann. SuppL, 3. 1, 290, 307, 1865; C. Pape, Pogg. Ann., 120. 337, 679,
1863 ,- 122. 408, 1864 ; 123. 277, 1864; A. Sella, Gott. Nachr., 311, 1891 ; S. Canizzaro, i\^MOw
Cimen^o, (I). 7. 321, 1858.
808 INORGANIC AND THEORETICAL CHEMISTRY
* S. Meyer, Sitzher. Akad. Wien, 109. 405, 1901 ; Ann. Phyaik, (4), 2. 135, 1900 ; E. van
Aubel, »6., (4), 4. 420, 1901 ; A. Winkelmann, Wied. Ann., 49. 401, 1903 ; J. J. J. Gamier, Comjpt.
Rend., 35. 278, 1852 ; H. V. Regnault, Ann. Chim. Phys., (3), 1. 129, 184J ; L. Schiiz, Wied. Ann.,
46. 177, 1892 ; W. Spring, Bull. Acad. Belgiqve, (3), 11. 355, 188G ; P. Bachmetjeff and M. Waa-
charoflF, Journ. Russian Phys. Ghem. Soc, 25. 115, 1893 ; R. Durrer, Phys. Zeit., 19. 8(i, 1918 ;
M. Trautz, Zeit. anorg. Chem., 95. 79, 1916 ; Zeit. Elektrochem., 95. 79, 1910 ; W. A. Tilden,
Phil. Trans., 203. A, 139, 1904.
§ 14. The Meaning of Dulong and Petit's Rule
The fact that the atomic heats of all elements are approximately the same,
led Dulong and Petit to infer that the thermal capacity of all atoms is the same.
This means that every atom of a solid — no matter of what kind — requires the same
amount of heat to raise its temperature 1°. J. P. Joule's rule means that each
elementary atom retains the same capacity for heat when it is combined as it had
when free. The number and kind of other atoms present and their mode of
combination seem to have no influence on the numerical value of this property.
The observations of Neumann and Joule indicate that the constituent atoms of a
solid compound behave as if the soUd were a mechanical mixture of its component
atoms, and each atom were free to vibrate independently of the others. i
According to the kinetic theory, temperature is proportional to the kinetic
energy of the molecules ; and consequently, Dulong and Petit's rule points to a
similar relation. It must be added that we can form no real conception of the
" temperature of an atom "or of the " temperature of a molecule." All our con-
ceptions of temperature are based on the properties of atoms and of molecules en
masse. R. Clausius 2 supposes the specific heat of an element to be made up of two
magnitudes (i) the heat c^ required to raise the kinetic energy of the molecules ;
and (ii) the heat € required to perform internal work. Clausius calls c^ the true
specific heat of the solid. Hence if M be the atomic weight of an element, the
observed atomic heat MC is equal to M(Cv-\-€). It is often stated that at the
absolute zero of temperature, —273°, all atomic motion must cease. This is a mere
assertion, of no intrinsic value, and probably wrong. The statement migU be
true of the translatory motion of the molecules — such that R. Clausius' Mc^ is zero
— because of the convergence of the specific heats of the elements to zero as the
temperature approaches absolute zero. The same fact also shows that the internal
work € becomes very small at absolute zero, and the fact that Dulong and Petit's
rule is so nearly exact at ordinary temperatures coupled with the assumption that
at the same temperature the kinetic energy of the molecules is the same, leads to
the inference that when the temperature of elementary solids is raised from absolute
zero, the internal work per atom is approximately the same.
W. Jankowsky ^ argued that the heat of a chemical reaction is developed for the most
part by the conversion of potential energy into heat, and that at the absolute zero, the
energy content of a substance is entirely potential, and that there must be an absolute
upper limit or maximum temperature where the energy content of a substance consists
entirely of heat.
Among the evidence which indicates that] the atoms of a solid^even at absolute
zero, probably oscillate about a position of equilibrium, the following may be cited :
(i) The low coefficient of thermal expansion of solids shows that the volume would
be very little changed if the solids were cooled to absolute zero ; (ii) it is not pro-
bable that solids would lose their compressibility at absolute zero ; (iii) the natural
frequency of the vibrations of the atoms of a solid calculated by different methods
shows no signs of ceasing at absolute zero, (iv) P. Debye's effect^ in which the
intensity of the higher orders of the X-ray spectrum of crystalline solids increases
as the temperature of the crystal is lowered, points in the same direction.
L. Boltzmann, in a paper Ueher die Natur der Gasmolekule (1876), has shown that
THE KINETIC THEORY OF ATOMS AND MOLECULES 809
the kinetic and potential energies of the molecules of a monatomic solid vibrating
about a position of equiUbrium are equal in magnitude. Consequently, the total
energy of a vibrating system — called for convenience an oscillator — is shared equally
between the average kinetic and potential energies, and is twice the value of either
alone>
This interesting result follows by considering the motion of a particle under the influence
of a central attractive force moving on an orbit about its position of equilibrium. If the
particle were at rest in any part of its orbit, it would tend to move to its centre of attraction,
and in so doing, would acquire such a velocity that its kinetic energy would be the same as it
possessed when oscillating in its former orbit. Hence, a particle oscillating about a centre
of rest possesses both kinetic and potential energy, and on the average, the one is equal to
the other, provided that the time average of its kinetic energy is equal to that of the
potential energy. This is the case if the potential energy is zero as the particle passes through
its position of equilibrium. In reality, the equipartition theorem applies only to the kinetic
energy, but if the average kinetic and potential energies are equal, each will make the same
contribution to the specific heat.
L. Boltzmann assumed that the atoms of a solid have natural periods of vibration,
so that if a monatomic gas be in contact with a solid, the bombardment of the
gaseous molecules produces a state of thermal equilibrium when the mean kinetic
energy of vibratory motion of the atoms of the solid is equal to the mean kinetic
energy of the translatory motions of the molecules of the gas. With a solid, the
average kinetic energy of the atoms in each state will be the same ; but the average
kinetic energy per atom of a monatomic gas is ^KT per atom, hence, the sum of the
kinetic and potential energy of the solid will be 2x^KT, or ^KT per atom ; and if
there are N atoms in a gram-atom of the solid, the total kinetic and potential energy
will be 3NKT or 3RT per gram-atom, where NK—R is nearly equivalent to two
calories per gram-atom per degree. Accordingly, the atomic heat, wCv, must be
3i2=6 nearly, or with a more exact value of R, 5*95. This interesting argument
shows how the atomic heats of the monatomic solids are nearly twice the molecular
heats of the monatomic gases ; and it furnishes a brilliant deduction of Dulong and
Petit's rule for solids.
The agreement between the result of Boltzmann's assumption and Dulong and
Petit's observation, shows that the atoms of a monatomic solid probably vibrate
so that their energy is equally divided between the kinetic and potential energy.
If the oscillations of the atoms are not harmonic in character, the time averages of
the kinetic and potential energies will not generally be equal. The agreement in
question also shows that the opposing forces — attraction and repulsion — between
the atoms just balance one another so that as two atoms approach one another the
attractive forces gradually diminish, and the repulsive forces gradually increase
until the latter predominate.
The discrepancies between Boltzmann's ZR and Dulong and Petit's constant.
— There must be a flaw somewhere, because the theory does not explain (i) how the
solid elements with a low atomic heat — carbon, silicon, and boron — have normal
atomic heats at high enough temperatures ; nor (ii) how all solids give abnormally
low values at low temperatures. Many attempts have been made to explain the
discrepancy between theory and fact. It may be necessary to consider :
{a) The time required for the atoms to adjust themselves to a change of temperature.
— L. Boltzmann assumed that the atoms take a long time to adjust themselves
to the temperature — but no corresponding variation of specific heat with tempera-
ture has been detected ; and the specific heats of solids are so related with the
melting points that if the specific heat changed with time, the melting point
ought likewise to change. Such a phenomenon has not been observed even in
the case of artificial minerals and natural minerals formed aeons ago. While the
translational energy may be rapidly distributed between the internal motions of
a molecule during a collision, yet, if the distribution is slow, so that it becomes
appreciable only after millions of collisions, the number of collisions per second
810 INORGANIC AND THEORETICAL CHEMISTRY
is so great — a million occurs in about one-seven-thousandth of a second according
to G. J. Stoney — that even when the exchange is slow, a second of time is a com-
paratively long interval.
{b) The oscillations of the atoms are not harmonic. — L. Boltzmann assumes that the
vibrations of the atoms is harmonic ; and this assumption is probably valid for most
metals far from their melting points ; but if the amplitudes of the vibrations of the
atoms are large, oscillations may be no longer harmonic. I. Langmuir 5 has
emphasized the fact that if the oscillations of the atoms are not harmonic, the time
averages of the kinetic and potential energies will not be equal. The average
kinetic and potential energies will be equal, only when the motion is harmonic,
in which case, the restoring force acting on the atom is proportional to the dis-
placement from the position of equilibrium. If the restoring force increases more
slowly than the displacement, the potential energy will be greater than the kinetic,
and from the principle of equipartition, the atomic heat will be greater than 3R ;
conversely, if the restoring force increases more rapidly than the displacement, the
atomic heat will be less than SR. The remarkable closeness of the atomic heats of
the elements to the value 3R, is taken to show that the forces to which the atoms of
a solid are subjected vary approximately with the displacement of the atoms from
their position of equilibrium.
There must then be both attractive and repulsive forces acting between the atoms. On
the average, these opposing forces must just balance each other. As one atom approaches
another the repulsive force must gradually increase and the attractive force decrease until
the repulsive force greatly predominates. We cannot consider that the repulsive forces
in solids are exerted only during collisions between atoms, for under these conditions there
would be no potential energy and the atomic heat would be |i?.
(c) The congealing of molecules to more rigid systems.— 'R. A. Millikan (1912) ^
considers that it may possibly be assumed that as the temperature is reduced, the
atoms of the solid are frozen, so to speak, into rigid systems of continually increas-
ing size — where each system is endowed with the kinetic energy of agitation appro-
priate to its temperature — before absolute zero is attained, it might be possible for
the total energy of the whole mass to become that of a single molecule of the sur-
rounding gas. C. Benedicks (1913) has also shown that the equipartition law is
avoided by assuming that the solids are not always monatomic, but at low tempera-
ture form atomic complexes, which change the number of degrees of freedom. The
equipartition law applies only to free atoms. However, from Joule's law, it appears
probable that the rule for atomic heats applies to atoms in combination as well as
free.
(d) Another explanation of the reduction in the atomic heats below 3R when the
temperature is low was suggested by A. Einstein (1907) . It is based on the so-called
quantum theory of energy ; and has been remarkably successful.
References.
1 W. Sutherland, Phil. Mag., (5), 32. 550, 1891.
2 R. Clausius, Pogg. Ann., 116. 100, 1862 ; W. Sutherland, Phil Mag., (5), 32. 550, 1891 ;
W. Jankowsky, Zeit. Elektrochem., 23. 368, 1917.
' W. Jankowsky, Zeit. Elektrochem., 25. 325, 1919.
* L. Boltzmann, Sitzber. Mad. Wien, 74. 555, 1876 ; H. Petrini, Zeit. phys. Chem., 16. 97,
1895 ; E. J. Routh, The Dynamics of a System of Rigid Bodies, London, 2. 54, 1892.
^ I. Langmuir, Journ. Amer. Chem. Soc, 38. 2236, 1916.
« R. A. Millikan, Science, 37. 119, 1913 ; C. Benedicks, Ann. Physih, (4), 42, 1333, 1913 ;
F. Richarz, Zeit. anorg. Chem., 58. 356, 1908 ; 59. 146, 1908 ; J. J. van Laar, Proc. Acad.
Amsterdam, 11. 765, 1909 ; 12. 120, 133, 1909 ; 13. 454, 636, 1910 ; J. Duclaux, Compt. Rend.,
155. 1015, 1912; A. H. Compton, Phys. Rev. (2), 6 377, 1915; F. Schwers, ih., (2), 8. 117, 1916.
THE KINETIC THEORY OF ATOMS AND MOLECULES 811
§ 15. The Quantum Theory of Energy and Dulong and Petit's Rule
An observer who does not allow himself to be led in his work by any hypothesis, how-
ever cautious and provisional, renounces beforehand all deeper understanding of his own
results.— M. Planck (1914).
An attempt to imagine a universe in which action is atomic leaves the mind in a hopeless
state of confusion. — J. H. Jeans (1914).
J. H. Jeans, in his The Dynamical Theory of Gases (Cambridge, 1904) , shows that
Maxwell-Boltzmann's theorem of the equipartition of energy is based upon the
assumption that there is no interaction between matter and aether, whereas every
ray of light which reaches the eye is evidence against the truth of the assumption.
With ordinary diatomic transparent gases two (rotational) degrees of freedom
appear to be directly affected by the translational motions during a collision ; with
the coloured gases there appear to be motions which consume energy in the pro-
duction of sethereal vibrations. In 1906, M. Planck, in his Vorlesungen vher Theorie
der Wiirmestrahlung (Leipzig, 1906), assumed that the interchange of energy between
the sether and a vibrating atom is not a continuous process, but takes place fer
saltum — that is, discontinuously — by jumps in definite amounts hv, where v represents
t he Schwingungszahl or the frequency of the atomic vibrations, and y^ is a universal
constant in the same sense that e, the unit of electrical change, is a universal constant.
The constant h — called Planck's constant — seems to be a fundamental imit which
regulates and controls the ceaseless ebb and flow of energy in the world of matter.
For brevity, write €=hv. This means that for any given temperature, a certain
amount of energy is associated with the vibrating atom, and that this amount is
a function of the vibration-frequency v of the atom ; and energy can be absorbed
or emitted by a vibrating system spasmodically, and only in amounts e or in integral
multiples of this magnitude such as €,2e, 3€, . . ., but not in intermediate quan-
tities, say, Je, ^e, fc, . . . This virtually means, said H. Poincare (1911) ,i that a
physical system can exist only in a finite number of states, it leaps from one of these
states to another without passing through a continuous series of intermediate states ;
and, adds M. Planck : 2
The continuity of all dynamical effects was formerly taken for granted as the basis
of all physical theories and in close correspondence with Aristotle, was condensed in the
well-known dogma—natura nonfacit saltus — nature makes no leaps. However, present-day
investigation has made a considerable breach even in this venerable stronghold of physical
science. This time it is the principle of thermodynamics with which that theorem has
been brought into collision by new facts, and iinless all signs are misleading, the days of
its validity are numbered. Nature does indeed seem to make jumps- — and very extra-
ordinary ones.
The ration or unit of energy e is called a quantum, and hence this hypothesis is
called the quantum theory of energy. According to this remarkable hypothesis,
the vibrating atoms radiate definite loads hv of energy which, for any given
vibration frequency, v, are indivisible. M. Planck inferred that the average energy
€ possessed by an oscillating, unit, with two degrees of freedom,
hv hv
^
Average energy = , ; or, Average energy = . . (12)
— e A
per degree of freedom ; and three times this value for a monatomic oscillator with
three degrees of freedom instead of the average value 3^:1" per atom deduced by an
application of Maxwell-Boltzmann's theorem which assumes that the evolution
or absorption of energy is a continuous process. Here u is written in place of the
fraction hvjkT.
M. Planck follows the theory of probability in deducing his formula ; D. L. Chapman
starts from J. H. van't Hoff's well-known expression Q/RT^=^{d log k)/dT. From the
quantum law, if the resonators of vibration period v are attached to the molecules of a gas.
812
INORGANIC AND THEORETICAL CHEMISTRY
then there will be vibrators possessing amounts of energy 0, hvy 2kv, . . . , but no
vibrators with intermediate amounts of energy. Let the number of vibrators with 0, hv,
2hv, . . . amounts of energy be respectively Wq, n^, ^2, • • • '^m' "^^^^j from J. H. van't
Hoff's rule, 7nhvdT/kT^ = d log (w^^/Wq), and by integration between T and oo, and writing
u in place of hv/kT, it follows that M^ = «Qe "''*". Consequently, the mean energy of a
vibrator is (/i»'e-" + 2;ive-2w + 3/ii/e-3w+ . . .)'(i +e-" + e-2w+ . . .), which, by division,
reduces to (30) above. It might also be added that F. R. von Bichowsky ^ has shown that
(i) Planck's radiation law, (ii) the quantum theory, and (iii) the equipartition law are not
independent, because if any two be assumed the third will follow ; and further, S. Ratnowsky
has shown that if J. W. Gibb's assimaption (that the free energy of a system cannot be
generated until the magnitude of the co-ordinates fixing the energy of the system has
reached a certain value, and is thereafter given off continuously) be made, Planck's
radiation law follows directly without the assumption of the quantiun hypothesis. Other
attempts to establish a theory of radiation without quanta have been made by M. Brillouin,
A. Byk, H. L. Callendar, and R. C. Tolman.
In one modification of the hypothesis, the oscillator is supposed to absorb energy
continuously until an amount hv has been absorbed, when it has a chance of emitting
the whole of this unit. Otherwise, energy will continue being absorbed until it
reaches 2hv, Shv, . . . Only when the amount of energy reaches an exact multiple
of hv is the oscillator in a condition to emit the whole of its energy. *
It is an open question what are the receptacles of energy in a solid. As H. A. Lorentz
(1913) 5 has shown, the phenomena of light niake it highly probable that energy quanta
can have no individual and permanent existence in the
aether, and cannot be regarded as accumulations of energy
in minute spaces travelling about with the velocity of
light. It seems more probable that the energy of solids
is localized in the eltistic vibrations of the solid, and that
the mean energy of an oscillator is equal to the mean
energy of an aether vibration of the same frequency. No
reason can be assigned why the electric charge e always
acts as if it were atomic, or why electrons, each with a
fractional charge— say, ^e- — -do not exist ; so also no
reason can be assigned why energy can change only by
complete quanta.
1000
200 400 600 800 1000 K
Fia. 9. — Values of the Function
^v/(e" — 1) at Different Tem-
peratures.
tion to the radiation of heat.
In 1907, A. Einstein,^ in his paper Die Plancksche
Theorie der Strahlung und die Theorie der spezijischen
Wdrme, extended Planck's atomic theory of radia-
He assumed that the longer heat waves emitted and
absorbed by solids are due to vibrations of the constituent atoms about a mean posi-
tion of rest. A. Einstein further assumed that the energy of the solid does not reside
solely in the kinetic energy of the atoms, but the vibration-frequency v of each atom
has three degrees of freedom, and the energy of these vibrations is governed by
M. Planck's law, and A. Einstein thus deduced a formula analogous with that of
M.Planck;
Bv
Average energy =3R ^ . . • • (13)
e JL
for the energy of the vibrating atoms of a solid. If j3v is very small, the function is
approximately 2>RT, and the expression corresponds with Dulong and Petit' s law,
which requires the atomic heat of monatomic solids to be proportional to the
temperature. For all other values of j3v the function is less than 2tRT. The values
of the function are plotted in Fig. 9. At any given temperature, the value of
the function differs more and more from the value of T as the value of ^v is
increased. Differentiating for dEldT, the atomic heat, d,, he obtained :
Atomic heat =3R
2«M
u^e
(14)
(e«-l)2
where u, for convenience, has been written in place of ^v/T, and ^ is written in
THE KINETIC THEORY OF ATOMS AND MOLECULES 813
place of hjk ; h is the atomic gas constant represented by i?/iV, when It is the
ordinary gas constant, and N (approximately 6-06x1023) denotes the number of
atoms per gram-atom. It will be evident that when the fraction u=^vlT is very
large, v will either be very large, or T very small, and Cv will be virtually zero ;
and when u is small, Cv=^?tR. For example, if j8v/T be greater than 10, Cv=^R
XO-004 ; and if it be less than unity, Cv will be less than 3i? XO'92. In the former
case, the specific heat approaches unity, and in the latter case, Cv is nearly normal.
The numerical values of the constants h and ^.— Seven different lines of argument '
show that Planck's constant /i, is equivalent to (6'5543±0-0026) X 10"^' ergs per second, and is
the same for all substances. For the yellow D-sodium line with a wave-length 0*5896jLt, it
follows that V is 3 X lOio/O'SSOe x 10"* or 5-088 x lO^S so that hv for this radiation is
6-62 X 10-" X 5-088 xlOi* = 33'7x 10-13 ergs per second. The numerical value of j3 is
4-865x10-".
W. Nernst and F. A. Lindemann (1911) ^ have shown that Einstein's equation
is in fair agreement with their observations of specific heats at low temperatures,
although discrepancies appear as the temperatures approach absolute zero ; and
they tried to rectify Einstein's equation by introducing a new term. So that the
atomic heat C^ becomes
3„ w2ew Jt^2el«
Atomic heat =~R -, — -„ + , ,
2 (e«— l)2^(ei«_i)2
(15)
on the assumption that the solid is a mixture of oscillating atoms half of which
have the vibration frequency v and half the frequency Jv. M. Planck and
A. Einstein assumed that all the oscillating atoms had a frequency v ; and W. Nernst
and F. A. Lindemann's assumption is a first approximation to a summation ex-
tending over an infinite number of values of v. W. Nernst and F. A. Lindemann's
equation represents the observed atomic heats of solids — aluminium, copper, silver,
lead, mercury, zinc, iodine, and the diamond — down to the lowest temperatures.
A few numbers selected from Nernst's tables for silver and the diamond are indicated
in Table XXIV.
Table XXIV.- — The Atomic Heats of Silver and the Diamond at Different
Temperatures.
Silver.
Diamond.
/3w=221.
/Si/ -1940.
Temperatures.
Cp (calc).
Cp (obs.).
Temperatures.
Cp (calc).
Cp (obs.).
350
1-59
1-58
30
000
0-00
63-8
2-98
2-90
92
0-01
003
100
4-77
4-86
205
0-62
0-62
200
5-77
5-78
243
0-97
0-95
273
6-02
6-00
306
1-69
1-68
331
6-12
6-01
358
2-08
2-12
535
6-45
6-46
413
2-55
2-66
589
6-57
6-64
1169
5-41
5-45
According to the form of quantum hypothesis now under consideration, oscillat-
ing atoms cannot absorb energy unless it comes to them with a certain degree of
intensity equal to hv, or some whole multiple thereof. As the temperature rises,
the number of molecules which take up loads of energy from the low intensity heat
waves increases rapidly in accord with the equipartition law, and the need for ab-
sorbing energy in integral multiples of hv. Molecules of chlorine and bromine begin
to absorb this energy at a lower temperature than the transparent diatomic gases
because (i) the bond of union between the respective atoms is weak, and their
814
INORGANIC AND THEORETICAL CHEMISTRY
frequency v — and consequently also their quantum hv — is small ; hence (ii) the
quanta or loads of energy hv absorbed by these oscillators are correspondingly small ;
and (iii) the temperature at which the kinetic energy of the diatomic oscillators
attain the value hv is correspondingly low. Diatomic hydrogen molecules at a low
temperature act like monatomic molecules because the rotatory motions at any
given temperature correspond with a definite frequency v, and when the energy
of impact falls below this value of hvj no energy can go into these rotations, and
energy is solely distributed among the three degrees of freedom corresponding with
translatory motion.
The quantum hypothesis gives a qualitative explanation : (i) how the atomic
beats of the elements approach zero as the temperature falls, and (ii) how abnormally
low values appear at a higher temperature with the elements of low atomic
weight. From measurements of atomic heats, it seems as if, as the temperature
rises, different kinds of atoms can take on their normal load at different stages,
the heaviest atoms take it on first, the lighter atoms last. With a given rise of
temperature there is a corresponding increase in the vibratory energy of the atoms
of an element, and at a sufficiently low temeprature, only a definite fraction of the
atoms can take on the normal quota hv. The higher the vibration frequency v,
the higher the temperature at which energy can be absorbed. Again, other things
being equal, with a falling temperature, the greater the vibration frequency v, the
sooner will atomic heats lower than 3i2=6 calories begin to appear.
Observations do not agree with the assumption that at absolute zero the atoms
of hydrogen have no latent energy. Consequently, A. Einstein and 0. Stern (1913) ^
have examined the hypothesis that the rotating molecules have ^hv units of energy
at the absolute zero, and instead of M. Planck's expression (30), they write :
Average energy =
hv
1
2
. (16)
The photoelectric effect and the emission of electrons by the action of the X-rays
are also in accord with the assumption that the latent energy of the electrons on a
metal is ^hv per degree of freedom. The average energy
plotted according to the equipartition law, where the average
energy =kT per degree of freedom, is shown by the curve
I, Fig. 10; according to M. Planck's formula (30), by II,
Fig. 10, and according to A. Einstein and 0. Stern's
formula (35), by III, Fig. 10, The formula for the specific
heats of gases at different temperatures derived by differentiat-
^ . . ■ ■ . ■ ■ ing the above formula, agrees well with A. Eucken's observa-
empe/a ure. . ^JQ^g j^ might be added, however, that P. Ehrenfest (1913)
Fig. 10. T e iifiec of obtained almost as good an agreement without their assump-
Energy of Solids. ^Aon. W. H. Keesom (1913) applied an argument simUar to
that used by P. Debye, and deduced a formula for the
specific heats of gases similar to that obtained for solids. It was found generally
that while the rate at which the specific heat (or the energy per degree) decreases
to zero as the temperature is lowered, the total energy does not necessarily become
zero at absolute zero — rather does there exist at this low temperature a latent
energy whose magnitude is JAv.
References.
1 H. Poincare, Dernikrespensies, Paris, 1913 ; Jmirn. Phys., (6), 2. 5, 1912 ; J. H. Jeans, Ann.
Physik, (4), 17. 132, 1905; (4), 20. 197, 1906; (4), 22. 180, 1907.
2 M. Planck, Phil. Mag., (6), 28. CO, 1914 ; 1). L. Chapman, Annnnl Jiejiorts of the Progress of
Chemistry^ London, 11. 3, 1914 ; Lord Rayleigh, Proc. Rm/. Soc, 83. A, 92, 1909.
3 F. R. von Bichowsky, Phys. Rev., (2), 11. 68, 1918; R C. Tolman, ih., (2), 3. 244, 1914;
S. Ratnowsky, Ber. deut. phys. Ges., 16. 232, 1916; 15. 64, 1915; A. Byk, Ann. Physik, (4),
THE KINETIC THEORY OF ATOMS AND MOLECULES 815
42. 1417, 1913; H. L. Callendar, Phil. Mag., (6), 26. 787, 1913; (6), 27. 870, 1914; M. Brillouin,
Ann. Phys., (1), 1. 13, 163, 433, 1914. "^
* R. A. MiUikan, Science, 37. 199, 1913.
^ H. A. Lorentz, Die Theorie der Strahlung und der QtiarUen, HaUe, 10, 1914.
« A. Einstein, Ann. Physik, (4), 22. 180, 1907.
' R. A. Millikan, Phys. Bev., (2), 2. 142, 1913; (2), 7. 365, 1916; R. T. Birge, ib., (2), 14.
361, 1919.
8 W. NerastandF. A. Lindem&nn, 8 itzher. Akad. Wien, 120.347,1911; Zeit. Elektrochem.,!!.
817, 1911.
• W. H. Keesom, Proc. Akad. Amsterdam, 12. 98, 1913 ; Suppl. Comm. Phys. Lab. Leiden,
30a, 1913 ; P. Ehrenfest and H. K. Onnes, ib., 37, 1914 ; Proc. Akad. Amsterdam, 13. 789,
1914 ; A Einstein and 0. Stern, Ann. Physik, (4), 40. 551, 1913.
§ 16. Debye's Theory of Atomic or Specific He^ts
The final object of mathematical research is a knowledge of the principles of science.
— K. Weierstrass.
In the solution of mathematical problems, the object of which is to represent the progress
of nature, we are led by very rapid methods to results which are often overlooked, and which
now and then excite our surprise by the paradoxical form in which they are presented ;
but when gxiided by simple reasoning, we return step by step over the course which was so
quickly bridged by calculation, we end by perceiving the action of the principles which
have given birth to these results.- — R. J. Hauy (1822).
In a paper, Zur Theorie der spezijischen Wdrmen, P. Debye (1912) i argued that
the whole heat energy of a solid resides in the energy of the vibrations of the con-
stituent atoms, and that each vibration has exactly the energy allotted to it by the
quantum theory, and therefore the heat vibrations are the same as light vibrations
of identical frequency. He supposes that it is not likely there is only one value
for the vibration frequency of an oscillating atom, and that it is more probable that a
whole series of values of v exists. It is, however, necessary to postulate an upper
limit to the range of vibration frequencies, and P. Debye assumes that the number
of frequencies for N atoms per unit volume cannot exceed 3iV. At the higher tem-
peratures, where each frequency has the same average energy kT — corresponding
with both kinetic and potential energy — the total energy will be SNkT corresponding
with Dulong and Petit's law. As the temperature is lowered, the average energy
with the higher frequencies is less than with the lower frequencies, so that, when the
temperature is low enough, only those atoms having vibrations of very low frequency,
with hv smaU, can obtain enough energy to vibrate. These low frequencies must
correspond with ordinary sound vibrations, and accordingly, P. Debye identifies
the thermal oscillations of the atoms with the elastic vibrations of the solid, and he
shows that the maximum frequency v for monatomic solids can be calculated from
the elastic constants of the material. P. Debye utilized the quantum hypothesis
and deduced an expression for the average energy of the individual frequencies :
Atomic energy =9it I J-\ — )
x/e"*—!
By differentiating this expression for the atomic heat, C^ at constant volume, he
obtained the
X
3f
Atomic heat=3i? 121 ) / r [ . . (17)
\ eT—V
where u is put in place of hvjkT, and x in place of pv=hvlk, so that xjT^u. P. Debye
called X the characteristic temperature of the particular solid, or it can be called
Debye's constant. Debye shows how the numerical value of x can be calculated
816
INORGANIC AND THEORETICAL CHEMISTRY
from the elastic constants of the solid. This equation is taken to mean that the
atomic heat of a monatomic solid is a function of the ratio xlT, where a; is a
characteristic temperature for each substance, and is dependent upon its density
and elastic constants. The above expression cannot be integrated, but P. Debye
has shown that it can be reduced to W. Nernst and F. A. Lindemann's formula by
a series of approximations. At very high temperatures, when the value of x becomes
very small, and the value of T/x is large, the value of the expression in brackets
approximates to unity ; and the atomic heat reduces to 3R. P. Debye's equation
is more complicated than the empirical relation of W. Nernst and F. A. Lindemann,
but the latter is a good approximation to P. Debye's at low temperatures. For
low values of T/x, P. Debye's constant x is large, and the atomic heat Cv then
approximates to
Atomic heat=322x77-94f-) .... (18)
meaning that at sufficiently low temperatures, the atomic heat varies as the third
power of the absolute temperature ; or by integration, the total energy of mon-
atomic solids near absolute zero is proportional to the fourth power of the
absolute temperature. This agrees with the radiation law deduced by J. Stefan
in 1879, and somewhat later by L. Boltzmann. These investigators showed that
the total energy radiated by a black body is 5'7 XlO~^(T^—To^) ergs persq. cm. per
second, when T represents the absolute temperature of the radiator and Tq the
absolute temperature of the body receiving the radiation.
Working from A. Einstein's (1911) relation between the elastic constants of a
solid and the vibration-frequency of its atoms, P. Debye (1913) has shown that the
characteristic temperature x can be calculated from the formula
a;=35-74 X IQ-^w-iD-iK-iF-i
(19)
where k is the compressibility coefficient ; D, the density ; w, the atomic weight ;
and i'' is a function of the coefficient of linear expansion a such that
/2(l+a)y /_14- \
^3(1 -2a)
3(1
(20)
Some data given by P. Debye are indicated in Table XXV. E. H. and
E. Griffiths 2 found P. Debye's formula to be more accurate than any other existing
Table XXV. — Debye's Constants.
m
D
kxIO'^
F
X
Aluminium
271
2-71
1-36
10-2
399
Copper
63-6
8-96
0-74
10-5
329
Silver
107-9
10-53
0-92
15-4
212
Gold .
197-2
19-21
0-60
24-7
166
Nickel
58-7
8-81
0-57
7-38
435
Iron
56-9
7-85
0-62
5-86
467
Cadmium
112-4
8-63
2-4
7-89
168
Tin .
119-0
7-28
1-9
8-50
185
Lead
207-1
11-32
2 0
610
172
Bismuth
. 208-0
9-78
3-2
8-98
111
Palladium
106-7
11-96
0-57
18-8
204
Platinum
195-0
21-39
0-40
171
226
expression for reproducing their observations of the specific heats between —165°
and atm. temperatures, although here, systematic divergences occur at the higher
temperatures. In general, observations of the specific heats of various metals at
THE KINETIC THEOKY OF ATOMS AND MOLECULES
817
low temperatures agree remarkably well with the values computed from Debye's
formula, even though the elastic constants for evaluating j3v have been determined
at room temperatures. According to W. H. Keesom and H. K. Onnes (1914), the
specific heat of copper is 0-0396 at —258-49° ; 0-1155 at -252-89° ; 0-2340 at
—246-63° ; and 08700 at —232-78°, and therefore decreases more rapidly on a
falling temperature than it should do according to P. Debye's third-power law.
The curve for aluminium calculated from P. Debye's formula is indicated in the
diagram, Fig. 11, where the circles represent the observed specific heats. A. Eucken
and F. Schwers (1913) found that Debye's third-power law holds very well for the
minerals fluorspar and pyrite from about —260° to about 187°. The application
of the quantum theory to the explanation of low temperature specific heats by
A. Einstein and by W. Nernst and F. A. Lindemann can be regarded as preparatory
to that of P. Debye. H. von Jiiptner and E. Rasch 3 have suggested simpler
formulae for the change in the specific heats of solids with temperature which gave
very good results, but they have no known theoretical foundation. J. H. Jeans
(1914) considers that both from its complete naturalness and from its agreement
with experiment, P. Debye's treatment of the specific heats of solids seems destined
to be final.
Various attempts have been made to improve Debye's theory. For example,
M. Born and T. von Karman (1912), and H. Thirring (1913),* assume that the
atoms are arranged in the solid like a space-lattice. There are probably motions
of the atoms or molecules other than
vibratory and translatory movements
which contribute something to the
specific heat. For instance, (i) the sub-
atomic electrons may have their energy
increased when a gas is raised to a high
temperature, but, if so, the increment is
too small to have any appreciable effect
on the specific heat of a monatomic gas.
A. Eucken (1914) ^ found that the value
of Cv for helium is virtually constant
between temperatures ranging from
—256° to 2350°. Again, (ii) the mole-
cules may have rotational movements.
The rotation of atoms is illustrated by the movement of the optical axes of a
crystal as a whole when the crystal is rotated in the hand. The agreement
between theory and observation when the electronic and rotary movements are
neglected shows that they are too small to have an appreciable influence on the
specific heat — unless perchance the specific heats are determined near the melting
points of the solids. J. H. Jeans (1914) considers that this is due to the circum-
stances that the forces opposing the rotational movements of the atoms inside the
solid are so large that the corresponding vibrations are of high frequency, and so
normally possess but little energy. A. E. Oxley (1914) observed that the specific
heats of sodium and mercury near the melting points are in excess of the theoretical
values ; and E. Griineisen (1913) argues that the forces which prevent the atoms
rotating are relaxed, and then an additional term — similar to Einstein's (35) —
should be added to the theoretical specific heat formula to allow for vibrations
which depend upon the rotations of the atoms.
3-^
0"
Fia.
•5
^'
o
/
1 f
^ 7
y
Abso/ut
e Tempemt
ures
100=
200=
300=
400°
11. — The Effect of Temperature on the
Atomic Heat of Aluminium.
References.
1 P. Debye, Ann. Physik. (4), 39. 752, 1913; E. H. and E. Griffiths, Phil. Trans., 214. A,
319, 1914.
2 E. H. and E. Griffiths, Phil. Trans., 214. A, 319, 1914; A. Eucken and F. Schwers, Ber.
deut. phys. Ges., 15. 578, 1913; W. H. Keesom and H. K. Onnes, Comin. Phys. Lab. Leiden, 143,
1914; 147, 1915.
VOL. I. 3 G
818 INOKGANIC AND THEORETICAL CHEMISTRY
» H. von Juptner, Zeit. Elektrochem,, 19. 71 J, 1913; 20. 10, 105, 1914; E. Rasch, Mitt.
Konig. Materialpruf., 320, 1914.
* M. Born and T. von Karman, Phys. Zeit., 13. 297, 1912 ; 14. 15, 65, 1913 ; H. Thirring,
ib., 14. 807, 1913 ; 15. 127, 180, 1914.
* A. Eucken, Sitzber. Akad. Berlin, 123. 682, 1914 ; J. H. Jeans, Report on Radiation and the
Quantum Theory, London, 1914 ; A E. Oxley, Proc. Cambridge Phil. Soc, 17. 450, 1914 ;
E. Griineisen, Molekulartheorie Jester Korper, Bruxelles, 1913.
§ 17. The Kinetic Theory o! SoUds
We must infer that constituent parts of all bodies are in perpetual motion. — R. Watson
(1789).
Nature gives no evidence of absolute rest. All matter, so far as we can ascertain, is
ever in motion, not merely in masses, as with planetary spheres, but also molecularly
throughout its most intimate structure. — W. R. Grove.
Many phenomena commonly associated with liquids and gases — vaporization,
crystallization, dissolution, difiusion, chemical action, etc. — are also manifested by
solids under the right conditions of temperature and pressure. The solid state is
not a condition of molecular inactivity and rest. The reason why the phenomena
which indicate molecular activity in solids are so often overlooked, is due to the fact
that these changes are usually very slow. In solids, the translatory motion of the
particles must be very greatly hampered by adjacent molecules ; and, except possibly
in the case of amorphous solids, it is highly probable that the forces of cohesion
cause the molecules in the solid to oscillate about a fixed position of equilibrium so
that their movements are restricted ; and it is doubtful if the molecules change
their locality in the same sense that the molecules of liquid and gases are con-
tinually moving from one part of the mass to another.
When a liquid is cooled below its freezing temperature, a certain amount of heat
is evolved — latent heat of fusion — as the liquid solidifies. The solidification of a
liquid must therefore be attended by a reduction in the mean kinetic energy of the
molecules ; and the intermolecular attractive forces then probably restrict the migra-
tions of the molecules to oscillatory or vibratory motions about their positions of
equilibrium. The low compressibility of solids, and the comparatively slow rate at
which one solid difiuses into another, show that the molecules of a solid have a
comparatively low mobility. One molecule can get away from contact with another
molecule only very very slowly, if at all. Solid diffusion, however, seems to be
confined to those systems in which solid solutions can be formed ; raising the
temperature or subjecting the system to a uniform pressure also appears to augment
the speed of difiusion.^
The fact that many solids evaporate or sublime very slowly, shows that their
molecules probably do possess a certain mobility. Thus, zinc at 370°, though still
a solid, volatilizes to such an extent that a clean copper plate placed just above
appears on the under side to have been coated with brass. The fact that most
solids retain their shape for indefinitely long periods, unless prevented by chemical,
mechanical, or physical actions, shows that the molecules of solids have a very
limited mobility — e.g. some ancient jewellery appears to be the same now as when
first engraved. On the other hand, a mass of pitch may be so brittle as to be readily
fractured by a blow, and yet, when placed on an inclined plane, it gradually loses
its shape, and, following the solicitations of gravity, begins to flow (not slide) down-
wards ; similarly, a long glass tube or rod supported at both ends, gradually sags
in the middle. The substances, glass and pitch, are therefore regarded as extremely
viscous liquids. Accordingly, it is not possible to draw a sharp line of demarcation
between amorphous solids and liquids.
The term solid is therefore ambiguous in that it has at least two meanings. Amorphous
solids are frequently harder and more brittle than the same substance in the crystalline
THE KINETIC THEORY OF ATOMS AND MOLECULES 819
condition, and yet the amorphous state is indistinguishable from the liquid state ; while
a liquid can be readily distinguished from a crystalline solid, it is sometimes said that all
true solids are in a crystalline state, and that amorphous solids are super-cooled liquids.
This view is discussed in G. Tammann's Kristallisieren unci Schmelzen (Leipzig, 1903).
From his investigation of the structure of the co-called amorphous solids, R. Gross infers
that there are probably no true amorphous solids, only crystals and liquids with varying
degrees of viscosity up to the high viscosity of glasses.
The stress and strain of solids. — Many properties of solids — e.g. tenacity,
hardness, etc. — depend on intermolecular forces and also on the grouping of the
particles. In general, the particles are in stable molecular equilibrium because, like
the so-called conservative system, they tend to restore any work done upon them.
For instance, any displacement of the particles within the limits of elasticity produces
a counter or restitutional pressure equal and opposite to the distorting stress. The
effect of cohesion or the attraction of the particles of a solid for one another must
also be attributed to the same molecular forces ; and the cohesive forces are
measured by the amount of force which must be applied in order to overcome them.
The term stress is applied to a force or system of forces which acts upon a body
or system of bodies producing an alteration of form. The change or alteration in
form which is produced by the application of a stress is called a strain, the magnitude
of the stress is usually referred to unit area of surface across which it acts. For
instance, if a bar of metal oi n sq. cm. sectional area and fixed at one of its ends,
sustains a load of w kilograms, uniformly distributed, the longitudinal stress is wjn
kilograms per sq. cm. ; and if a portion of the bar
increases in length from 100 to lOO'Ol cm., and the
increase be uniformly distributed over the por-
tion lengthened, the longitudinal strain will be
(100-01-100)/100=0-0001 cm. per sq. cm.
According to R. Hooke,^ 1676, ut tensio sic vis —
strain is proportional to stress — a relation known as
Hooke's law. Consequently, the strain will be k
times the stress, where ^ is a constant. The observed
relations between the two variables indicate that when
the stress is a compression, the curve showing the
corresponding change of volume. Fig. 12, does not
approach the strain-axis so rapidly as when the
stress is a tension. Hence, the intensity of the
force resisting compressive strains decreases more rapidly than is the case with
the force resisting tensile or dilative strains.
Sirs//? (Change cf Volume)-
Fig. 12.— Hooke's Law.
If a body subjected to a stress experience no strain, it woiild, if it existed, be called a
perfectly rigid body. There are no such bodies.* Consequently, every solid can sustain
stress or transmit force only by suffering strain. A body is said to be perfectly elastic if,
when subjected to a given stress at a given temperature, it experiences a definite strain
which does not increase when the stress is prolonged, and which disappears completely
when the stress is removed. If the form of the body be permanently altered when the stress
exceeds a certain value, the body is said to be soft or plastic, and the state of the body when
the (permanent) alteration is just going to take place, is called the limit of perfect elasticity.
According to the British Standard Specification,
The elastic limit is the point at which the extensions cease to be proportional to
the loads. In a stress-strain diagram plotted to a large scale it is the point where
the diagram ceases to be a straight line, and becomes curved. The yield point is the
point where the extension of the bar increases without increase of load. In practice,
the yield point is the load per sq. in. at which a distinctly visible increase occurs in
the distance between the gauge points on the test-piece, observed by using dividers ;
or at which when the load is increased at a moderately fast rate there is a distinct drop
of the testing machine lever, or in hydraulic machines, of the gauge finder.
If the stress, when it is maintained constant, causes a strain which increases with time,
the substance is said to possess viscosity or to be viscous. According to J. C. Maxwell, a
viscous material is fluid when any stress, however small, produces a constantly increasing
820 INORGANIC AND THEORETICAL CHEMISTRY
strain ; and he draws a distinction between elasticity of bulk and elasticity of shape. The
latter is peculiar to solids. A body possesses elasticity of bulk or volume elasticity, when,
on removal of the stress, it returns to its original volume, even though the form of the surface
be permanently altered. Under a compressive stress, the elasticity of bulk may far exceed
the elasticity of shape. According to Lord Kelvin : •
If we reckon by the amount of pressure, there is probably no limit to the elasticity
of bulk in the direction of the increase of pressure for any solid or fluid ; but whether
continued augmentation produces continued diminution of bulk towards zero without
limit, or whether for any or every solid or fluid, there is a limit towards which it maj^ be
reduced in bulk, but smaller than which no degree of pressure, however great, can
condense it, is a question which cannot be answered in the present state of science.
The volimae elasticity is also called the cubic elasticity or bulk modulus, or the resistant <
to compression, and it is represented in dynes per sq. cm.
If other units are employed it is convenient to remember that a megabar is equivalent
to 10' dynes per sq. cm., or to 0*987 atm., or to 750'15 mm. of mercmy at 0°, sea-level,
and latitude 45°. One gram per sq. cm. is equivalent to 981 dynes per sq. cm.
When a uniform pressure of dp dynes per sq. cm. reduces the volume v of a body to
v—dv the compressibility is measured by the fractional change of volume dv/v per dyne of
applied pressure per sq. cm. The volume elasticity is the reciprocal of the compressibility.
Hence
dv
_ . ^ . Volume elasticity v, = -r- ; or, ^=—v~-
Compressibility, ^=-£; or, ^=-~-£- ^^^^^ elasticity ^ = ^; or, ^
where the negative sign indicates that the volume decreases as the pressure increases. The
volume elasticity is therefore the quotient of a given stress p, by the strain it produces.
For example, if a liquid contracts 50 millionths of its volimae when subjected to a pressure
of 1000 dynes per sq. cm. the modulus of elasticity is 2 X 10' dynes per sq. cm. The com-
pressibility measures the sensitiveness of a body to changes of pressure under the specified
conditions. These results apply to gases, liquids, and solids. The temperature is supposed
to be constant, and the result is called the isothermal elasticity of the gas. If the operation
be conducted adiabatically, pvy ==^a, constant, and by differentiation and rearrangement of
terms, vdpjdv—yp. Hence, the adiabatic elasticity of a gas is y times the pressiu-e, where
y denotes the ratio of the two specific heats of the gas. Similarly, by diJfferentiation of
Boyle's law, pv=a, constant, when the temperature is constant, vdp/dv=p. Hence, the
ratio of the adiabatic and isothermal elasticities of a gas is equal to the ratio of the two
specific heats of the gas.
If a solid bar, supposed perfectly elastic, be exposed to longitudinal stress, the ratio of
the stress to strain or the longitudinal elasticity or the coefficient of resistance to extension, is
called Young's Modulus • — symbolized E. The longitudinal elasticity or Yoimg's modulus
is represented in dynes per sq. cm. This constant is sometimes referred to as the modulus
of elasticity ; though this term is liable to confusion there are so many kinds of elasticity.
If a wire of a sq. cm. cross-sectional area be stretched by a force of F dynes, and its length
changes from I to l-\-dl. Young's modiolus is IF/a.dl. If Yoimg's modulus for a bar of the
given material be independent of the direction in which the axis of the bar is taken, the
material is said to be isotropic, all other materials are said to be ceolotropic. If a bar be
exposed to simple longitudinal traction, the ratio of the lateral contraction to the longitudinal
extension of an isotropic elastic solid— each measured per unit of length- — is called Poisson's
ratio,' and symbolized a. Thus, if the diameter of a bar under a uniform longitudinal
stress changes from 10 to 9*9997 cm., the lateral contraction is 0-0003 cm., and if the longi-
tudinal strain be 0*0001 cm., Poisson's ratio will be 0*3.
A shear is a particular kind of strain in which there is an extension in one direction
combined with an equal compression in a periDendicular direction, as when a sphere is con-
verted into an elUpsoid, or a square into a rhombus. The shear is measiu-ed by the tangential
stress required to produce unit shear, i.e. a shear of one radian, 6, or 57*3°. The resistance
to transverse distortion, the so-called rigidity, n, is therefore n=F/d. The rigidity, or shear
modulus, is expressed in dynes per cm. ; for example, if a tangential stress of 10^ dynes per
sq. cm. deflects a steel rod through an angle of 0*7°, then i'' = 10«, and ^=0*7/57*3=0*0122
radians ; and n=Fld, or the rigidity is 8*2 x lO^^ dynes per sq. cm. This magnitude is
sometimes called the modulus of torsion, but this term is also applied to another concept,
and is best not employed for rigidity.
The relation between the elastic constants are as follows. If ^ be used to denote the
coefficient of compressibility ; E, Young's modulus, or the longitudinal elasticity ; n, the
rigidity ; and a, Poisson's ratio, the relations * between these constants for a homogeneous
isotropic substance are E = 2n{l-{-ff) ; and )8iS=3(l — 2(r).
THE KINETIC THEORY OF ATOMS AND MOLECULES 821
Assuming that the molecules of the different metals show a similarity in be-
haviour in passing from the rigidity of rest at absolute zero, to the vanishing point
of rigidity at the melting point, W. Sutherland ^ found empirically that the relation
between rigidity and temperature can be represented by the simple parabolic
formula :
where n denotes the rigidity at the absolute temperature B° \ iV is a constant supposed
to represent the rigidity at absolute zero ; and T is the absolute melting temperature
of the metal. The following values of 10 -^iV, that is, 10 "^ times the rigidity at
absolute zero, were calculated from the observed values of w at 0° :
Cu
Ag
Au
Mg
Zn
Al
Sn
Pb
Fe
Ni
Pt
452
295
284
164
426
264
200
118
771
781
661
By plotting njN as ordinates and djT as abscissae from zero to unity in each case,
a parabolic curve is obtained on which the values for the different metals fall. This
result establishes the relation between the rigidity of the metals and their melting
points, and according to W. Sutherland shows that " rigidity is in its essence a
kinetic phenomenon almost as simple in character as the elasticity of perfect gases."
The older writers distinguished between cohesion — ^the mutual attraction of
particles of the same substance — and adhesion — the mutual attraction of particles
of different substances — but there is nothing to show that there is any distinction
in kind between the two phenomena.
It is the custom to explain the cohesion of solids as an effect of intermolecular
attraction. The laws of gravitational attraction explain the movements of planets
at great distances, and a similar law for molecular distances served P. S. de Laplace
to explain the surface tension and capillary action of liquids. Many attempts have
been made to correlate the elastic constants with molecular attraction. There are
difficulties in the application of the gravitational law of inverse squares to solids.
According to R. A. Fessenden's hypothesis cohesion is primarily due to the ionic
charges of electricity. The magnitude of the electrical attraction can be calculated
from the magnitude of the charges on the atoms, and the number of atoms per c.c.
Measurements of the electro-chemical equivalent of silver show that the atoms in one
c.c. of silver have a total ionic charge of about 1000 coulombs. The diameter of the
silver atom approximates 10~'* cm. and there are about 10^^ atoms per c.c. Con-
sequently, the total quantity of electricity on a single atom is about 10~24 coulombs,
or, on a single layer of atoms, 10""^ coulombs per sq. cm. Two adjacent layers of
atoms may be regarded as two plates of a condenser 10 ~8 cm. apart and charged
with 10 ~^ coulombs of electricity. Assuming the charges to act as if concentrated
at the centres of the atoms, the force required to separate the two layers will be
44 X 10^ dynes. According to G. Wertheim, the tensile strength of silver is 37 XlO^
dynes per sq. cm., a number in fair agreement with the calculated in view of the
approximations made concerning the size of the atoms.
The metals with the smallest atomic volumes usually have the greatest tensile
strength. It follows that if the atoms in one rod are twice the diameter of those in
another rod, and if the charges behave as if located at the centres of the atoms, then
since the charge on each atom is the same, and the smaller atoms are as close again
as the larger atoms, twice as much work will be required to shear the rod made of
the smaller atoms through a certain angle as with the rod with the larger atoms ;
there will also be twice as many atoms in the smaller rod to be sheared ; conse-
quently, the force required to produce a given shear, i.e. the rigidity of the rod will
vary inversely as the fourth power of the atomic diameter, or as the four-thirds
power of the atomic volume.
The rigidity and Young's modulus vary with temperature, and therefore these
properties can be compared only when the metals are in corresponding states.
822 INORGANIC AND THEORETICAL CHEMISTRY
R. A. Fessenden found empirically that Young's modulus and rigidity are approxi-
mately two-thirds the atomic volume, so that the rigidity and Young's modulus vary
inversel)^ as the square of the atomic volumes. The relation between atomic volume
V and rigidity was pointed out by G. Wertheim ; R. A. Fessenden plots the two
curves: Rigidity, w=28xl0i2i;- 2 ; and Young's modulus, E=78xlOi2tj-2^ and
found the observed values for a number of metals fall close to the values so calcu-
lated. If the attractive force between the atoms with the same charge varies in-
versely as the square of the distance, the attraction of the smaller atoms in a given
rod will be 2^ times as great per atom as with the larger atoms. There will also be 2^
times as many atoms. Hence the tenacity will also vary as the four-third power of the
atomic volume. Assuming the change of the tensile strength with temperature is
proportional to the melting point T° measured from absolute zero, R. A. Fessenden
gives : Tensile strength=0*5208ri;=| kilograms ; and, comparing the calculated
results with G. Wertheim's observations, he finds that —
Fe
Cu
Pt
Ag
Au
Al
Zn
Sn
Pb
Calc. .
. 74
48
48
29
29
18
16
5
4
Obs. .
. 65
41
35
29-6
28-5
18
15-7
3-4
2-36
W. C. Roberts- Austen has shown that the addition of a metal of small atomic
volume to one of large atomic volume will increase the tensile strength of the latter
provided chemical combination does not supervene.
Several have assumed that the cube root of the molecular or atomic volume v
of a metallic element is proportional to the average distance r between adjacent
atoms. For example, G. Wertheim 10 showed that the longitudinal elasticity, E,
increases as the magnitude r decreases ; and H. Tomlinson further showed that for
a number of metals, the product E (grams per sq. cm.) into r^ approximates to a
constant, 1711 X 10®. Hence, H. Tomlinson assumed that the modulus of longitudinal
elasticity varies inversely as the seventh power of the average distance between
the atoms ; that is, as the seven-thirds power of the atomic volume.
If the intermolecular attraction diminishes rapidly as the molecules move apart,
the more a solid is strained by a tensile stress, the less should be the force required
to maintain that strain. This is in direct opposition to the observed facts. Simi-
larly with a compressive stress : the greater the strain, the less should be the com-
pressive strength required to maintain that strain. This also does not agree with
observation. The molecules of a solid under no stress are presumably in a state
of equilibrium but still rotating or oscillating about definite mean positions from
which they do not depart except when constrained to do so by the application of
an external force which is greater than the force or forces which determine the
position of equilibrium of the vibrating molecules. A disturbance of that condition
by compression appears to be opposed by repulsion as a restoring force. This is
shown by the elastic compressibility of solids and liquids. By compression, the
particles of a solid or liquid can be brought closer together, and the mutual repulsive
forces, when developed, result in the storage of energy in the strained body which
makes the particles return to their former position when the pressure is relieved.
Similarly with elastic tension, where the separated molecules are drawn together
by attraction as a restoring force when the tension is released.
The fact that a solid resists both compression and dilation, is usually taken to
demonstrate the existence of an intermolecular force which changes from a repulsion
at small distances to an attraction at greater distances. This means that the
intermolecular force either changes instantaneously from attraction to repulsion
and vice versd, or else there is a neutral zone in which some or all the molecules of
a solid exhibits no cohesion when a solid gradually passes from a state of compression
to a state of tension. There is no evidence of such a state of neutrality. The inven-
tion of an intermolecular force which changes abruptly from attraction to repulsion
is not very satisfactory. The identification of the repulsive forces with the thermal
oscillations of the molecules or the resilience which attends molecular impacts,ii
THE KINETIC THEORY OF ATOMS AND MOLECULES 823
leads to the assumption that the thermal oscillations do not cease at absolute zero
since the elasticity at that temperature is not markedly different from what it is
at ordinary temperatures.
The assumption that the attractive and repulsive forces are independent of one
another is fairly old, and it was developed by G. Mie and E. Griineisen 12 on the
assumption that both forces are inversely proportional to some power of the distances
of the molecules apart.
The molecules of a solid may conceivably possess axes along which the attractive
forces are particularly active. In gases and liquids the molecules are in constant
rotation, and the haphazard translator^ motions as well as the distances between
the centres is so great that the directional forces are inappreciable ; but in crystals
directed forces predominate. The orderly configuration of the molecules in a
crystalline solid shows that the intermolecular forces are directed forces.
There is nothing incompatible with a law of intermolecular attraction which
refers the elastic strength of a solid in tension or compression to the increasing
resistance which a given configuration of the molecules offers to deformation.
Under the influence of a gradually increasing compressive stress, the deformation
' of the molecular configuration continually increases as the particles are forced closer
together, but at a certain point, allotropic or chemical change may occur so that
the atoms rearrange themselves to form new molecules. This is shown very well
by P. W. Bridgman and G. Tammann's experiments on the effects of pressure
on single and mixed solids. When the deformation becomes too great, a form
stable under the new conditions replaces the former configuration. J. C. Maxwell,i3
in an essay on the Constitution of Bodies, considers that the molecules of a solid
oscillate about mean positions so that with certain groups of molecules, the con-
figuration is never very different from the mean stable positions about which the
oscillations occur. This will be the case even if the solid be in a state of strain
provided the amplitude of the oscillations does not exceed a certain limit ; if it
exceeds this limit, the oscillating molecules do not return to their former configura-
tion, but begin to oscillate about new positions of stability in which the strain is
less than in the original configuration. There are probably many groups of such
molecules with oscillations of different amplitude. Thus, the breaking up of any
one configuration depends partly upon the magnitude of the strain on the original
configuration, and partly on the amplitude of the oscillation.
References.
1 J. Johnston and L. H. Adams, Amer. Journ. Science, (4), 35. 205, 1913 ; C. E. van Orstrand
and F. P. Dewey, Prof. Paper U. S. Geol. Sur., 95. 81, 1915.
2 R. Gross, Jahrb. Rod. Electron., 15. 305, 1918.
^ R. Hooke, A Description of Helioscopes, etc., London, 31. 1676 ; Lectures dePotentia Resti-
tutiva or Spring, London, 1678.
^ J. C. Maxwell, Theory of Heat, London, 302, 1908 ; C. Chree, Phil. Mag., (5), 32. 233,
1891 ; British Standard Specification, 56, 1911.
^ Lord Kelvin, Encyc. Brit., 7. 796, 1878 ; Mathematical and Physical Papers, London, 3. 7,
1890.
« T. Young, A Course of Lectures on Natural Philosophy an d the Mechanical Arts, London,
1. 135, 1807.
' S. D. Poisson, Traiti de m^canique, Paris, 1833.
8 G. F. E. Searle, Experimental Elasticity, Cambridge, 18, 1908.
» W. Sutherland, Phil. Mag., (5), 32. 31, 215, 524, 1891.
10 G. Wertheim, Ann. Ghim. Phys., (3), 12. 385, 1844 ; (3), 23. 52, 1849, H. Tomlinson, Phil.
Trans., 174. 1, 1883 ; Proc. Roy. Soc, 38. 42, 488, 1885 ; R. A Feasenden, Jowm. Franklin Inst..
142. 187, 1896 ; Chem. News, 66. 206, 1892.
1^ P. G. Tait, Trans. Roy. Soc. Edin., 33. 65, 1886.
12 G. Mie, Ann. Phys., (4), 11. 687, 1903 ; E. Gruneisen, ih. (4), 39. 257, 1912.
" J. C. Maxwell, Encyc. Brit., 6. 310, 1877.
824 INOKGANIC AND THEORETICAL CHEMISTRY
§ 18. Reactions between Solids — Spring's Experiments
In W. Spring's experiments,! two cylinders of copper were placed end to end
with clean surfaces in contact, and heated to 400° for some hours under pressure ;
the cylinders united together so completely that when the resulting cylinder was
broken, the fracture did not pass through the jointed surfaces. The melting point
of the copper was far above the temperature of the compressed metals, and thermo-
scopic observations have shown that the result cannot be explained by the rise of
temperature due to the compression. Cylinders of aluminium, bismuth, cadmium,
tin, gold, and lead behaved similarly, but the experiment was not successful with
antimony and platinum ; G. Spezia (1911) failed to obtain any signs of combination
with a mixture of finely divided silver and copper after being a month under
a uniform pressure of 8000 atm. W. Spring further reported that precipitated
alumina dried at 140° appeared to flow like a liquid when subjected to a pressure of
5000 atm., but neither natural nor artificial silica showed any signs of flow. Cylinders
of different metals were united by compression, e.g. with cylinders of copper and
zinc, an alloy of brass, about 18 mm. thick, was formed about the plane of contact.
Similar remarks apply to cylinders of iron and zinc, copper and cadmium, tin, or
bismuth. Powdered metals when subjected to a pressure of about 10,000 atmo-
spheres, gave coherent masses which appeared as if they had been fused. By re-
peatedly filing and compressing a mixture of bismuth, cadmium, and tin in the
correct proportions, Wood's alloy was formed ; similarly with Rose's alloy (lead,
bismuth, and tin) ; mixtures of copper and zinc furnished brass ; copper and tin
gave bronze. The criterion for judging the formation of the alloy were " general
appearance and fusibility." W. Hallock 2 showed the futility of the melting-point
test because an alloy may be formed by heating the constituents to a temperature
above the melting point of the alloy, but below the melting point of any single
constituent. W. Rosenhain and P. A. Tucker could obtain only a conglomerate of
tin and lead by compression at 35 tons per s«q. in,, and similar results were obtained
with copper and silver by G. Spezia. ^ In agreement with G. Massing, J. Johnston
and L. H. Adams say :
Compression alorie does not result in the production of true alloys. In those systems
in which the metals crystallize out from the melting as pure components, there is no essential
difference between the conglomerate produced by compression and that obtained by fusion.
In all other systems, the immediate effect of compression is slight, and consists solely in
the fact that compression brings about very intimate contact between the particles, the
result of which, in turn, is the formation at the contact of a compound (if such be possible)
or the promotion of diffusion (when the metals can form mixed crystals).
J. Johnston and L. H. Adams contend that in Spring's experiments the results
must be attributed not to the pressure itself, but rather to the grinding stress to
which the material was subjected, for W. Spring frequently speaks of his materials
squirting out around the piston which did not fit the cylinder very well. For
the effects recorded by W. Spring can be more simply observed by grinding the
materials in a mortar ; the alloys produced by W. Spring's process are agglomerates.
Pressure can favour the formation of alloys by bringing into good contact metals
capable of forming solid solution, this process going on very slowly by the diffusion
of one metal in another. A uniform pressure can fuse only those metals whose
melting points are lowered by pressure, and few metals other than bismuth satisfy
this condition. On the other hand, a grinding process may produce results which
are not obtained by a uniform pressure— e. 7. calcium carbonate ground in a mortar
loses a little carbon dioxide, but not if this compound is subjected to a uniform high
hydrostatic pressure.
This relation between the volume of liquid and solid phases and the lowering
of the melting point dT by a pressure dy, is given by Clapeyron's equation :
THE KINETIC THEORY OF ATOMS AND MOLECULES
825
when V2 denotes the specific volume of the liquid, v-^ that of the solid, T^ the melting
point on absolute scale of temperature, and A the latent heat of fusion. This means
that a pressure acting uniformly on the solid and liquid phase of a single substance
raises or lowers the melting point according as the process of melting is accompanied
by an increase or decrease of volume.
The effect of a non-uniform pressure on solids. — The volumes of most of the
silicates increase during fusion.-* For instance, the specific gravity of a sample of
fireclay was 2*627 before fusion, and after fusion 2'4:70. This corresponds with an
expansion of about 6 per cent. Consequently an increase of pressure should raise,
not lower, the fusion temperature of the clay.
It has been assumed that the latent heat of fusion A is constant, and independent
of temperature. Experiment shows that this assumption is generally valid, although
G. Tammann 5 and others have discussed the possibility of the latent heat of fusion
changing from a positive to a negative value. G. Tammann has also shown that
since a liquid is usually more compressible than a solid, a positive value of V2,—Vi
will diminish with an increase of pressure, and, after passing through zero, will
gradually assume an increasing negative value. It is not probable that the dis-
crepancy between theory and practice is to be attributed to either of these possi-
bilities, rather is the formula not applicable to the case under discussion. Theory
assumes that the pressure is uniformly exerted in all directions, whereas the inter-
stices between the grains of a pulverulent material would give the pressure the
character of a shearing stress. The solid and liquid phases do not therefore sufEer
the same increase of pressure.
H. W. B. Roozeboom ^ uses the following illustration : If Vg and Vp Figs. 13 and 14,
represent the vapour pressure curves of the solid and liquid states at a pressure p, and
Temperatures
C B A
Fig. 13.
V g and V'l represent the vapour pressure curves of the solid and liquid at a higher pressure,
when the two phases- — -liquid and solid — are subjected to the same pressure, the point of
intersection A of the Vg and Vi curves will represent the melting point of the substance
luider a pressure p ; similarly, the point of intersection B of the curves V g and V'l will
represent the melting point of the substance under a pressure higher than p. When the
solid phase alone is subjected to the increased pressure it will melt at the point of inter-
section G of the Vi and V g curves. The temperature OCJ is always less than OA whether
OB be greater or less than OA . Consequently, the melting point oj a solid will always he
lowered when the pressure acts on the solid hut not on the liquid. E. Rieke calculates that by the
application of a tensile or compressive stress p, the lowering of the melting point =vT^p^l2E,
where E represents the elasticity of the solid in the direction of the applied stress p ; A the
latent heat of fusion ; 7'^ the absolute melting point ; and v the specific volume of the solid.
J. Johnston and L. H. Adams have shown that pressure decreases the stability
of a phase which then exhibits an increased tendency to pass into another phase ;
otherwise expressed, pressure acting only on a solid phase increases its vapour
826 INORGANIC AND THEORETICAL CHEMISTRY
pressure, and the solubility in a given solvent. Pressure also lowers the melting
point in accordance with the relation :
§=-^ «
when Vi is the specific volume of the solid and the temperature and pressure in
question. Since the quantities on the right are always positive, the application of
an excess o{ pressure on the solid phase alone always lowers the melting point.
If dTi refers to the depression of the melting point when the pressure dp acts on the
solid phase alone, and dT the depression when both phases are subjected to the
same pressure dp, the combination of equation (1) and (2) furnishes the relation
dTi vi
dT iJ2'~'^i
meaning that the ratio of the lowering of the freezing point of the solid phase,
when this alone is subjected to pressure, to that observed when the same given
pressure acts on both phases, is equal to the ratio of the specific volume of the solid
phase to the change of specific volume on freezing. This result shows how much
the melting point is lowered when the pressure acts on the solid phase alone.
For example, with a uniform pressure, the melting point of ice is lowered 0*00752
per atm., but by an unequal pressure the melting point is lowered 0*09° per atm. —
i.e. twelve times as much. J. Johnston further estimates that if D denotes the
specific gravity of the solid, T the normal melting point on the absolute scale, the
lowering of the melting point is 0'0242T^/Z)A per atm. Thus, if potassium melts
at 62° {i.e. 335° K.), and its heat of fusion is 15-7 cals., and specific gravity 0*87,
its melting point will be lowered 0*59° per atmosphere unequal pressure. It
is therefore important to distinguish clearly between the effect of uniform and of
non-uniform pressure in all discussions on the effects of compression on solid
systems ; neglect to do so has given rise to some apparently contradictory state-
ments.
Chemical action between compressed solids. — W. Spring claimed to have
crystallized amorphous substances like bismuth, zinc (130°), manganese dioxide,
zinc and lead sulphides, mercuric iodide, and transformed plastic or monoclinic
sulphur into the rhombic varieties. These statements have been contradicted
by C. Friedel, E. Jannettaz, and J. Johnston and L. H. Adams. The latter say
that compression alone will not in general produce crystallization, or transform
one modification of a substance into another. It is true that a uniform or nearly
uniform pressure will tend to produce any reversible (enantiotropic) transformation
in favour of the system with the smaller volume, but it often does not do so because
the reaction velocity is very small. In irreversible (monotropic) changes, no
positive statement can be made as to the influence of pressure except by empirical
trials. The effects produced by uniform pressure are comparatively slight. If
the pressure, uniform or non-uniform, be such as to make the substance melt at the
temperature of the investigation, crystallization or recrystallization may ensue ;
but pressures up to 15,000 atm. do not convert calcite — sp. gr. 2"71 — into the denser
aragonite — sp. gr. 2*93 ; nor marcasite — sp. gr. 4 "9 — into the denser pyrites — sp. gr.
5'0 — and this even at 425° under 2000 atm. pressure when the change takes place
under ordinary pressures at about 450°.
W. Spring claims to have formed sulphides and arsenides of the metals by the
alternate compression and filing of intimate mixtures of the metals with sulphur
or arsenic respectively. C. Friedel considered that sulphides were really formed
in Spring's experiments because of the change of colour, which he regarded as a
more certain indication than the evolution of hydrogen sulphide when treated with
hydrochloric acid, and because he found that a mixture of zinc and sulphur gave rise
to the evolution of this gas. W. Hallock, however, obtained sulphides at ordinary
THE KINETIC THEORY OF ATOMS AND MOLECULES 827
temperatures with copper and sulphur an inch apart and with a wad of cotton
wool in between. He claims that it is the vapour of sulphur which attacks the
copper. The same remark applies if the sulphur is replaced by mercuric chloride.
There is no reason to doubt that combination did occur even if the reactions did
not go to completion, as W. Spring, judging from a microscopic examination, thought
really occurred. Hence, although chemical combination no doubt occurred in
these experiments — just as the trituration of two solids will sometimes induce
combination — yet it does not follow that pressure, fer se, promotes chemical
action. Further examples of the chemical union of solids were reported by
W. Spring to occur by compressing mixtures of sodium carbonate and barium
sulphate ; sodium sulphate and barium carbonate ; potassium nitrate and sodium
acetate ; lead chloride and potassium nitrate ; mercuric chloride and potassium
iodide ; arsenious oxide and cadmium nitrate ; etc. Dry sodium nitrate and zinc
chloride or sulphate react when shaken together forming zinc nitrate ; the heat of
the reaction decomposes part of the latter forming zinc oxide and brown vapours.
T. von Hagen studied the effect of pressure on powders with the object of preparing
tabloids by compression.
While chemical reaction may be favoured by close contact, by the grinding
action which attends the application of pressure, and by the slow diffusion which is
possible in certain systems, J. Johnston and L. H. Adams have pointed out that
VV. Spring's tests for chemical combination were defective. For example, the
small cylinder obtained by compressing a mixture of powdered and dried anhydrous
sodium carbonate and barium sulphate to 3000 atm. was pulverized and " washed
completely in cold water ; the insoluble residue filtered off, and analysed."
W. Spring's results, therefore, show the composition of the compressed mass after the
addition of water, but not the composition of the dry solid phases, for it is known
that the system, BaS04+Na2C03^Na.2S04+BaC03, when in the presence of water
reaches equilibrium after the elapse of a certain time. In the case of potassium
nitrate and sodium acetate, the mixture was left four months in a desiccator, and he
noticed the mixture was deliquescent. The original salts do not deliquesce ; hence,
said W. Spring, some deliquescent potassium acetate must have been formjed. Here
again, the equilibrium was tested after exposure to water vapour, but there is no
indication of the state of equilibrium of the dry solids. Mere trituration of the
pure and dry salts in the presence of water vapour suffices to start the deliquescence.
In general, only those reversible reactions — physical or chemical — which are
accompanied by a diminution in volume are favoured by an increase of pressure.
The fact that substances can react in the solid state by trituration or grinding, or by
the application of a non-uniform pressure, shows that the molecules of solids can be
brought close enough for chemical union. The older chemists did not believe that
solid substances could react chemically, and their experience is summarized in the
oft-quoted phrase : corpora non agunt nisi soluta — substances do not react unless
they are in solution. J. L. Gay Lussac combated this dogma as far back as 1846.
He said :
II est certain, au contraire, que tous les corps solides, liquides, ou a^riformes agissent
les iins sur les autres, mais que, des trois etats des corps, Fetat solide est le moins favorable
k I'exercice de I'affinit^.
References.
1 W. Spring, Bull. Acad. Buy. Belg., (2), 45. 746, 1878 ; (2), 49. 323, 1880 ; (3), 5. 221), 492,
1883 ; (3), 6. 507, 1883 ; (3), 10. 204, 1885 ; Bull. Sac. Chim., (2), 10. 204, 1885 ; Ann. Soc.
Geol. Belg., 15. 156, 1888 ; Zeit. phys. Chem., 2. 536, 1888 ; 15. 65, 1894 ; G. Spezia, Atti Accad.
Torino, 45. 1, 335, 1911 ; W. Hallock, Amer. Journ. Science, (3), 34. 277, 1887 ; C. Friedel, Bull.
Soc. Chim., (2), 39. 626, 1883 ; (2), 40. 526, 1883 ; W. Spring, ib., (2), 40. 520, 1883 ; E. Jannettaz,
ib., (2), 40. 51, 1883 ; Bull. Soc. Geol. France, (4), 12. 227, 1884.
2 W. Hallock, Atner. Journ. Science, (3), 37. 402, 1889 ; W. Rosenhain and P. A. Tucker,
Phil. Trami., 209. A, 89, 1909 ; G. Spezia, Atti Accad. Torino, 45. 1, 1910 ; G. Masing, Zeit. anorg.
Chem., 62. 265, 1910.
828 INORGANIC AND THEORETICAL CHEMISTRY
• J. Johnston and L. H. Adams, Amer. Journ. Science, (4), 35. 204, 1913 ; Journ. Amer.
Chem. Soc, 34. 563, 1912 ; 66. 361, 1912 ; J. Johnston, Journ. Franklin InsL, 183. 1, 1917.
* G. Bischof, Neues Jahrb. Min., 17, 1845 ; C. Doelter, ib. ii, 141, 1901 ; C. Bams, Phil.
Mag., (5), 35. 173, 1893 ; J. Thoulet, Bull. Soc. Min., 3. 34, 1880 ; F. Niess, Ueber das
Verhalten der Silicate beim Uebergange aus dem Oluthflusaigen in den f eaten Aggregatzu^tand,
Stuttguart, 1889.
5 G. Tammann, Kristallisieren und Schmelzen, Leipzig, 162, 1903.
« H. W. B. Roozeboom, Die heterogenen Oleichgetvichte, Braunschweig, 1. 213, 1901 ; E. D.
Williamson, Phys. Rev., (2), 10. 275, 1917 ; J. Johnston, Journ. Amer. Chem. Soc, 34. 789, 1912 ;
J. Johnston and L. H. Adams, Amer. Journ. Science, (4), 35. 205, 1913 ; J. Johnston and
P. Niggli, Journ. Geol, 21. 602, 1913 ; E. Rieke, Ann. Physik, (3), 54, 731, 1895 ; Centr. Min., 97,
1912 ; J. H. Poynting, Phil. Mag. (5), 12. 32, 1881 ; W. Ostwald, Lehrbuch der allgemeinen
Chemie, Leipzig, 2. ii, 374, 1902 ; G. N. Lewis, Proc. Amer. Acad., 36. 145, 1900 ; 43. 268, 1907 ;
Zeit. phys. Chem., Zb. 346, 1900; 61. 139, 1908; H. le Chatelier, ib., 9. 335, 1892; H. von
Hagen, Zeit. Elektrochem., 2b. 376, 1919; R. Wegscheider, Zeit. anorg. Chctn., 93. 95, 1915;
M. Hasselblatt, ib., 93. 75, 1915 ; G. Tammann, ib., 92. 37, 1915 ; P. W. Bridgman, Phys. Bev.,
(2), 7. 216, 1916; E. D. Williamson, *6., (2), 10. 140, 1917.
§ 19. The Vibration Frequency o! Atoms and Molecules
The more boldly we advance beyond experience, the broader the survey we obtain,
the more surprising the facts we discover, but the greater the likelihood of one going astray.—
L. BOLTZMANN (1!
In recent years, the mathematical treatment of the theory of solids has been
largely based upon the quantum theory principally in connection with the specific
heats of solids, but also in connection with the co-relation of the physical properties
of solids with a property called the periodic time or the vibration frequency —
Schwingungszahl — of the atoms or molecules. W. Sutherland i made an attempt
in 1890. He showed that if the molecules of a solid vibrate about a mean position,
it can be assumed that at the melting point the vibratory motion will just break
down ; and the vibration frequency or the period of vibration of the elements at
their melting points becomes
Vibration frequency v^=Ka. / — —
where Z is a constant. W. Sutherland's argument is somewhat as follows :
Let a molecule or atom of mass M and mean specific heat C, be heated from absolute
zero to its melting point 2'^ ; it will receive heat MCT^, and this will be proportional to
the kinetic energy \MV^ of the molecule, where V is the velocity of vibratory motion at
the melting point, provided the body undergoes no expansion when heated. By Dulong
and Petit's rule, MG is almost constant for the elements, and therefore V. the mean velocity
of the vibrations at the melting point is proportional to NT^jM. If D denotes the density
of the element, M/Z) will represent the molecular volume, v ; and if a denotes the mean
coefficient of linear expansion of the substance between absolute zero and T^, the increase
in the linear dimensions of the space occupied by the molecule when heated from zero to
Tjf^ will be aT^t'». This represents the length or amplitude of the vibrations just as
the molecule is going to leave the vibratory state characteristic of the solid, on the
assumption that the amplitude of the vibrations of the particles is augmented as
the temperature rises from absolute zero to the melting point, at which temperature the
amplitude becomes comparable with the distances of the molecules apart. At the melting
point, the crystalline form of the solid is destroyed. Hence, the vibration frequency of
the molecule at the melting point is proportional to aT^{MID)klT^IM. Assuming that
aT^ is a constant, a relation verified by E. GrUneisen,^ it follows that if the constants bo
collected together the vibration frequency is proportional to {MID)\{{T^M)h.
W. Sutherland estimated the relative vibration frequencies of the elements of
the alkali family to be in the proportion :
Li
Na
K
Bb
Cs
0-21
0*43
0-66
0-96
1-23
THE KINETIC THEORY OF ATOMS AND MOLECULES
i.e. approximately as 1 : 2 : 3 : 4'5 : 6 ; and for the alkaline earth metals :
Be
Mg
Ca
8r
Be
0-35
0-70
104
1-62
1-88
or approximately as 1, 2, 3, 4:'5, 5*3. Similarly, for other groups when data are
available. Consequently, the periods of vibration of the molecules at their melting
points show simple harmonic relations. Analogous results were obtained with
some compounds of the elements. A precisely similar expression^was obtained
by F. A. Lindemann ^ for the relation between the vibration frequency, y, and
the melting point, Tm- F. A. Lindemann gave the constant the empirical value
2-06x1012, which W. Nernst afterwards altered to 3-08x1012. The vibration
frequencies of the atoms of the solid have been determined by a number of inde-
pendent methods and the average values so obtained agree fairly well with one
another.
(1) From the melting point. — The vibration frequency v by means of W. Suther-
land's formula, with W. Nernst's constant.
i;=3-08xl0i2,
V M
Mv^'
or, v=
3-08x1012
^Xl0i2 /j^
vi V M
(1)
where M denotes the molecular or atomic weight ; T^, the absolute melting tempera-
ture ; and v, the atomic weight. The agreement between the observed vibration
frequencies and those calculated by F. A. Lindemann' s formula is very fair :
Al
Cu
Zn
Ag
Pb
Diamond
vx 10-12 (Observed)
8-3
6-6
4-8
4-5
20
40
j^X 10-12 (Calculated)
8-4
7-5
4-8
4-8
1-9
35
The low atomic weights and the high melting points of these elements, diamond,
boron, and silicon, give these elements abnormally high vibration frequencies.
Lithium has a low atomic weight, but its low melting point gives its vibration
frequency a normal value. The atomic frequencies of the elements calculated
from F. A. Lindemann's formula * are indicated in Table XXVI.
Table XXVI. — Atomic Vibration Fbequencies of the Elements.
Element.
l/XlO-12
Element.
vxlO-12
Element.
vxl(ri2
Hydrogen .
4-88
Iron
9-11
Antimony
3-22
Helium-liquid
0-66
Cobalt
8-87
Tellurium .
2-69
Lithium
10-65
Nickel
8-86
Iodine
1-82
Beryllium .
23-65
Copper
7-40
Xenon-liquid
0-85
Boron
28-10
Zinc
4-79
Caesium
1-12
Carbon-graphite .
27-70
Gallium
2-82
Bariiun
2-66
Carbon-diamond .
31-70
Germanium
6-23
L anthanum
3-04
Nitrogen
2-50
Arsenic
4-20
Cerium
2-86
Oxygen
2-54
Selenium-grey .
2-79
Praseodymium
3-24
Fluorine-liquid
1-80
Bromine-liquid •
1-70
Needy mium
3-11
Neon-liquid .
0-34
Krypton-liquid .
1-90
Samariimoi .
3-76
Sodium
4-31
Rubidium
1-54
Tantalum
5-72
Magnesimn .
7-88
Strontium
3-44
Timgsten
6-06
Aluminium .
8-33
Yttrium .
4-07
Osmium
5-96
Silicon
10-50
Zirconium
4-63
Iridiimi
5-47
Phosphorus, red .
6-72
Columbium
6-73
Platinum
4-76
Phosphorus, white
3-83
Molybdenum
5-57
Gold
3-69
Sulphur-rhombic .
4-30
Ruthenium
6-99
Mercury
1-38
Sulphur-monoclinic
4-24
Rhodium .
7-01
ThaUivun
2-00
Chlorine-liquid
2-24
Palladium
6-16
Lead
1-99
Argon-liquid
1-32
Silver
4-80
Bismuth
1-80
Potassium .
2-53
Cadmium .
3-01
Thorimn
3-06
Calcium
4-28
Indium
2-37
Uranium
4-67
Scandium
6-84
Tin .
2-50
Chromium
9-23
Titanium
9-17
Vanadium
9-26
Manganese
8-35
830
INORGANIC AND THEORETICAL CHEMISTRY
The numbers are based on the assumption that the molecules of the solid are mon-
atomic. It is also assumed that the molecular structure remains unchanged—
without polymerization or dissociation — in cooling from the melting point to
absolute zero. If the molecular weight changes from M to nM, the molecular
volume will be increased n times. Consequently, in virtue of the terms mi and v^,
the frequency calculated by F. A. Lindemann's equation must be divided by
n^+i, or nK The molecular volume is usually computed from density deter-
minations at ordinary temperatures ; and it is assumed that the molecular volume
is a constant, whereas it is not likely that the molecular volumes of a substance
at absolute zero and at the melting point are the same ; hence the formula
probably needs a correction term for the change in the molecular volume
of temperature.5 It must also be emphasized that the constant has been
evaluated empirically, and that a series of approximations have been made in the
100 150
Atomic Weights.
Fig. 15. — Vibration Frequencies and Atomic Weights.
250
deduction of the formula. The different allotropic forms of an element may have
difierent frequency numbers. The atomic frequencies of the elements also exhibit
periodic properties, as W.Sutherland (1890) and later W. Biltz (1911) have shown.
The curve, Fig. 15, recalls L. Meyer's atomic volume curve, Fig. 14, Cap. VI
Thorium, manganese, tin, and tellurium do not sit on the curve. Argon and
potassium occupy their anomalous position; hydrogen seems to fall with the halogens.
The unique chemical character of carbon in its great capacity for forming compounds
seems to correspond with its exceptionally large vibration frequency. H. S.
Allen ^ has shown that if 'N be the atomic or Moseley's number for an element, a
large proportion of the elements furnish values of the product Nv=^7ia, where n
is a whole number, and a is a constant, approximately 21*3 X 10^2 ; and for a few
others — lithium, beryllium, sodium, magnesium, phosphorus, sulphur, potassium,
arsenic, rubidium, indium, iodine, mercury — n involves fractions. E. Griineisen
THE KINETIC THEORY OF ATOMS AND MOLECULES 831
found the product of the linear expansion coefficient, a, and the absolute tempera-
ture of the melting point to be a constant ; H. Alterthum substituted the reciprocal
of a for Tm in Sutherland and Lindemann's formula, and obtained
v=4-2xlOii\/^'^ , .... (2)
^ Mavi
The deviations of the results with this formula from others are probably due to
the use of values of a determined at ordinary temperatures.
(2) Fro7n the elastic constants of the solid. — The relation between the natural
vibration frequency v of the atoms, the atomic weight M, the compressibility j8,
and the density Z) of a solid was computed by A. Einstein (1911) to be
v=3-3xW^y^^ . . . . (3)
showing that the forces which produce the thermal oscillations of the solid are
the same in kind as those which produce elastic oscillations — e.g. sound waves.
A. Einstein's constant was 2 '77 X 10^, but the empirical value 3'3 X 10*^ was found to
give more satisfactory agreement with the values obtained by the preceding
formula. F. A. Lindemann modified A. Einstein's formula to
v=3-6S X lO'^M-iD-ip-i
E. Griineisen (1912) "^ obtained a relation (4) between the coefficient of
linear expansion, a ; the compressibility, j8 ; and the specific heat at constant
volume, Cv, in gram calories per degree. When this relation was combined with
A. Einstein's specific heat formula, an expression for the vibration frequency is
obtained :
v=2-92xlOiV^5, .... (4)
3avi
This shows a good agreement with values obtained by other methods.
(3) From the frequencies of the longest heat waves in the dispersion s'pectrum. —
The natural vibration frequency v of the atoms of the elements for any temperature
can be calculated from the abnormally low specific heats. In order that the specific
heat may be less than normal at ordinary temperatures — say, 27°, when T=300 —
the vibration frequency v must lie between 6*5 XlO^^ and 6*5 X 10^2^ Vibration
frequencies of this range lie in the infra-red spectrum, and therefore all elements
with abnormally low specific -heats must have a value of v in this part of the
spectrum. Adapting a table from N. R. Campbell's Modern Electrical Theory
(Cambridge, 242, 1913), it has been found
S and P
Fl
0
Si
.B
H
C
Gv
. 5-41
5-0
3-99
3-81
2-68
2-32
1-79
i/X 10-12
. 7-1
9-1
14-3
150
20-0
23-1
250
These values of v are in fair agreement with such observations as have been made
on the absorption frequencies of the atoms in compounds of these elements. The
agreement is surprising, because, as N. R. Campbell says, it might have been antici-
pated that the forces under which the atoms would vibrate in the compounds
would be entirely different from the vibrations of these same atoms when they
form part of a solid element. While the absorption bands in the infra-red probably
represent the free vibrations of the atoms and not of the electrons inside the atoms,
the vibrations of transparent bodies in the ultra-violet are probably contributed by
the electrons. The spectrum shows that the vibrating atoms have many degrees
of freedom, which are very important optically, but they possess such high values
of V that they do not absorb appreciable amounts (quanta) of energy when the
body is heated.
832
INORGANIC AND THEORETICAL CHEMISTRY
According to S. Pagliani,® the vibration frequency r of a molecule can be
calculated from the vibration frequencies vi and V2 of its component atoms by the
formula v=f(vi+»'2) J ^^^ i* is connected with the entropy cf) of the compound by
the expression v=9'6xl0i2<^.
The specific heats of compounds. — ^At ordinary temperatures the additive nature
of the specific heat laws of F. E. Neumann, J. P. Joide, and H. Kopp show that
the heat energy of the molecules is mainly derived from the vibrations of the indi-
vidual atoms. At high enough temperatures, the vibrational energy of the atoms
approaches the value 3RT ; but at low temperatures, W. Nernst ^ assumes that
the vibrations of the molecules play a more important part than the vibrations of
the atoms in the molecule. W. Nernst further assumes that the heat energy of a
compound in the solid state is made up of the energy due (i) to the motions of the
molecules relative to one another, and (ii) to the internal energy of the molecules
owing to the vibrations of the atoms in the molecule. The first contribution is
calculated by P. Debye's formula ; the second by A. Einstein's. Each calculation
involves a knowledge of the characteristic vibration frequencies — the first, symbolized
by vi, is given approximately by F. A. Lindemann's formula ; the second,
symbolized by V2> obtained by the optical measurements of H. Rubens' residual
rays — Reststrahlen ^^ — by repeated reflexions from the surfaces of solids. It is
assumed that the frequency of such infra-red radiations corresponds with the
frequency of vibration of an electrically charged ion which may be identified with
an atom, and that the forces which control the heat vibrations of the atoms of a
solid are the same, whether specific heats or the reflexion of infra-red radiations are
involved. Consequently, the specific heat of the compound will be :
2C,:
d (Debye's function) d (Einstein's function)
dT
dT
According to W. Nernst, for KCl, ^v-^^im and|/Si^2=2I3'5, where j8=4-78
X 10^1. The calculated mean values of 0^, at different temperatures are compared
with the observed values of Cp in Table XXVI. The agreement between hypo-
thesis and observation is good. The analogy between the results with the
atoms of the elements and the molecules of compounds has led H. S. Allen to
infer that
The forces binding the atoms in the molecule are similar in character to those which
bind the molecules of the solid, that is, the forces of chemical affinity are of the same
nature as the forces of molecular cohesion.
Table XXVII.— Specific Heat of Potassium Chloride.
Internal vibrations.
Calculated
Observed
2C,.
. Difference
between
C^ and C^,
Temperature
Cp from
Debye's
formula.
C^ from
Einstein's
formula.
22-8
301
48-3
70-0
235-0
5600
1-04
1-87
3-52
4-67
5-81
5-93
0-046
0-25
1-43
2-89
5-55
5-87
1-086
2-12
4-95
7-50
11-68
12-70
1-16
1-96
5-70
7-68
11-78
13-08
0 04
-008
0-37
0-04
0-05
019
It is found that if the single ions are situated at the points of the space lattice of a
crystalline solid, the mean atomic weights can be substituted in the formula for
calculating the frequencies. The agreement is then not nearly so good as when
the residual rays of H. Rubens are assumed to be reflected from the molecules as
THE KINETIC THEORY OF ATOMS AND MOLECULES 833
a whole. The calculated results with mercurous chloride and with water agree
with observations only when more complex molecules than HgCl and H2O are
respectively assumed.
If the values of the vibration frequencies are known, it is possible to calculate
the molecular weight of elements and compounds from Lindemann's formula by
substituting molecular weight and molecular volume in (33). The observed values
for crystals of lead, silver, zinc, copper, aluminium, and carbon in the diamond
agree with those calculated on the assumption that these elements are monatomic ;
with sulphur and graphitic carbon, the molecules are more complex — with sulphur
probably Sg. Similarly with crystals of sodium chloride, NaCl ; potassium chloride,
KCl ; potassium bromide, KBr ; silver chloride, AgCl ; lead chloride, PbCl2 ;
and benzene, CgHg. The molecular weights so determined agree with the formulae
usually assigned to these compounds, whereas with crystals of water, silica, and
mercurous chloride, the ordinary formulse are doubled so as to furnish H4O2, Si204,
and Hg2Cl2 respectively.
The relation between the vibration frequency and the atomic heat of fusion. —
If A denotes the latent heat of fusion of a crystalline solid, and M the atomic
weight, the number of calories required to melt a gram-atom of the solid will be
If A. The energy required to melt the solid may be regarded as equivalent to the
work required to rupture the bonds which hold the crystal units in position, and
enable the molecules to move freely amongst themselves. According to the quantum
theor)'-, the energy of a solid is the energy of the oscillators which it contains, and
H. S. Allen 11 assumes that the latent heat of fusion is equivalent to the energy
required to counterbalance the energy of a certain number of oscillators. The
average amount of energy associated with a vibration frequency v, at a temperature
T, is RTul(e^—l), where u is put in place of hvjRT, and h is Planck's constant, R
the gas constant per gram-molecule. If c denotes the ratio of the number of
oscillators to the number of atoms, there will be cN oscillators in a gram-molecule,
where N denotes Avogadro's constant. The total energy of the atoms in a mon-
atomic solid is accordingly
,,. cNRTu MX u
^^=-^^'''''NRT='^l
NR is the value of the gas constant per gram-molecule, i.e. 1'989 calories per degree.
The vibration frequencies of a number of elements are known, and consequently the
values of c can be computed. H. S. Allen finds the value of c for nickel, cobalt,
rubidium, sodium, potassium, iron, silver, lead, copper, and palladium to be nearly
unity ; for aluminium, mercury, cadmium, and platinum, IJ ; for zinc, 1 J ; for
tin, If ; and for gallium and bismuth, 2J. H. S. Allen's expression for the latent
heat of fusion A resembles H. Crompton's formula indicated in the next section.
The relation of H. Crompton's formula to H. S. Allen's is seen by writing the
latter
t^=NR- "
Tc 6«-l
when, as H. S. Allen shows, the expression w/(ew— 1) is not far from unity, the right
side of the equation is approximately constant and equivalent to H. Crompton's
relation. A great deal of the work which has been done with the classical
doctrine of energy can now be translated into the language of the quantum
theory.
References.
1 W. Sutherland, PJiit. Mag., (.5), 30. 318, 1890 ; (5), 32. 524, 1891.
2 E. Griineisen, Ann. Physik, (6), 39. 257, 1912.
* F. A. Lindemann, Phys. Zeit., 11. 609, 1910.
VOL. I. ' 3 H
834 INORGANIC AND THEORETICAL CHEMISTRY
* H. S. AUen, Phil. Mag., (6), 34. 478, 1917 ; S. Pagliani, Atti Accud. Lincei, (5), 24. 835,
948, 1915 ; W. Biltz, Zeit. Elektrochem., 17. 670, 1911.
5 W. Sutherland, Phil. Mag., (5), 30. 318, 1890 ; E. Gruneisen, Ann. Physik, (4), 39. 298,
1912.
« H. S. Allen, Proc. Roy. Soc, 94. 100, 1917 ; Phil. Mag., (6), 34. 478, 1917.
7 E. Gruneisen, Ann. Physik, (4), 39. 257, 1912.
8 S. Pagliani, Atti Accad. Lincei, (5), 24. i, 943, 1915.
• W. Nemst, Vartrdge liber die kinetische Theorie der Materie und der Elektrizitdt, 79, 1914;
The Theory of the Solid State, London, 85, 1914; H. S. AUen, Phil. Mag., (6), 35. 338,
404, 1918.
1° H. Rubens and H. Hollnagel, Sitzber. Akad. Berlin, 26, 1910 ; H. Rubens, ib., 513, 1913 ;
H. Rubens and H. von Warenberg, ib., 169, 1914 ; H. Rubens, Verh. deut. phys. Ges., 13. 102,
1911.
11 H. S. Allen, Proc. Phys. Soc, 28. 204, 302, 1916.
§ 20. Empirical Relations between the Properties of Solids
We cannot attain to a real theory of chemistry until we are able to connect the science
by some hypothesis with a general theory of dynamics. — A. C. Brown (1874).
The kinetic theory of solids is in the earlier stages of its development in that it is
based upon very imperfect knowledge and arbitrary assumptions. There have
been quite a number of attempts to develop the subject. i The greatest success
has been obtained by considering the properties of solids at low temperatures as a
limiting case in the same sense that gases have been more amenable to mathematical
treatment when in a rarefied condition. The amplitude of the oscillating molecules
at low temperatures is probably small, and but a relatively small number are vibrating.
This assumption is in conformity with the fact that many properties of solid bodies
at very low temperatures are small, and vary almost proportionally with the tem-
perature— this applies, for example, with the specific heats, coefficients of thermal
expansion, the temperature coefficients of compressibility, etc., at low temperatures.
C. M. Guldberg (1868) sought to establish the relation
pv=RT--p log -
as the equation of state of ideal solids in the same sense that pv=RT is the equation
of state of ideal gases. In C. M. Guldberg's equation, R, jS, and Vq are constants
which have specific values for different substances — Vq is considered to be the
specific volume of the substance at absolute zero. It is further shown that if w
denote the weight of a cubic metre of the solid ; E, the modulus of elasticity ; a,
the coefficient of thermal expansion ; and A, the latent heat of fusion, these constants
are related so that E=wK ; awj3=424:0 ; j8=4xl04A. The agreement between
the observed values and those calculated by these relations is remarkably good,
and it shows that some intimate relation probably subsists between the latent heat
of fusion, the coefficient of thermal expansion, and the elasticity or compressibility
of solids.
E. Griineisen has also found empirically that the quotient of the coefficient of linear
expansion, a, and the specific heat at constant pressure, Cp, of a number of metals is nearly
a constant — -that is, a/6'p = a constant ; he also found empirically that the product of the
atomic volume, v, and the coefficient of linear expansion, a, divided by the product of the
coefficient of compressibility, jS, and the atomic heat, Cv, is a constant' — that is, avj^Cv
=a constant ; and that the product of the coefficient of linear expansion, a, and the
absolute melting temperature, Tm'> is nearly constant — that is, aTm — a constant,
R. Pictet 2 assumed that the mean values of the amplitudes of the vibrating molecules
in a melting solid are always the same ; that the product of the mean distances of the
molecules and the coefficient of linear expansion, a, is proportional to the absolute
temperature of the melting point, Tm ; and that the product of all three variables is constant.
THE KINETIC THEORY OF ATOMS AND MOLECULES 835
In ignorance of the absolute distances between the vibrating particles, and since the volume
of a solid is proportional to the cube of the linear dimensions, the cube root of the atomic
volume V may be substituted for the distances between the particles, and Pictet's rule
assumes the form aTmV^^a constant. This rule applies very well for the heavier metals
where the mean value of the constant lies between 4 and 5. The constant for aluminium
is 6-6 : for magnesium, 5'7 ; tin, 28 ; antimony, 2'2 ; and bismuth, 2-05, although there
is some uncertainty as to the exact values of the melting points and the coefficients of
expansion of some of the metals. It has been proposed to use Pictet's rule for the calcu-
lation of melting points of elements. W. Sutherland, however, foimd the empirical rule
aT^M*=0-04 to 0-05 — say, 0-045— gives better results for all the metals tried, with the
exception of antimony, bismuth, and tin. The rule applies to sodium, magnesitun, and
aluminium — iridium gives a constant 0*037.
J. W. Richards ^ foimd empirically that the latent heats of fusion of about 15 elements
is one-third the amount of heat required to raise the metal from absolute zero to its melting
point within an error of from 5 to 10 per cent. If the average atomic heat of an element
between —273° and its melting point is approximately 6-4, Dulong and Petit's rule, then, the
amount of heat in a gram-atom of the metal at its melting point is 6'4Tjn. Assuming that
the latent heat of fusion is one-third the total heat in the metal at its melting point, the
latent heat of fusion will be one- third of 6-4Tm ; and, from Pictet's rule, the latent heat of
fusion =: 9 -S/av* cals. per gram-atom. Richards found that this agreed well with all the
metals for which the necessary data were available with the exception of aluminium, e.g.
for copper 9-5/aV^ = 3006, and dividing by the atomic weight to get the amount per
kilogram, the latent heat of fusion is 46*2 — the observed value is 43*0.
H. Crompton's formula* is MA = r38T»i2'v nearly; here M denotes the molecular
weight of the solid ; Sv, the sum of the valency bonds ; A, the latent heat of fusion of
unit weight of the substance ; and Tm is the absolute temperature of fusion. H. Crompton's
formula also gives a constant ranging between I'S and 2-56 for a nimiber of organic com-
pounds, but for normal or non-associated liquids he obtained satisfactory results with
AZ)=0-099Tm— the variations from the mean value of the constant 0*099 ranged from
0-080 to 0-142. He has also shown that for a series of metals the atomic heat of fusion
divided by the absolute temperature of the melting point, Tm, is almost independent
of the nature of the metal, for the quotient lies between 1 -84 and 4-82. P. Walden obtained
similar result's by using the molecular heat of fusion of 33 compounds ; here the quotient
fell between 12-5 and 148 with a mean value of 13-5. If the compoimds are associated in
the liquid state, the constant assumes a lower value. G. Tammann, however, examined
a far larger number of compounds than P. Walden, and only about a quarter fell between
the limits indicated by P. Walden. P. W. Robertson,^ also starting from Pictet's rule,
obtained the expression wXlTm%/v = & constant, which gave better results than
H. Crompton's equation for all the elements with atomic weights over 40 for which data are
available. The constant ranges from 0-87 to 1-28 — ^gallium, bismuth, and bromine gives
values respectively 2*05, 1-75, 1-63. The formula also applies to compounds when tu
denotes the molecular weight.
E. Obach ^ foimd a relation between the specific inductive capacity, K, and the latent
heat of evaporation A of a number of related organic liquids, such that A/iC=a constant.
From Trouton's rule, MA/Tj = a constant, where Tj, denotes the absolute boiling point,
it follows that MK/Tf, is also a constant, and accordingly, the absolute boiling point of a
series of related bodies is proportional to the molecular inductive capacity. P. de Heen found
the absolute boiling point at 760 mm., and the coefficient of expansion a at 0° are inversely
proportional, or the product aTj is a constant. Hence also oKM must be constant, or
the coefficient of expansion at 0° is inversely proportional to the molecular inductive
capacity. Accordingly, the product oKM represents P. de Heen's intermolecular work,
or that part of the heat spent in physical dissociation as contrasted with heat required for
chemical dissociation. The approximate proportionality between the latent heat of
vaporization and the dielectric constant of a liquid observed by E. Obach, ^ was extended
by P. Walden to a relation between the internal pressure and the dielectric constant.
W. C. McC. Lewis, however, foimd that there is no direct proportionality between the
cohesion or intrinsic pressure and the dielectric capacity. Although substances with a
large dielectric capacity have generally a large intrinsic pressure, the relationship is rather
more complex than is implied in E. Obach's or in P. Walden's rule. W. C. McC. Lewis
has also shown that the approximate relation indicated by E. Obach and P. Walden foUows
from the assumption that the cohesion or intermolecular attraction is an electromagnetic
and not an electrostatic effect. A. P. Mathews also postulated that, while the intramolecular
affinity uniting the atoms in a molecule is an electrostatic effect produced by valency
electrons, intermolecular cohesion is a magnetic effect produced by the valency and other
electrons.
According to H. Tomlinson, if w denotes the atomic weight of an element of
density D, and Cj^, the thermal capacity per unit mass, then the thermal capacity
836
INORGANIC AND THEORETICAL CHEMISTRY
C per unit volume is C=DCm. ', and by Dulong and Petit's rule, t<;C^= constant.
Since, as shown above, £'r7=constant, and r is proportional to {wID)'^, it follows
that EC~'^=& constant ; this means that the cube of Young's modulus varies as
the seventh power of the thermal capacity of the atoms per unit volume. It follows
also that the same constant is equal to Er"^.
According to A. H. Stuart,^ the relation between the coefficient of longi-
tudinal elasticity E ; the density D ; the specific heat Cp ; and the coefficient
of thermal expansion a, is Cp^l'iiaEID ; and if Dulong and Petit's rule be
CpW=Q'26, the atomic weight of an element is equal to \i\BjaE. The applica-
tion to the few metals for which reliable data are available is indicated in Table
XXVIII.
Table XXVIII.
Metal.
D
E
kilograms
per sq. mm.
axlO«
Atomic weight.
Calculated.
Observed.
Aluminium .
2-6
6,710
2-313
23-7
27-1
Copper
8-9
12,140
1-666
63-4
63-57
Gold .
19-3
9,650
1-443
196-0
197-2
Iron .
7-9
18,500
1-210
50-0
55-84
Platinum
21-5
17,044
0-902
197-0
195-2
Silver .
10-5
7,141
1-921
108-0
107-88
Tin .
7-3
4,170
2-234
111-0
119-0
J. D. van der Waals' equation of state for solids. — E. H. McCrea (1907) ^ tried
to adapt J. D. van der Waals' equation to the solid state, but with no definite
result. K. Eisenmann deduced
^ kT_'i
-1
as the equation of state of solids with spherical atoms — here h is Boltzmann's
constant. E. Kohl (1913) deduced a formula for the two specific heats : Cp—C^
=3aMXDI{Dg—Di), where a denotes the coefficient of linear expansion ; M, the
molecular weight ; A, the latent heat of fusion ; D, the specific gravity ; and
Dg and Di, the respective specific gravities of solid and liquid at the melting point.
It is assumed that J. D. van der Waals' equation is applicable to the solid, and that
the variation of the energy with temperature is independent of the state of aggre-
gation. The results are not good.
I. Traube also applied J. D. van der Waals' equation to highly compressed
gases and liquids by assuming that the observed volume is made up of two compo-
nents : (i) the volume of the molecules ; and (ii) the co-volumes or intermolecular
spaces. I. Traube calculated the J. D. van der Waals' constants a and h of the
metals by introducing two values for the volume v at two temperatures near 0°.
The one value of v was obtained by making v=wlD, where w denotes the atomic
weight, and D the density, and the other value of v was calculated from the co-
efficient of cubical expansion 3a. The values of a and b can be computed by
substituting these values of v and T in J. D. van der Waals' equation a{v—h)lv^
=RT, when the external pressure is negligibly small in comparison with ajv^,
I. Traube' s results are shown in Table XXIX.
THE KINETIC THEORY OF ATOMS AND MOLECULES 837
Table XXIX. — Values of J. D. van der Waals' a and b for Solid Elements.
Elements.
-^
b
a
litre atm.
^=^.atm.
3a
-.4,
Potassium
45-00
42-230
17-3
8,080
0-0002490
0-0040
Sodium
23-58
22-360
10-2
18,300
0-0002160
0-0042
Lead .
18-20
17-760
18-1
50,800
0-0000882
0-0036
Thallium
17-22
16-810
16-2
54,600
0-0000924
0-0039
Magnesium .
14-00
13-690
14-2
72,200
0-0000819
0-0037
Cadm.ium
13-02
12-710
12-3
72,200
0-0000930
00039
Tin . . .
16-23
15-900
17-9
67,800
0-0000675
00033
Aluminium .
10-50
10-310
13-1
117,800
0-0000696
0-0038
Mercury (liquid)
14-66
14-010
7-4
34,400
0-0001810
0 0041
Copper .
7-13
7-034
11-9
233,100
0-0000504
0-0037
Silver .
10-25
10-072
13-2
125,700
0-0000576
0-0033
Gold .
10-21
10-083
19-7
176,200
0-0000435
0-0035
Platinum
9-30
9-200
19-4
320,000
0-0000270
0 0035
Palladium
8-51
8-464
35-2
22,380
00000354
0 0033
Osmium
7-12
7-050
16-2
486,600
0-0000197
00036
Iron
6-60
6-520
12-2
319,700
0-0000366
0-0037
Nickel .
21-27
21-040
44-1
97,300
0-0000384
0-0032
Bismuth
17-88
17-690
37-7
117,700
0-0000396
0-0037
Antimony
23-09
13-020
54-9
319,700
0 0000345
0-0033
Arsenic
19-84
19-310
16-65
42,200
00000174
0-0039
Tellurium (cryst.) .
16-83
19-310
19-1
35,500
00001032
0-0039
Selenium (cryst.) .
16-83
16-200
10-0
41,400
0-0001480
0-0039
Sulphur
15-50
14-960
4-6
41,400
00001370
0-0046
Phosphorus .
16-89
15-500
41-5
16,200
0-0003750
0-0046
Silicon
12-91
12-820
41-5
248,700
00000231
0-0033
Diamond
3-41
3-406
63-6
545,800
0-0000375
0-0031
The numbers in the last column are remarkably concordant when the heterogeneous
character of the data is taken into consideration ; the average of these numbers
approximates to ■^^. The value for carbon is exceptional. The values of K=alv^
are about one-third those obtained by T. W. Kichards ^ on the assumption
that the stress, P, under which a substance rests can be measured by the heat C
absorbed per gram-molecule when the volume changes dvjdt. Whence P=CdTldv.
Accordingly, K^^^CdTjdv ; but if K=alv^, it follows from van der Waals'
equation a{v-b)/v^=RT, or K(v—h)=RT, and that
273
1
dv
c
1
dv
1
v-b
dT
"SRT'
Ui,
v-b
dT
"273
.r^-^rrol or, ^^^
which says in words : at a given temperature T, the coefficient of expansion of the co-
volume of a solid element, or the change per unit volume of the co-volume, is in general
constant, viz. gyg. This rule applies to the metals (monatomic) and the metalloids
(polyatomic), but not to the halogens. It also follows that the atomic volume is
nearly ZR cals. at 0°, when T=273°.
The relation between the latent heat of fusion and the coefficient of expansion.—
Again, the greater the coefficient of expansion of a metal the lower the melting
point, 10 although the results may be disturbed by changes in the complexity of the
molecules near the melting point. The rule is illustrated by data in Table XXIX.
It has been shown that for the solid metals J. D. van der Waals' equation, assumes
the forms a{v—b)jv'^=RT, and dvl(v—b)dT^^~^', substituting the latter value of
v—b in the former, it follows that at 0°, when T=273,
vWdl)"^
where the term in brackets represents the change in volume which occurs per unit
838
INORGANIC AND THEORETICAL CHEMISTRY
volume when the temperature alters 1° — i.e. the coefficient of cubical expansion 3a ;
and the term a/v represents the internal heat of vaporization, i.e. the observed
latent heat of vaporization less the work done in expansion against atmospheric
pressure during the change of state, and
Internal latent heat, A= - 4-RT ; or, - =X—RT
V V
Consequently, from the preceding expression, 3{y—RT)a=R, or the product of
the internal molecular heat of vaporization, and the coefficient of cubical expansion
is equal to the constant 12— this is illustrated in Table XXIX., where the product
3a(A— i^T) is approximately constant for the elements which follow Dulong and
Petit's rule. H. F. Wiebe ii assumed that the coefficient of cubical expansion
3a is proportional to the amount of heat required to raise an element of atomic
weight w from its melting point T^ to its boiling point Tft. If G denotes the specific
heat of the element, the product C'w(J'6—T^)= constant, which is nearly 2-03.
Again, assuming that the amount of heat required to raise the temperature from
absolute zero to the melting point T^, — or the thermal energy of the atoms of an
element just before it melts — is almost inversely proportional to its coefficient of
Table XXX.— Relation
BETWEEN Coefficient of Expansion and Internal Heat
OF Vaporization
Element.
Coeff. cubical
expansion 3a.
Melting
point.
Boiling
point.
A-/2r
Za(\^RT)
Phosphorus
0-000375
317
_
6,600
2-5
Potassium .
0-000249
328
960
9,200
2-3
Sodium
0-000216
363
1090
10,500
2-3
Selenium
0-000148
490
. —
14,550
2-1
Siilphur
0-000137
388
. — ,
15,650
21
Tellurium
0-000103
760
. —
20,350
2-1
Cadmium .
0-0000930
594
1110
22,900
2-1
Thallium
0-0000924
563
1110
22,800
21
Lead
0-0000882
601
1170
24,100
21
Magnesium
0-0000819
905
1190
24,600
2-0
Aluminium
0-0000696
929
1470
30,300
21
Tin .
0-0000675
505
1300
26,800
1-8
Silver
0-0000576
1234
1510
31,200
1-8
Copper
0-0000504
1355
1970
40,500
20
Gold
0-0000403
1337
2270
46,800
2-0
Bismuth
0-0000396
540
. — ,
50,300
20
Nickel
0-0000384
, —
2170
44,800
1-7
Iron
0-0000366
. —
2680
55,200
2-0
Palladium
0-0000354
.
2460
50,600
1-8
Antimony
0 0000345
710
. —
51,100
1-8
Platinum
0-0000270
. —
3670
75,500
20
Graphite
0-0000237
—
. —
76,700
1-8
Silicon
0-0000231
. —
. —
78,000
1-8
Osmium
0-0000197
. —
4900
100,600
2-0
Arsenic
0-0000174
—
. —
101,700
1-8
Diamond
0-00000375
■ —
—
45,200
1-7
expansion, a, so that the product of these two constants are approximately
constant ; and wCTfnCL=a. constant. Both rules apply to some families of elements,
but there are many exceptions.
W. Spring ^^ found that for a number of related isomorphous elements — iron and
aluminium ; cobalt and nickel ; sulphur, selenium, and tellurium^ — the product of the
increase in volume which occurs on heating the element from 0° to 100° into the atomic
weight is a constant— e.g. for sulphur, 0'035408 x 32=:l-2330 ; selenium, 0-017610x78
= 1-3657 ; tellurium, 0-010634 X 127 = 1 -3505. P. de Heen also has shown that the product
THE KINETIC THEORY OF ATOMS AND MOLECULES 839
of the coefficient of expansion and the absolute temperature of fusion is a constant or a
multiple of this constant for compounds which are chemically related, and crystallize in
the same system.
The above relations are either wholly empirical, or else they are based upon
quite arbitrary or empirical assumptions as to the constitution of matter. Some
of them may at times be useful in making approximations when observational data
are wanting. They are also instructive in revealing the probability of the near
discovery of a great law which will co-relate all the so-called physical constants of
the metals, and include these empiricisms as special cases of restricted application.
There is another side to this. By the substitution of one or more of these various
formulae in others, it is possible to obtain an indefinitely large number of equations
correlating various physical properties. The changes can also be rung with the
different forms of the gas equation. Consequently, a game of permutations and
combinations can be played in physical chemistry just as in organic chemistry a
similar game is possible with the methyl, ethyl, propyl, . . . radicles. C. F.
Schonbein i^ said :
I regard the discovery of thousands and thousands of new organic compounds in the
same light as I do the infinite number of figures which may be produced by the kaleidoscope.
What would the world say of a man who should take the trouble to shake for whole years
that plaything, and describe minutely all the shapes (pretty as they might be) he had
obtained from his operation ?
Eeferences.
1 F. Richarz, Zeit. anorg. Chem., 58. 356, 1908 ; 59. 146, 1908 ; G. Mie, Ann. Physik, (4),
11. 657, 1903 ; 0. Sackur, ib., (4), 34. 465, 1911 ; E. Grtineisen, ib., (4), 39. 257, 1912 ; (4), 26.
211, 393, 1908 ; (4), 33. 33, 65, 1910 ; Phys. Zeit, 12. 1023, 1911 ; Verh. deut. phys. Ges., 13.
426, 491, 836, 1911 ; 14. 322, 1912 ; H. Alterthum, ib., 15. 65, 1913 ; H. Polanyi, ib., 15. 156, 1913 ;
R. Ortway, ib., 15. 773, 1913 ; M. Thiesen, ib., 10. 410, 604, 947, 1908 ; I. Traube, Phys. Zeit.,
11. 231, 1911 ; S. Ratnowsky, ib., 14. 269, 1912 ; 16. 41, 1914 ; Ann. Physik, (4), 38. 637, 1912 ;
W. Sutherland, Phil. Mag., (5), 30. 318, 1890 ; W. Nernst, The Theory of the Solid State, London,
1914 ; C. M. Guldberg, Forh. Vid. Sels. Christiana, 140, 159, 1867 ; 15, 1868 ; Zeit. phys. Chem.,
16. 1, 1895; 32. 116, 1900; OstwaWs Klassiker, 139, 1903; A. C. Brown, B. A. Rep.,
45, 1874.
2 R. Pictet, Coinpt. Rend., 88. 855, 1879 ; W. Sutherland, Phil. Mag., (5), 30. 318, 1890 ;
W. Herz, Zeit. anorg. Chem., 89. 397, 1914.
3 J. W. Richards, Journ. Franklin Inst., 143. 379, 1897.
4 H. Crompton, Journ. Chem. Soc, 67. 315, 1895 ; 71. 925, 1897 ; Chem. News, 58. 237,
1903 ; P. Walden, Zeit. Elektrochem., 14. 713, 1908 ; G. Tammann, Zeit. phys. Chem., 85. 273,
1913 ; N. Deerr, Chem. News, 76. 234, 1897.
5 P. W. Robertson, Journ. Chem. Soc, 91. 1233, 1902.
« E. Obach, Phil. Mag., (5), 32. 113, 1891 ; H. Davies, ib., (6), 24. 415, 1912 ; E. B. Ro8a,i6.,
(5), 31. 188, 1891 ; W. C. McC. Lewis, ib., (6), 28. 104, 1914 ; A. P. Mathews, Journ. Phys. Chem.,
17. 481, 1913 ; P. Walden, Zeit. phys. Chem., 66. 407, 1909 ; S.Tereschin, Wied. Ann., 36. 792,
1889 ; R. Schiff, Liebig's Ann., 234. 338, 1886.
7 A. H. Stuart, Proc. Inst. Mech. Eng., 1155, 1912.
8 R. H. McCrea, Chem. News, 95. 101, 1907 ; E. Kohl, Chem. ZerUr., ii, 742, 1913 ; I. Traube,
Ann. Physik, (3), 61. 380, 1897 ; (4), 5. 548, 1901 ; (4), 8. 267, 1902 ; Zeit. anorg. Chem., 34.
412, 1903 ; 40. 372, 1904 ; Ber., 42. 86, 1909 ; Ber. deut. phys. Ges., 11. 231, 1909 ; C. Benedicks,
Zeit. anorg. Chem., 47. 455, 1905; K. Eisenmann, Ann. Physik, (4), 39. 1165, 1913.
9 T. W. Richards, Proc. Amer. Acad., 30. 3, 1901 ; Zeit. phys. Chem., 40. 169, 1902.
10 T. Carnelley, Ber., 11. 2289, 1878 ; I. Traube, Zeit. anorg. Chem., 34. 413, 1903.
11 H. F. Wiebe, Ber., 12. 788, 1879 ; Chem. News, 40. 154, 1879 ; T. Carnelley, Journ. Chem.
Soc, Z5. 565, 1879.
'2 W. Spring, Bull. Acad. Roy. Belgique, (3), 2. 33, 1881 ; P. de Heen, Recherches toucJiani la
physique comparee et la theorie des liquides, Paris, 193, 1888.
13 G. W. A. Kahlbaura and F. V. Darbishire, The Letters of Faraday and ScTioenhein, London,
206. 1899.
840 INORGANIC AND THEORETICAL CHEMISTRY
§ 21. The Kinetic Theory of Liquids
It must be remembered that in every dynamical investigation, what the mathematician
really investigates is not the problem presented by nature, but some simplification of it. . . .
For any pxu^ose which is of use to man, the approximation arrived at by the simpler
problem is suflScient, wherever the errors are of such a nature that they are not cum\ilative.
Nevertheless, it shoxild be clearly recognized that it is a mechanism illustrating nature,
and not nature itself, that has been mathematically investigated. — G. J. Stoney (1895).
The molecules of a liquid seem to have less freedom for movement than gases,
although the molecules are sufficiently mobile to allow the liquid to take up, more
or less quickly, the shape of the containing vessel. The mobility of liquids is
indeed their most obvious quality. It is very probable that the kinetic energy of
the translatory motions of the molecules of a liquid and its vapour, at the same
temperature, are equal.i A molecule can, in time, travel to any part of the liquid
mass. The rate of diffusion of one liquid in another shows that the movements
are rather slow, probably because a molecule in its travels must be continually
abutting against other molecules. The number of molecules in a gram-molecule
of a compound is approximately 6x1023; and for a non-associated liquid, say
fluobenzene with a gram-molecular volume of 91*7 c.c, the volume of the molecular
domain is 1 "5 X 10 ~ 22 qq Consequently, the average distance between the centres of
adjacent molecules is of the order 7 X 10~® cm., which is the same order of magnitude
as the diameter of the molecules. Further, while the intermolecular attraction
with gases is relatively small, the attraction between the molecules of a liquid must
be greater because of the closer packing. Again, approximately 1700 c.c. of steam
at 100° condenses to one cubic centimetre of liquid water at the same temperature,
so that the average distance between the molecules of the liquid must approximate
at least to the cube root of 1700, that is, about one-twelfth of the average distance
between the molecules of the gas. The slow evaporation of liquids also shows
that the molecules possess a certain mobility, and that the velocities of the moving
molecules are not all the same.
The resistance offered by liquids to compression is very great ; a fluid can
support a stress only when the stress is uniform in all directions. For example,
one c.c. of water at about 10° is reduced 0*000048 c.c. per atmosphere, or each
additional atmosphere pressure brings the particles >v^ 0*000048 =0-036, i.e. about
one-thirtieth nearer to one another, provided the volume of the molecules remains
constant. The fact that a liquid readily changes its shape but strongly resists
any force tending to diminish its volume, is taken to indicate that the potential
energy depends only on the mean distance between the molecules, and not on their
configuration. The small variation in the volume of liquid with increasing pressure
has led to the idea that the molecules are \ery close together. On the other hand,
this result, at first sight, is not parallel with the effect of temperature on the volume
of a liquid, for water at 10° will contract nearly six times as much in cooling to 4°.
P. S. de Laplace (1806) - assumed that the molecules of a liquid attract one another
with a force which extends over a very short distance which he called the radius
or sphere of molecular attraction. The range of this action has been shown to be
nearly 5 XlO"^ cm., for many experimenters have found that cohesion is inappreci-
able at greater distances. It is assumed that if it were not for the relatively great
intermolecular attractive forces, the molecules of a liquid would travel in approxi-
mately straight lines ; as it is, they are supposed to describe curved paths.
In any given liquid, the attractive forces tending to draw the molecules closer
together are balanced against the centrifugal forces or the tendency of the molecules,
so to speak, to move tangentially outwards away from the curved path. This
tendency to move outwards is proportional to the square of the mean velocity F
of the molecules, and inversely as the radius of curvature r of the path ; for, with a
molecule of mass M moving on a circular path, the centrifugal force is equivalent
to MV^/r ; and, if the attractive force for unit distance be F, then, for equilibrium.
THE KINETIC THEORY OF ATOMS AND MOLECULES 841
MV^Ir=Fr. The radius of curvature is probably very small and of molecular
dimensions ; accordingly, the intermolecular attraction R must be proportionally
great. If the speed of the molecules be reduced by cooling, the centrifugal tendency
is lessened, the attractive forces predominate, and the molecules move closer to-
gether into a new position of equilibrium. Accordingly, the liquid contracts in
volume ; and it also becomes more viscous owing to the greater difficulty experienced
by the molecules in moving away from the sphere of one another's influence against
intermolecular attraction. Isothermal compression also brings the molecules
closer together — this favours the inward attractive forces ; but the kinetic energy
of the molecule is at the same time augmented — this increases the speed of the
molecules, and strengthens the centrifugal forces which make the liquid resist
further compression.
The cohesive attraction of the molecules of liquids. — The resultant of the mutual
attraction of molecules simulates an inward pressure which opposes a resistance
to the forces tending to enlarge its volume. Consider two layers of molecules in
proximity to one another ; the resultant attraction of molecule for molecule will be
proportional to the number of attracted and attracting molecules, that is, to the
square of the number of molecules. The number of molecules in unit volume of a
liquid or solid will be proportional to the density, D, so that if a be the constant of
proportion, the intermolecular attraction, the internal pressure, the cohesive
pressure, or the intrinsic pressure, P — as Lord Rayleigh ^ called it — will be
P=aD^ ; and if v denotes the volume occupied by one gram of the substance,
P=alv^, which is the term employed for intermolecular attraction in J. D. van der
Waals' equation. The relation av~^ also follows directly from the assumption that
the molecular attraction varies inversely as the fourth power of the distance of the
molecules apart. The magnitude of the intrinsic pressure for water, determined
by two independent methods, is very great — approximately 11,000 atm. ; carbon
disulphide, 2900 atm. ; ethyl alcohol, 2400 atm. ; and for ethyl ether, 1300 atm.
P. S. de Laplace ^ was surprised at the magnitude of his own estimate of this pressure,
for be obtained what he called une aussi prodigueuse value qui ne peut pas Hre admise
avec vraisemblance — approximately 10^2 tons per sq. in. The intrinsic pressure is a
measure of the cohesion of a liquid or solid, that is, it measures the force — tensile
strength — required to separate one portion of a liquid from another. Laplace's
constant K is equivalent in meaning to J. D. van der Waals' av~^ ; and T. Young
estimated that the surface tension ex of a liquid is equal to one-third the total
cohesive force, P, into the radius r of the molecule, or G=^^rP.
The intrinsic pressure of liquids. — The intrinsic pressures of a liquid calculated
from J. D. van der Waals' equation and from A. Dupre's relation between the internal
latent heat of vaporization are somewhat discrepant. The former gives 10,500 to
11,000 atm., the latter 23,900 atm. The discrepancy has been attributed (i) to the
invalidity of the assumption that the density of the liquid in the surface layer is
the same as in the bulk of the liquid, and G. Bakker therefore assumes that there
is a gradual decrease in density in passing from the liquid to the vapour phase ;
and (ii) to the assumption that the molecular forces have no temperature effect,
whereas it is certain that the intrinsic pressure P decreases as the temperature
rises, and W. C. McC. Lewis ^ shows that the relation between the intrinsic P pressure
and temperature T probably has the form P=a-\-h log T-\-cT, where a, b, and c
are constants. The different expressions which have been suggested for evaluating
the intrinsic pressure have been compiled by W. C. McC. Lewis.
In 1869, A. Dupre suggested that the intrinsic pressure is equal to the internal
work done per unit volume of liquid against the molecular attracting forces when
the molecules are drawn apart until they are outside the range of their mutual
attractions. Hence, A. Dupre assumed that the work done against intrinsic pressure,
P, is equal to the internal latent heat per unit volume A, or P=X ; but
W. C. McC. Lewis emphasized that this can be true only when P is independent of
temperature, and that the true relation probably has the form of H. von
842
INORGANIC AND THEORETICAL CHEMISTRY
Helmholtz's free energy equation P=X-\-T(dj>jdT)^ ; andH. Davies has shown that
dPjdT is equal to —^Rjvc ; to —Rjh ; or to — 14*8^c/^cj which enables P to be
calculated from observed values of A and the critical constants.
A. Dupre showed that if ^ denotes the coefficient of compressibility, and a the
coefficient of cubical expansion, J. D. van der Waals' equation furnishes the relation
P=—TalP, on the assumption that P is independent of the temperature. The
original edition of J. D. van der Waals' thesis also showed that if v be the volume of
a given mass of vapour, Vi the volume of the same mass of liquid, a the coefficient
of gaseous expansion, and j3 the compressibility of the liquid, then, at 6°,
P^=v{l-^ad) /viP, and I. Traube has applied this result to liquid and solid elements.
T. W. Richards assumed that the heat C required to raise the temperature of
unit volume by the amount dT, is equivalent to the intrinsic pressure P such
that P=CdTldv, on the assumption that P is independent of temperature and
volume. C is here the atomic heat — approximately 6. The numerical values
obtained with the elements are very large. I. Traube showed that T. W. Richards'
expression gives very nearly three times the values obtained by J. D. van der Waals'
expression, and he therefore writes P=\CdTldv. C. Benedicks also deduced the
expression P=av~'^ of J. D. van der Waals' equation, and assuming that a is inde-
pendent of temperature and a denotes the coefficient of cubical expansion, he found
P=R{a-'^-\-2T)v-^, or approximately P=Rlva, which is virtually equivalent
to T. W. Richards' expression only if C were put equal to R. R. H. Davies obtained
the expression P=4:RlaVc, where Vc denotes the critical volume, and a the
coefficient of cubical expansion, which H. Davies found to be equal to {^Tc—T)—!.
At absolute zero, this reduces to P=SRTclvc.
J. D. van der Waals' equation may be written {j)-\-P)(v—h)=RT, where p
denotes the external and P the internal pressures, and h the limiting molecular
volume at the absolute zero. Accordingly, if the small pressure p be neglected
in comparison with the large pressure P, it follows that the intrinsic pressure,
P=RTI(v—h). C. M. Guldberg extrapolated Cailletet andMathias' linear diameter
to absolute zero, and found the densities of a number of substances at absolute zero
and at the critical temperature, the ratio being practically a constant 3*75 or
h=0'27vc, for substances. D. Berthelot also found for oxygen, chlorine, carbon
dioxide and sulphur dioxide, carbon tetrachloride, and benzene, h=0'2Qvc. The
results furnished by calculating the numerical value of P=RTI(v—b) at 0° on
these assumptions represent minimum values of P, because the values of h are not
likely to be independent of temperature, since 6 probably increases as the tempera-
ture rises. Taking P. Walden's or S. Young's values of Vc, the molecular critical
volume, h=0'27vc ; and v the molecular volume at 0°, the intrinsic pressures P
at 0° of 36 liquids were calculated by W. C. McC. Lewis in atm. per sq. cm.
Table XXXI. — ^Intrinsic Pressures of Liquids.
^c
b
V
v-b
P
Water ....
50-3
13-58
18-00
4-42
5084
Methyl alcohol
117-6
31-75
39-5
7-75
2898
Ethyl alcohol
166-9
45-06
56-1
11-04
2035
Carbon disulphide .
166-0
44-82
58-78
13-98
1610
Benzene
256-1
67-15
86-7
20-55
1093
Carbon tetrachloride
276-2
74-57
94-3
19-73
1138
Ethyl ether .
281-9
76-11
100-5
34-39
1653
Stannic chloride
351-8
94-99
114-5
19-51
1135
The relation between intrinsic pressure and surface tension.— The conception
of an internal, intrinsic, molecular, or cohesive pressure presupposes the existence
of attractive forces between the molecules diminishing rapidly with distance.
THE KINETIC THEORY OF ATOMS AND MOLECULES
843
G. N. Antonoff has deduced a relation between the surface tension and the internal
pressure on the assumption that the attractive force between the molecules is
electro-magnetic in nature, and that the molecules can be treated as if they were
electro-magnetic doublets. If P denotes the molecular or internal pressure ;
a, the surface tension ; w, the number of doublets or molecules in unit volume
of liquid ; D, the specific gravity of the liquid ; M, its molecular weight, and
1-64x10-24, the weight of an atom of hydrogen, then w=l'64xl024D/lf,
and therefore P=2crw*. Hence, if the surface tension of benzene be 32 dynes
per cm. ; the molecular weight, 78 ; and the specific gravity, 0'890 ; then
yi=6'8xl02i, and P=12xl06 dynes per sq. cm., a magnitude very nearly equal
to 1200 atm.
Since the surface tensions and intrinsic pressures of liquids are closely related,
it follows that when the exact connection between the various physical properties
of the one is known, their relations with the other follow as a matter of course.
S. W. Smith 6 assumed that the surface tension a of some liquid metals is related
with the intrinsic pressure P by the empirical formula : cr=20O4-0'0O437P, when
a is expressed in dynes per cm., and the intrinsic pressure P in megabars —
106 dynes — ^per sq, cm. P. Walden found for a number of organic liquids P=l-82or,
when the surface tension is determined at the boiling point. Otherwise expressed,
the intrinsic pressure of a liquid at its boiling point is proportional to the surface
tension at the boiling point. The relation between the surface tensions or intrinsic
pressure and some other physical properties of liquids is shown in S. W. Smith's
Table XXXII. The elastic moduli and the other properties run parallel with the
surface tensions.
Table XXXII.- — Relations between the Physical Properties of Solids and
Surface Tension.
Surface
Internal
Compressi-
Atomic con-
tension (fluid)
dynes per cm.
pressure
(solid)
megabars,
P
bility per
atomic vol.
per megabar,
Elastic
modulus.
centration,
sp. gr. -4-
atomic wt.
Hardness.
Lead
424
51,500
40-04
1,800
0-0550
1-0
Tin .
480
68,700
25-92
5,000
0-0619
2-5
Aluminium
520
119,300
13-13
7,200
0-0821
. —
Zinc .
707
108,900
13-70
. —
01077
6-0
Silver
858
161,900
8-65
7,300
00963
—
Gold
1018
178,500
4-79
8,100
0-0979
. —
Copper
1178
236,100
3-83
12,400
0-1364
8-0
Iron .
(1244)
239,000
2-84
20,900
0-1375
15-0
Nickel
(1538)
306,300
1-76
—
0-1408
—
The relation between intrinsic pressure and latent heat.— J. E. Mills (1902) 7
assumes that the resultant attraction between two molecules varies as the square
of the distance of the molecules apart, and is a mutual property of each pair of
molecules, so that the resultant attractive force F=k'm^jr^, where r denotes the
distance apart of the two molecules — each of mass m ; and k is the attraction
constant whose numerical value is dependent on the nature of each of the attracting
molecules. Consequently, the assumed law of molecular attraction differs from the
attraction of gravitation in being dependent on the nature of the molecules and
not solely dependent on their mass. If L denotes the observed heat of vaporization ;
E, the work of expansion against external pressure ; Di, the density of the liquid ; and
D that of the vapour, J. E. Mills finds that
L-E
\-^D
Constant
844
INORGANIC AND THEORETICAL CHEMISTRY
If a mass of liquid of volume v has n particles at a distance r apart ; and if V denotes
the volume of the same mass in the gaseous state when the molecules are a distance R apart,
the work W expanded in tearing the molecules asunder, against the assumed intermolecular
attraction, during vaporization will be
j F.dr=km^j ^; or, W=km^Q,-^^
■R It
Again, if L denotes the observed heat of vaporization, and E represents the work spent in
overcoming the external pressure of the atmosphere during vaporization, the work W
required to pull apart the n molecules of mass m will be W=nm{L—E). Note that v and
V respectively denote the volumes of the liquid and vapour, each containing n molecules
of mass m, r = %/v/n, and B. = \/y/n ; and that the density D^ of the liquid is D^ = nmjv,
and of the vapour, D=nm/V ; substituting for i), and D, in the two preceding values of
W ; and remembering that k, m, and n are constants. Mills' expression follows at once.
The work E spent in overcoming the external pressure by altering the distance
apart of the molecules is calculated from the equation ^=0'0000318^(7— v) cais.,
where p denotes the pressure of the air in millimetres of mercury, and V and v
the respective volumes before and after expansion. The values of the constant
for carbon disulphide at different temperatures selected from Mills' tables are
indicated in Table XXXIII.
Table XXXIII.- — Mills' Constant fob Carbon Bisulphide.
Temperature.
Density of Uquid,
^1
Density of vapour,
Latent heat,
L
External work,
E
Mills' constant.
0°
1-2921
0-000966
90-00
7-24
82-22
20°
1-2775
0-000350
89-06
7-48
82-38
40°
1-2321
0-0024
85-64
8-10
82-71
60°
1-2003
00044
82-87
8-42
82-85
80°
1-1684
0-0075
79-70
8-67
82-82
100°
1-1684
0-0120
76-14
8-85
82-61
120°
1-0997
0-0182
72-18
8-94
82-23
J. E. Mills also found that out of 435 observations on 26 different non-associated
liquids over ranges of temperature, only 30 differed from the mean value of the
constant for the specific substance in question by more than 2 per cent, and only
four of the 30 by more than 5 per cent. The data thus support the assumption
that the molecular attraction (i) is independent of the temperature ; (ii) varies
inversely as the square of the distance apart of the molecules ; and (iii) is a constant
for any particular substance.
Gr. N. Antonoff showed that the relation between the internal pressure and the
surface tension is the same whatever be the law of attraction between the molecules.
For many purposes, therefore, it is not necessary to specify the attractive forces
other than that they diminish rapidly with the distance. K. D. Kleeman 8 has also
shown that it is possible to obtain an infinite number of formulae for the surface
tension of a liquid, and each of these formulae corresponds to a law of molecular
attraction, so that any number of laws of attraction will give latent heat formulae
agreeing with facts. All these laws of attraction are but fragments of a general
law which must contain an arbitrary function of the distance between the attracting
molecules and their temperature. J. E. Mills argues that a change of temperature
does not change the nature or magnitude of molecular attraction, but rather
determines the orbit the molecules will follow in obedience to the attractive force.
In reply to R. D. Kleeman's proof that the molecular attraction must decrease at
a much greater rate with the distance of the molecules than is given by the law of
inverse squares, J. E. Mills postulates a mutual absorption or cancellation of the
whole or part of the attractive force when this attraction is exerted upon other
particles.
THE KINETIC THEORY OF ATOMS AND MOLECULES
845
W. Sutherland (1893) ^ used the hypothesis that the intermolecular attraction
varies inversely as the fourth power of the distance apart of the molecules ; R. D.
Kleeman (1910) examined a fifth power law ; A. Albertosi (1915) a sixth power
law ; and P, de Heen a seventh power law. The fifth power law proposed by
J. C. Maxwell in 1866 was abandoned by him in 1879. Newton's gravitational
law thus includes but one term in a possible series :
F=ar-^+br-^+cr-^+dr-5-\- . . .
where the coefficients h, c, dj . . . are so small that they can be neglected except
where r is itself small.
P. S. de Laplace's formula for the pressure P within a sphere of liquid of unit
radius is P=K-\-H, where Laplace's constant K involves the intrinsic pressure,
and the term H involves the surface tension. A high intrinsic pressure is accom-
panied by a large surface tension. The intrinsic pressure has been related with the
surface tension, the coefficient of thermal expansion, molecular volume, compressi-
bility, vapour pressure, viscosity, etc. If the attractive forces between the mole-
cules, or the intrinsic pressure, be very great, the coefficient of thermal expansion
as well as the compressibility and viscosity will be smaller than when the attractive
forces are small. Similarly also the latent heat of vaporization of a liquid will be
larger when the molecular attraction is great than when the molecular attraction
is small. Consequently, a large latent heat, a small compressibility, a small co-
efficient of thermal expansion, and low viscosity should run together. This is
illustrated qualitatively in Table XXXIV. An enormous number of changes have
Table XXXIV.— Some Effects of Molecular Attraction in Liquids.
Liquid.
Molecular latent
heat, Cals.
Coefficient thermal
expansion.
Compressibility.
Viscosity at 0°.
Mercury
12-4
0-00018
0-000004
0-0168
Water . ' .
11-1
0-00006
0-000053
0-0178
Alcohol .
9-6
0-00103
0-000112
0-0177
Carbon tetrachloride
7-9
0-00124
0-000125
00133
Chloroform
7-9
0-00127
0-000128
00071
Carbon disulphide .
6-8
0-00146
0-000174
0-0044
Ether
6-5
0-00166
0-000176
0-0029
Sulphur dioxide
6-2
0-00215
0-000303
been rung with formulae connecting the various physical properties of liquids. In
a general way, it may be said that when a relation has been found to obtain between
certain physical properties of a liquid over a certain range of temperature, it is almost
sure to be applicable to other liquids over a similar range of temperature. ^o This
probably also applies to the so-called associated liquids ; for the apparent failure with
these is mainly due to the use of incorrect molecular weights ; with corrected mole-
cular weights the associated liquids would also fall in line with other liquids. If the
physical properties are markedly constitutive, that is, are dependent upon the
inner structure of the molecule — the formulae are not general — e.g. viscosity, latent
heat, specific heat, surface tension. Of course every physical property is to some
extent constitutive, and small deviations are found, but in view of the sparsity of
accurate and comparable observational data, the general agreement just indicated
is rather striking.
Eeferences.
» J. E. Mills, Journ. Amer. Chem. Soc, 31. 1099, 1909.
2 P. S. de Laplace, Mecanique celeste, Paris, 10. 1, 1806 ; Suppl, Paris, 1807.
3 Lord Rayleigh, Phil. Mag., (5), 30. 285, 456, 1890 ; W. C. McC. Lewis, Tran.9. Faraday
Soc, 7. 94, 1911 ; Phil. Mag., (6), 28. 104, 1914.
846 INORGANIC AND THEORETICAL CHEMISTRY
* P. S. de Laplace, Mecanique celeste, Paris, 4. 389, 1805 ; J. C. Maxwell, Encyc. Brit., 5. 56,
1875 ; T. Young, Phil Trans., 95. 65, 1805.
' W. C. McC. Lewis, Kolloid Zeit., 7. 197, 1910 ; Phil. Mag., (6), 20. 602, 1910 ; (6), 22.
193, 1911 ; (6), 28. 104, 1914 ; H. Da vies, ib., (6), 24. 415, 1912 ; (6), 23. 657, 1912 ; W. Sukhod-
sky, ib., (6), 23. 955, 1912 ; G. Bakker, Ann. Physik, (4), 17. 475, 1905 ; A. Duprc, Theorie
mecanique de la chaleur, Paris, 1869 ; Ann. Chim. Phys., (4), 6. 283, 1865 ; J. 1). van der Waals,
Over de Continuiteit Dan den Gas- en Vloeistoftoestand, Leiden, 1873 — not in the London or Leipzig
editions; I. Traube, Ber. deut. phys. Ges., 11. 231, 1909: Zeit. anorg. Chem., 34. 413, 1903;
C. Benedicks, ib., 47. 455, 1905 ; C. M. Guldberg, Zeit. phys. Chem., 16. 1, 1898 ; B. Berthelot^
Compt. Rend., 130. 713, 1900; P. Walden, Zeit. phys. Chem., 66. 407, 1900; S. Young, Proc.
Boy. Soc. Dublin, 12. 374, 1910.
« S. W. Smith, Journ. Inst. Melah, 12. 168, 1914 ; 17. i, 65, 1917 ; O. N. Antonoff, Phil.
Mag., (6), 36. 377, 1918 ; (6), 38. 417, 1919.
' J. E. MiUs, Journ. Phys. Chem., 6. 209, 1902; 8. 383, 593, 1904; 9. 402, 1905; 10. 1,
1906; 11. 132, 594, 1907 ; 13. 512, 1909; 15. 417, 1911 ; 18. 101, 1914; 19. 257, 650, 1915;
21. 101, 345, 623, 1917 ; Phil. Mag., (6), 20. 629, 1910 ; (6), 22. 84, 1911 ; Journ. Amer. Chem.
Soc, 31. 1909, 1909 ; J. E. Mills and D. MacRae, ib., 32. 1162, 1910 ; A. P. Mathews, Journ.
Phys. Chem., 17. 520, 1913 ; D. Tyrer, Phil. Mag., f6), 23, 112, 1912.
« R. D. Kleeman, Phil. Mag., (6), 19. 783, 1910; (6), 21. 83, 1911 ; J. E. Mills, ih., (6),
22. 84, 1911 ; Journ. Phys. Chem., 15. 417, 1911 ; G. N. Antonoff, Phil. Mag., (6), 36. 377,
1918 ; (6), 38. 417, 1919.
» W. Sutherland, Phil. Mag., (5), 36. 507, 1893; (6), 17. 669, 1909; R. D. Kleeman, ib.,
(6), 19. 793, 1910: (6), 21. 83, 1911; (6), 21. 83, 1911; A. Albertosi, Journ. Chim. Phys.,
13. 379, 1915; J. C. Maxwell, Phil. Trans., 156. 249, 1866; H. Tomlinson, ib., 174. 1,
1883 ; P. de Heen, Bull. Acad. Roy. Belgique, (3), 4. 12, 1882 ; Recherches touchaiit la
physique comparee et la theorie des liquides, Paris, 2. 7, 1888 ; A. Wertheim, Ann. Chim. Phys.,
(3), 12. 285, 1844 ; E. Gruneisen, Ann. Phys., (4), 39. 257, 1912 ; H. Chatley, Proc. Phys. Soc,
27. 443, 1915.
1" D. Tyrer, Journ. Phys. Chem., 17. 717, 1913.
§ 22. The Surface Tension and Surface Energy of Liquids and Solids
Capillary force may be considered as the first degree of chemical affinity. — M. l'Hermite
(1855).
The molecular forces at the surface of a liquid do not produce the same effects
as in the interior. In the interior of a liquid, the molecules are attracted equally
in all directions, those at the surface are attracted inwards, in a direction perpen-
dicular to the free surface. The surface tension, of course, is not the cause but
rather a consequence of the internal pressure. The effect of this inward pressure
is in many respects analogous with what would obtain if the surface was enveloped
by an elastic membrane in a state of tension. The tension is called the surface
tension or surface pressure of the liquid. For equilibrium, the surface tension of a
given liquid is the same in all directions and at all points on the surface. There is
a difference between the tension of the imaginary membrane and the actual tension,
or inward pressure of the free surface of the liquid, in that when the surface of the
liquid increases, it does so by exposing fresh particles, and not by stretching the
old surface of the liquid in the sense that indiarubber would be stretched. The
elastic membrane is merely an analogy to assist the mind in forming a concept of a
number of different phenomena ; but it does not enable us to form a satisfactory
mental picture of the distribution of the molecules in the surface film of a liquid.
It is therefore convenient to regard liquids as if they were enclosed in a contractile
membrane in a state of uniform tension which makes the free surface the smallest
that circumstances will allow. The sphere has the smallest surface of any given
mass of liquid, consequently, drops of liquid assume the form of a sphere, unless
they are so large that gravitation can exert an appreciable influence, and flatten
the drop to a thickened disc. The smaller the drops the more nearly do they
approach perfect spheres.
ExAMPLES.^ — (i) Mercury globules on a flat board show the phenomena well, (ii) By
suspending a drop of, say, ortho-toluidine in a 3 per cent, solution of sodium chloride large
spherical drops 5 to 6 cm. diameter are easily obtained since the effect of gravity on the
THE KINETIC THEORY OF ATOMS AND MOLECULES 847
drop is cancelled, so to speak, because the drops are suspended in a liquid of the same
density they are themselves. Similarly, (iii) soap-bubbles in air are spherical, and they
would remain suspended therein and not sink if it were not for the actual weight of the
film of soap solution.^
In the case of liquid crystals, the surface tension of the crystals tends to make
them assume the form of spherical globules, but this is opposed by the special
molecular force — the Gestaltungskraft of 0. Lehmann — which makes normal crystals
assume their characteristic forms. Hence, photographs show that the liquid
crystals of some substances are spherical, while others have pyramidal forms with
more or less rounded edges.
The pressure in the space enclosed by a spherical liquid surface.— Imagine the liquid
sphere to be divided into two hemispheres by an imaginary plane. Let r be the radius
of the sphere ; p, the pressure at any point inside; and a, the surface tension. The only
forces acting on the hemisphere ABCD, Fig. 16, are (i) a thrust on the plane
face ABC, due to the pressure of the liquid in the half sphere not shown in the
diagram, and which is equal to the pressure p X the area of ABC, viz. ttt^ ; and
(ii) the tension of the surface acting round the edge of the circle ABC, and which
is equal to a times the perimeter ABC, that is, to <r.27rr. These two forces are
in equilibrium, and therefore balanced so that p7r7'^ = a2Trr. Hence, the pressure
p = 2(r/r, that is, the pressure is inversely proportional to the radiiis of the
sphere ; it is produced by surface tension only ; and is the excess of the internal
over the external pressure. If a soap-bubble be in question, there are two
surfaces, each of which exerts a pressure 2(r/r, so that the pressure within a soap-bubble
is 4(r/r. The same reasoning coupled with hydrostatic pressure can be applied to gas-
bubbles rising in a liquid.
A liquid boils when the bubbles of vapour formed in the interior can rise and
escape at the surface. The pressure of the vapour within a bubble at a depth h
below the liquid is equal to the pressure of the atmosphere plus the pressure equal to
the weight of a corresponding column of liquid of height h, and 2or/r. If the
pressure of the vapour within the bubble be less than this, the bubble must collapse.
The smaller the bubble, the smaller the value of r, and the greater the pressure of
the vapour. Hence the temperature of a liquid must be much higher to maintain
small than to maintain large bubbles in equilibrium. If there are no facilities for
producing bubbles which are fairly large to start with, the temperature may rise
above the boiling point until a large bubble is formed — this may give rise to bumping,
or portions of the liquid may be projected from the vessel, with explosive violence.
The presence of a small piece of capillary, closed at one end, and
filled with air, enables the vapour of the liquid to be formed in the
interior, and fairly large bubbles to be formed at the open ends.
This explains how the presence of these capillary tubes, porous
earthenware, charcoal, etc., enables liquids to boil steadily without
bumping.
Surface tension is evidence of the presence of cohesive forces
between the molecules of a liquid. J. C. Maxwell 2 illustrated surface ^jq 17, — sur-
tension by supposing a film of liquid S, Fig. 17, to be stretched face Tension
on a wire framework, and attached to a light wire AB, and a weight of Liquids.
W. Assume the part of the wire wetted by the film is of unit
length. The force exerted per unit length of the stretched film is termed the surface
tension. It is usually more convenient to consider not the actual surface tension
or force acting on the surface of the liquid, but rather the energy associated with
that surface, the so-called surface energy. If W just supports the stretched film,
the surface tension a of the film is equal to JTF, because there are two surfaces to
the film. If the weight suffices to stretch the film a distance s, the work performed
against surface tension will be the product 2as. The potential energy of the surface
of the film has therefore increased by an amount 2gs ; and each side of the film
has increased by an areas ; hence, the increase of energy per unit area is 2gsI2s—g.
Hence if W denotes the surface energy per unit area, W=(7. Consequently,
=^B
848
INOKGANIC AND THEORETICAL CHEMISTRY
the surface tension of a liquid is equal to the work done in enlarging the surface of
the liquid by one sq. cm., under isothermal conditions, although it is not permissible
to define surface tension as the energy per unit area of the surface of a liquid. The
surface tension in dynes per cm. is numerically equal to the surface energy expressed
in ergs per sq. cm. ; the surface tension may be expressed in dynes per cm. ; in grams
per cm. ; or in mgrm. per mm. In converting from one system of units to another,
the number of grams multiplied by 980 furnished the equivalent in dynes ; and mgrm.
per mm. multiplied by 9 8 changes the result into dynes per cm.
There are several more convenient methods of measuring surface tension. When a
capillary tube is plunged vertically in a liquid, the surface tension inside the tube ranges
over the liquid within the circumference of the capillus or internal bore, and the liquid rises,
A, Fig. 18, or falls, B, Fig. 18, below the level of the liquid outside the tube, according as the
liquid wets, A, Fig. 18, or does not wet, B, Fig. 18, the walls of the tube. The vertical
component of the surface tension between the liquid and the walls of the tube at the level
of the liquid within the tube acts along a length equivalent to the circumference of the internal
bore, i.e. 27rr ; and at an angle a — called the angle of
contact, or edge angle — between the normal to the free
surface, and the tangent to the liquid surface, where it
meets the solid walls. Consequently, the vertical com-
ponent of the surface tension at the point of contact
between liquid and solid is equivalent to 27rr X o- x cos a.
This is balanced by the weight of the column of liquid
within the tube which is above the level of the liquid
outside the tube ; or, if the liquid does not wet the tube,
by the column of liquid which is held back by the
surface tension. If h denotes the height of this column,
the volume of this liquid is hTrr^^ and if D be the density
of the liquid, and g the acceleration of gravity, the
weight of this column of liquid is hXnr^xDxg dynes.
When these two forces are balanced, or in equilibrium,
a x27rrxcosa=hx irr^xDxg or a^^ghrD/cos a dynes,
which represents the surface tension in terms of the height h of the capillary effect ; the
radius r of the tube ; the density D of the liquid ; and the constant angle a. When the
contact angle is nearly 180°, cos a approaches unity, and the surface tension can be
experimentally measured from the formula <r = ^ghrD. Sometimes, the relation rh or 2<t/D
is called the specific cohesion of a liquid, and it is symbolized by a^, such that a^=2(rlD=rh.
For example, if the density of water is unity, and the surface tension 7*25, the value of
a 2, the specific cohesion is 14 '50. The surface tension can also be measured from the
wave-lengths of ripples on clean surfaces ; ^ from the number of drops which fall from a
given orifice in a given time ; etc.
Numbers ranging from 7*13 to 7*945 mgrm. per mm. have been published for the
surface tension of water at 18° ; and from 39"23 to 55*78 for mercury at 20°. The
surface tensions of a few liquids in dynes per cm. are :
Fia, 18. — Contact Angle.
r-\
Mercury
Water
Carbon di-
Chloroform
Carbon tetra-
Alcohol
Ether
(18°).
(20°).
sulphide (20°).
(20°).
chloride (20°).
(20°).
(20°).
440
75
32-7
25-68
21-6
26-3
15-9
The kinetic theory of vaporization would lead to the assumption that the proper-
ties of a liquid surface must be radically different if it is in contact with another
liquid instead of its own vapour. At the boundary surface of the gas-liquid system
the mean free paths of the molecules are of a very different order of magnitude, while
at the surface of the liquid-liquid system, the mean free paths will be of the same
order of magnitude and characteristic of each liquid. M. Planck and M. Cantor *
believe that the surface of a liquid when in contact with another liquid retains the
same properties as it had when in contact with its own vapour. Two liquids of
limited solubility are in equilibrium when each liquid is saturated with the other,
and D. Konovaloff showed that the saturated vapours emitted by each of the two
layers of liquid have the same vapour pressure, and the same vapour composition
although their own composition is very different. G. N. Antonoff believes that the
surface tensions of the layers of the two liquids of limited miscibility are identical
in the two contact planes. Lord Rayleigh's hypothesis led to the conclusion that the
THE KINETIC THEORY OF ATOMS AND MOLECULES 849
surface tension ai2 at the limiting surface of the two liquids with the respective surface
tensions cri and o"2 when in contact with their own vapour is given by y'(Ti2= ^/cri
— n/0'2 ; ^'^^ Gr. N. Antonofi did not find this formula to agree with experiment, and
he has shown that the interfacial surface tension o"i2 is equal to the difference of the
surface tensions against air of the two superposed liquids in equilibrium such that
(Ti2=o'i— o^2> so that two superposed liquids, of limited solubility and in equilibrium,
must be regarded as solutions in the same solvent ; and he also believes that they
contain an equal number n of molecules per unit volume so that ni=n2, by
A. Avogardo's hypothesis, for, says G. N. Antonoff, " if two solutions in the same
solvent have the same freezing and boiling points, as is the case with two superposed
liquids of limited solubility and in equilibrium, they must contain an equal number
of molecules in unit volume." This is possible only when the molecules of the
one component B form a compound with some of the molecules in the solution, and
any further addition of B would not increase the number of molecules present in
the solution. Hence, all properties depending on the number of molecules and not
on their dimensions remain invariable.
The relation between surface tension and temperature. — According to D. I.
Mendeleefi,^ one characteristic of a perfect liquid is that its surface tension should
be a linear function of its temperature, and this view was confirmed by A. L. Selby.
The decrease of the surface tension with a rise of temperature is usually represented
by a simple linear equation, for in most cases the surface tension a at 6° is equal to
(TQ{l~-ad), where ctq denotes the surface tension at 0°, and a is a constant charac-
teristic of particular liquids. ^ Thus,
Water from
Ether from
Alcohol from
Cadmium from
Lead from
0° to 70°.
2° to 25°.
5° to 72°.
365° to 431°.
389° to 498'
0-0021
0-0060
0-0034
0-00042
0-00029
The ordinary linear equation a=(7o(l+«^) and the parabolic equation a-
=GQ{l-\-ad-\-bd^) hold only over a very restricted range of temperatures, but they
probably become very inaccurate when extrapolated far beyond the range of
observation.
According to J. D. van der Waals,*^ the surface tension cr of a liquid at a tempera-
ture T, not far removed from the critical temperature Tc, is
where A is a constant independent of the temperature. At the critical temperature,
both G and daJdT are zero.
Just as the kinetic energy of an ideal gas remains constant during an isothermal
expansion, because the external work performed during the expansion is derived
from the heat which enters the gas, so, when the surface of a liquid is stretched
isothermally, the energy of the new surface is not equal merely to the mechanical
work performed during the stretching, for an allowance must be made for the energy
which enters or leaves the surface in the form of heat. The mechanical work per-
formed when the surface of a liquid is increased is spent against molecular forces
in bringing additional molecules within the surface-layer. If energy be expended
in work against the cohesive forces during the stretching, the surface of the liquid
will be cooled ; and if the process be conducted isothermally, the inflow of heat will
increase the surface energy ; and the total surface energy will therefore be the sum
of two factors ; (i) The mechanical work a performed against the intermolecular
forces — with water at 0°, for example, this is equivalent to 75 ergs ; and (ii) the
inflow of heat — bound energy — required to maintain the temperature of the film
constant during the stretching — with water at 0°, for example, this is equivalent to
about 40 ergs. Lord Kelvin (1858) ,8 by reasoning analogous to that employed in
deducing H. von Helmholtz's equation, showed that the free energy, W, or the energy
which is available for doing work, is then equal to the surface energy a less the heat
VOL. I. 3 I
Benzene.
Alcohol.
(Tq
. 29-4
34-3
a .
. 0.0035
0-0027
3a
. 000139
000124
a/Sa
2-5
2-2
850 INORGANIC AND THEORETICAL CHEMISTRY
energy absorbed from the surrounding bodies, which is represented by the product
of the absolute temperature and the temperature coefficient, dcr/dT, or the variation
of surface tension with temperature is given by
The change of surface tension with temperature is a linear function over a consider-
able range of temperature, such that the surface tension o- at a temperature 6 is,
(j=z(jQ—hd, where o-q and h are constants ; accordingly, under these conditions,
the temperature coefficient by the surface energy is a constant, or da/dT=~b,
where the negative sign means that the surface tension decreases with increasing
temperature.
P. S. de Laplace inferred that the ratio of the temperature coefficient of the sur-
face tension and the coefficient of thermal expansion should be a constant, and this
has been verified in a number of cases. For example, if gq be the surface tension at
0°, and the surface tension a at 6° is G=GQ{l—a9) ; and if 3a be the
coefficient of cubical expansion, the ratio a/Sa is nearly constant :
Phenol. Acetic acid. Nitrobenzene. Carbon disulphide.
40-6 28-9 48-2 54-6
0-0029 0-0038 0-0028 0-0029
000089 0-00116 0-00089 0-00121
3-3 3-3 31 2-0
Hence, the effect of temperature on the surface tension is largely an effect of the
change of density. R. D.Kleeman^ obtained the relation g=K{Di—D) between
the density of a liquid Dj and of its saturated vapour D, when Kisa, constant depen-
dent on the nature of the liquid. According to W. A. Kistiakowsky, at the absolute
boiling point, Tb, the capillary rise a^ in a tube of 1 mm. radium, is a^M=KTb,
where M represents the molecular weight, and Kisa, universal constant with the same
value, 0'00116 ± 0"004, for all non-associating liquids. By definition, their boiling
points are proportional, and the constant K=KTblMDb^, where Dj denotes the
density of the liquid at its boiling point. C. Schall found that the surface tension of
liquids varies approximately as the 2|'s power of the density : o-/Z)'=a constant.
P. S. de Laplace's assumption that the surface density is the same as the density of
the main body of the liquid failed to explain why the surface tension diminished with
temperature more rapidly than the body density. By assuming that the densities
of the surface film and the body of the liquid are markedly different, and that there
is a corresponding difference of stress in the two regions, it will be understood how
the rate of variation of surface density with temperature, as hinted at by C. Schall's
rule, can be more rapid than that of the body density. This is evidenced by the
heating of the upper film of liquid in a capillary tube producing a more marked effect
than heating the liquid lower down. E. H. Amagat ^^ also found that a stress of 3000
atm. diminishes the coefficient of thermal expansion of ether, between 0° and 50°,
from O'OOIT at one atm. pressure to about one-third this value at 3000 atm. pressure.
J. W. Gibbs (1876) also showed that the surface layer may be regarded as a special
phase with its own characteristic density and entropy.
The relation between surface tension and compressibility. — As a rule the surface
tension of highly compressible liquids is low,ii provided there are no changes in the
character of the molecules of the liquids under investigation. For example, repre-
senting the compressibility per atmosphere by jS, and the surface tension by a,
iSxlO* .
The product jSxo- is not a constant, but, according to T. W. Richards and
J. H. Mathews (1908), ^^ the product of the surface tension a of about thirty-seven
organic liquids with the cube root of the fourth power of the compressibility /c, is
Ether.
Acetone.
Alcohol.
Chloroform.
Benzene.
Water.
Mercury.
190
121
105
103
92
48
3-83
15-9
23-3
21-6
20-8
28-3
7.5
440
THE KINETIC THEORY OF ATOMS AND MOLECULES 851
constant. There are difficulties in comparing the compressibilities of different
liquids, because the results are in part determined by the shapes of the molecules.
I. Traube found that the product of the intrinsic pressure into the square root of the
atomic (or molecular) compressibility is approximately constant for 14 elements in
the solid state. Impurities in the metals may disturb the relation, for small traces
of foreign matter may have exerted marked effect on the internal pressure.
The relation between surface tension and latent heat of vaporization.— J. J.
Waterston 13 made one of the first attempts to connect the surface tension, o, with
the molecular latent heats of evaporation, MA, and the molecular volumes, v.
His expression MX^=k(jv^ — where k is a. constant — is but a rough approximation,
which, according to R. Eotvos, gives better results if the data for different substances
are determined for corresponding states, i.e. at their critical temperatures, or the same
temperatures reckoned downwards from the critical temperatures as zero. J. Stefan ^^
has shown that the observed heat of vaporization, L, expressed in suitable units, is
equal to the internal work performed in transforming the liquid into vapour ; and
that this, in turn, is equal to the product of the volume v of the liquid, and the
difference in the internal pressure P and the vapour pressure of the liquid p. Other-
wise expressed, L={P—p)v. G. Bakker (1888) regards P as the internal or cohe-
sive pressure per unit area across any section in the interior of the liquid. In that
case, Pdv represents the internal work done when a liquid expands by dv ; and, if
the liquid changes its state so that Vq volumes of liquid become Vi volumes of vapour,
at the same temperature, the observed latent heat will be L=Pdv-\-p{vi—V2) ;
and the internal latent heat will be X=fPdv. G. Bakker further assumed that
Kisa function of v such that P=Av-^, so that the internal latent heat
when the volume Vi is very large compared with Vq. If A is independent of tempera-
ture the internal latent heat will be identical with av~^ in J. D. van der Waals'
equation, and accordingly will correspond with the assumption that the molecules
attract one another inversely as the fourth power of their distance apart. The values
of P calculated from this relation and also from the assumption that P is equivalent
to the av~^ of J. T>. van der Waals' equation, agree in a number of cases ; but
G. Bakker believes that ^ is a function of the temperature such that A^a—TdajdT.
Assuming that G. Bakker's A is equal to J. D. van der Waals' a at the temperature
of vaporization, the
n TiT
Latent heat of vaporization, L= h v:r
vo M
where Vq, being small in comparison with Vj, has been neglected and ^(^i— Vq)
becomes pv^ ; and hence if M be the molecular weight of the vapour, and Boyle-
Charles' law obtains, pv=RTIM. Calculations ^^ based on this relation do not
agree very well with the observed latent heats.
P. Walden ^^ has indicated a number of relations between the surface tension and the latent
heat of vaporization L in calories. If Lb and Lm respectively denote the latent heats of
weight; and Cj and c^, the surface tensions at the boiling and melting points respectively :
vaporization and fusion; v, the molecular volume ; Z^jandD^, the density ; M, the molecular
Di,L^ = 34:'8a^ ; MLj,^3-C^4v(rf^; Dt^L^w— 7'2(r^, provided the liquids do not form more
normal complex molecules than is indicated in the normal formula weights. D, L. Ham-
mick gives the relation 6(rVld=L, where d denotes the molecular diameter; L the internal
latent heat ; and V the gram-molecular volume.
E. T. Whittaker found empirically that for about half a dozen liquids, the surface energy
o- of a liquid in contact with its own vapour is proportional to the product of the internal
latent heat A and the absolute temperature T. Accordingly, <r = kTX, where k is a
constant which R. D. Kleeman computed to be equal to 0"557Mii)c/Tc, when Dc and Tc
respectively denote the critical density and critical temperature. E. T. Whittaker'g rule
is but an empirical density approximation. Several other relations between the molecular
852 INORGANIC AND THEORETICAL CHEMISTRY
internal latent heat, MX, and the densities of the liquid, £>i, and saturated vapour, D,
have been proposed.^' C. Dieterioi (1908), for instance, suggested MX=kT log (DJD) ;
R. D. Kleeman and A. J. Batschinsky, MX=ki{Di^—D^), where k and ^^ are constants
dependent on the nature of the liquid. The changes have been rung on these relations by
substitutions with the various vapour pressure formulae. R. Clausius and E. Clapeyron's
equation, Trouton's relation, R. Eotvos' rule, and many other subsidiary relations have been
obtained. W. C. McC. Lewis deduced a relation between the latent heat of vaporization,
L, of a liquid and the product of absolute temperature, T, and the coefficient of thermal
expansion a divided by the product of the density D and the coefficient of compressibility
K at constant volmne, such that L=—Ta/DK. L. Henry found this relation agrees
very well with a number of observations but for the exceptional behaviour of water, ^^ the
alcohols, and the fatty acids, for which there is a large amount of evidence pointing to irregu-
larities in the molecular structure.
J. D. van der Waals ^"deduced a form of Trouton's rule from his equation of state MX/Tc
is a constant, where Tc denotes the critical temperature ; and further that the latent heat of
vaporization of all substances is independent of the temperature reckoned from the critical
temperature. This does not agree with experiment. Several other more or less empirical
relations have been obtained. Thus, P. de Heen ^o obtained from his theory of fluids
Ci — Cg=l-S33aX, when Ci and Cg respectively denote the specific heats of liquids and gas ;
a, the coefficient of expansion; and A, the latent heat of vaporization; A. Nadejedine,
X=kCp, when ^ is a constant ; (7, the specific heat ot the liquid ; and p, the pressure under
which evaporation proceeds ; O. Tumlirz, A = 0'67537yT'/Z), which can easily be reduced to
Trouton's rule.
The relation between intrinsic pressure and solubility. — P. Walden^i has
compared the intrinsic pressures of a number of liquids with their solubilities in water,
and found that they run parallel with one another, for the mutual solubility of
two liquids was found to be greater, the smaller the difference in their intrinsic
pressures ; and if this difference is very great, the two liquids are immiscible.
S. W. Smith (1917) illustrates this rule by the cases of silver and gold, and of zinc
and lead. The ratio of the intrinsic pressures with the former pair of metals is as
1 : 1-08, and of the latter, 1 : 213. The former are completely miscible, the latter
only partially so.
It can be shown thermodynamically that the specific heat of a liquid is indepen-
dent of the magnitude of the surface and that the amount of energy necessary for
the production of a new surface is dependent on the temperature. This, says
H. Freundlich, is because the surface energy is of a potential not a kinetic
nature. P. N. Pavloff 22 found that the effect of surface tension on the melting point
shows that very small particles have a greater surface tension and a smaller melting
point than coarser-grained particles. According to P. Walden, if M denotes the
molecular (or formula) weight of a substance which has a specific cohesion a^2 ^t
the melting point T^° K., then ilfa^s^s-gs J^.
The solvent powers of liquids for the indifferent gases have been found by
G. Geffcken,23 A. Ritzel, etc., to run parallel with the compressibility of the solvents.
There is also a parallelism between the lowering of the compressibility and the lower-
ing of the solubility when salts are added to water. The greater the compressi-
bility of the solvent, the greater the amount dissolved. A. Ritzel postulates that
every gas possesses what he calls a solubility pressure — Ldslichkeitsdruch — tt, for a
given fluid. The solubility pressure, which favours solution, is opposed by a counter
pressure which hinders solution. When the solution is saturated, the opposing
pressures are balanced. If j8 denotes the compressibility coefficient of the liquid,
Sy the solubility of a given gas in the liquid, and S the change in volume which
occurs when unit volume of the liquid is saturated with gas, A. Ritzel assumes that
meaning that the solubility is greater, the greater the solubility pressure, the greater
the compressibility of the solvent, and the smaller the change in volume which occurs
on solution. Without testing this equation quantitatively owing to lack of measure-
ments of the solubility pressure, tt, A. Ritzel calculates values of tt, and showij
THE KINETIC THEORY OF ATOMS AND MOLECULES
853
qualitatively that the preceding assumptions are valid. The results for the solutions
of carbon monoxide in some liquids are shown in Table XXXV.
Table XXXV — Solubility Pressure of Carbon Monoxide.
Solvent.
P
s
S
IT
Acetone
Chloroform .
Benzene
Alcohol
0-0001210
00001030
0-0000915
00000875
0-00211
0-00224
000231
0-00209
0-238
0-206
0-174
0172
4-15
4-48
4-39
411
The surface tension decreases as the compressibility increases, and the solvent
powers of liquids decrease as the surface tension increases. The surface tension
curve for the solubility of gases in binary mixtures exhibits a maximum where the
solubility curve shows a minimum ; but not always conversely owing possibly to
secondary disturbing effects.
Several attempts have been made to represent the surface tension of a mixture
by the mixture law : o-=aiori+a20'2+ • • •> where aj, a2, . . . denote the
fractional proportions of the two liquids, so that ai+a2=l, and cri, 0-2, .. •
their respective surface tensions. H. Kodenbeck found this rule to apply with
mixtures of alcohol with water or chloroform ; and chloroform with ether or petro-
leum. W. Sutherland 24 deduced, from the inverse fourth law of molecular attraction,
the expression :
cr fWiVcTi , WiVc
D2""V Di
where D^ and D2 denote the respective specific gravities of the components of the
mixture ; cti and ct2, the respective surface tensions : w^ and W2,, the proportions by
weight ; and cr and D, the respective surface tension and specific gravity of the
mixture. This is regarded as a special case of S. D. Poisson's formula o-^ai^cri
+2aia20-i2+a2^o'25 where 0-^2 is a characteristic constant for the interface of the
superposed liquids. P. Volkmann, C. E. Linebarger, and W. H. Whatmough found
the formula to be applicable in some cases, not in others.
Very little advance has been made in deducing relations between the chemical
constitution and the surface tension of chemical compounds. D. I. Mendeleeff 25 has
shown that the product of the molecular weight and the surface tension — unfortu-
nately called the 7nolecular cohesion — with certain homologous series of compounds
varies proportionally with the number of CH2 groups introduced ; and that the con-
stant 2a/jD generally varies in the same direction as the latent heat of vaporization.
The substitution of hydrogen by an equivalent of oxygen or the halogens raises the
coefficient. Dilute solutions have also been investigated. The surface tension of
a liquid is altered when a substance is dissolved therein. In general, the inorganic
salts slightly raise the surface tension of water ; hydrogen chloride or bromide and
ammonia lower the surface tension of water ; sulphuric acid and alkali hydroxides
raise the surface tension of water.
The surface tension of solutions. — In 1875, G. Quincke showed that the surface
tension of aqueous solutions, or, is a linear function of the concentration C expressed
in gram-equivalents per litre, then G=Gg-{-hC, where 6 is a constant — 01566 for
sodium chloride, and 0'1666 for potassium chloride. Otherwise expressed, the
so-called molecular rise of the surface tension of solutions defined by (a — (J^IC, is
perceptibly constant for dilute solutions. 26 G. Pann found that this proportionality
does not obtain with more concentrated solutions, and E. H. Archibald connected
the raising of the surface tension with the degree of ionization a of the salt in solution
by the expression G—(jy}=a(\~a)C-\-haCj where a and h are constants.
854 INORGANIC AND THEORETICAL CHEMISTRY
The relation between the concentration of the surface film and the body o£ a
solution. — The distribution of the solute between the surface film and the body
of the liquid is not necessarily the same. J. W. Gibbs,27 in his classical memoir On
the equilibrium of heterogeneous substances (1876), first showed the relation between
the concentration of the surface film and the surface tension :
dp da
''dc=-^dc
where u denotes the excess concentration in grams per sq. cm. surface over the
concentration C in the body of the solution ; dpjdC denotes the coefficient of the
change dp in the osmotic pressure which is attended by a change dC in the concen-
tration. This magnitude is positive for all solutions. The coefficient dajdC
represents the coefficient of the change in the surface tension for a small change dC
in the concentration of the solution. Since u and daldC are always opposite in
sign, the surface tension will increase with concentration if the concentration of the
surface film is less than in the interior of the liquid ; and conversely, for dilute
solutions obeying J. H. van't Hoff's osmotic pressure formula ^=jRCT, or dpldC=RT,
and therefore J. W. Gibbs' relation between the amount u of solute adsorbed by
a surface film and the change in the surface tension per unit change of concentration,
assumes the form :
-_^ ^a
""" RT'dC
This same expression follows directly from the principle of virtual work. A surface
s contains a gram-molecule of salt and the surface tension is a, so that the surface tension
is diminished da, when a little solute enters the surface, and the change of surface energy is
8d<r. To remove this amount of solute from a volume v of the solution against the osmotic
pressure p, requires the expenditure of energy vdp so that sd(r-\-vdp=0. If the gas law is
applicable, v^RT/p, and therefore da/dp^—RT/sp; but the osmotic pressure is directly
proportional to the concentration C, so that da/dC= —RT/Cs. Again, since s is the surface
which contains a gram-molecule excess of solute, and if u denotes the excess in unit area,
u = l/s. Hence, Te&TTanging terms, u— —(C/RT){d(rldC).
It will be observed that if dajdC be positive, *.e. when the increase of the surface
tension with increasing concentration of the solute in the film is negative, the surface
will contain less solute than the body of the liquid — this is called negative adsorption —
and the effect of the salt in raising the surface tension of the solvent will in conse-
quence be partially counteracted. Again, if dojdC be negative, the surface tension
will decrease with concentration, as is the case with many organic compounds — e.g.
the oleates, amyl alcohol, etc.— and u will be positive, meaning that the solute will
be absorbed by the surface so that the concentration of the solute in the surface film
will be greater than in the body of the liquid. This is called j^ositive adsorption.
The term adsorption is usually applied to this surface layer. Consequently, a solute
is positively adsorbed by a surface film when it lowers the surface tension of a solvent
towards its own vapour, and negatively adsorbed when it raises the surface tension of
the solvent, A small quantity of the dissolved substance can lower the surface tension
of a solution to a marked degree, but a solute cannot raise the surface tension very
much because in the latter case, the concentration of the solvent in the surface film
will be less than in the body of the liquid, and the extreme limit is attained when the
surface film is purely solvent. Here then u and dojdC can possess only a small
value. On the other hand, if the adsorption be positive, the whole of the dissolved
substance, under suitable conditions, will be concentrated in the surface layer,
and this can reduce the surface tension very much. There is a limiting case with a
solution containing 0'0O0022 gram-molecule of salicylic acid per litre, for all the solute
collects in the surface film. The concentration of the solute in the surface film has
been investigated by J. von Zawidsky,28 C. C. Benson, and S. R. Milner by producing
a copious froth which has a very large surface, and comparing the concentration
THE KINETIC THEORY OF ATOMS AND MOLECULES 855
of the solute in the froth and in the body of the liquid. Thus, with an aqueous
solution of auiyl alcohol, it was found that when C for the original solution was
0*0375 gram-molecule, C for the foam was 0*0394: gram-molecule, hence w=0*0019
gram-molecule. Near the critical point, where dajdC^O, u=0, so that with an
increase of concentration, the solute distributes itself in the interior of the liquid,
and does not enrich the surface layer. W. C. McC. Lewis' results 29 did not agree
with theory, and hence S. Arrhenius inferred that the phenomena of adsorption does
not depend on surface tension, and that all attempts to correlate these phenomena
are doomed to failure.
According to I. Traube,3o the more a solute diminishes or increases the surfacie
tension of a solvent, the smaller or greater is its intrinsic pressure, and the difference
between the surface tensions of a solvent and a solution is a measure of the intrinsic
pressure of the solution. P. Walden (1909) also showed that the greater the internal
pressure, the greater the surface tension, and the greater the tendency of the mole-
cules of a liquid to form complex aggregates ; and further, that the intrinsic
boiling pressure in atmospheres is 75"3 times the surface tension of the liquid at the
point.
The surface energy of liquids. — The volume energy of gases is a linear function
of the temperature for d{pv)ldT=R, where 72 is a constant; and in 1886,
R. Eotvos, in a paper Ueber den Zusammenhang der Oherjldchens'pannung der Fliissig-
keiten mit ihrem Molekularvolumen,^^ called attention to the fact that the surface
energy of liquids, d{(js)ldr, is likewise a function of the temperature r measured
downwards from the critical temperature, since as—kr, where A; is a constant.
This is Eotvos' rule. In illustration :
Ether (6° to 120°) ....
k
. 0-227
Chloroform (20° to 60°) .
. 0-230
Carbonyl sulphide (3° to 63°) . . '
. 0-231
Carbon disulphide (22° to 78°) .
. 0-237
Sulphur dioxide (2° to 60°)
. 0-230
W. Ramsay and J. Shields, however, have shown that a slight correction is necessary,
since the line representing the product of the surface tension a and molecular surface
s has its origin about 6° below the critical temperature, Fig. 19. Consequently,
GS=k{T—Q). In order to make his formula apply to all liquids, R. Eotvos made
the surface s such that the same number of molecules are distributed over the same
surface area, and obtained what has been called the molecular surface ; he also
assumed that the molecular surface of a liquid was equal to {Mv)^, where Mv denotes
the molecular volume of the liquid, that is, that volume of the liquid which contains
the same number of molecules, for if this volume were a cube, the edge would be
{Mv)^, and any face [Mv)^. Substituting this value of s in Eotvos' formula, what
W. Ostwald 1 calls the molekulare Oherflachenenergie — the molecular surface energy
— is obtained, namely
G{Mv)^=kT (1)
where r denotes the temperature, T, reckoned downwards from the critical tempera-
ture, Tc, that is, the difference between the critical temperature and the tempera-
ture of the experiment, r=Tc—T. The molecular surface energy, g{Mv)^,
represents the work necessary to enlarge the surface of a liquid by an amount propor-
tional to the molecular weight — the molecules being treated as spheres. The
formula assumes that the distribution of the simpler and more complex molecules
over the surface is the same as in the body of the liquid ; this assumption is justified
from the circumstance that the molecular surface energy of a mixture is the mean
of those of its constituents determined at the same temperature.
W. Ramsay and J. Shields tested Eotvos' formula for about fifty liquids at
temperatures ranging from —89 8° up to the critical temperature. It was found
856 INORGANIC AND THEORETICAL CHEMISTRY
that the liquids arranged themselves into two groups when the observed data were
substituted in the formula :
a{Mv)i=k{T-d); oT(T(Mv)^=k{Tc-T-d) . . (2)
The theory of the equations of R. Eotvos and W. Ramsay and J. Shields is somewhat
as follows : Assuming that the internal attraction of the molecules varies inversely as the
fourth power of their distance apart, or, what is the same thing, inversely as the volume of
unit mass — specific volume — the work of expansion W from a specific volume v of liquid to
a specific volume V of vapour against internal molecular forces, will be
r a / 1 1\
Internal work= / ^^dv ; or, Tr=a( — — j
where the constant a can be calculated from the internal latent heat of vaporization of the
liquid. If the internal heat of vaporization be denoted by A, it follows that W = X; and,
since the specific volume is inversely as the density, if Di and D respectively denote the
densities of liquid and vapour. The internal latent heat of vaporization, X=a{D^-~D). If
the density of the vapour be negligibly small in comparison with that of the liquid, A =al>i,
but by L. Cailletet and E. Mathias' rule, if the density of the vapour be likewise negligibly
small, D^=a—hT. By differentiating these two expressions with respect to T, dXjdT
=adDJdT ; and dZ>i/(^T = constant, meaning that the decrease in the latent heat or the
density of a liquid with temperature has a constant value for substances not too near
their critical temperatures, and which do not change their chemical character with
the change of temperature. Consequently, dX/dT = & constant. Again, according to
J. Stefan, ^2 neglecting thy influence of the vapour, a molecule passing from the interior
of a liquid to the surface will escape from one -half the molecular attraction, and if
it evaporates clear away from the surface, it will escape from the other half of this
attraction. Consequently, the work required to bring a molecule from the interior
to the surface is one-half the work required to transport a molecule from the interior
of the liquid to a point outside where the attraction of the liquid is no longer sensible.
The molecular surface energy a{Mv)^, or the work necessary to bring a number of mole-
cules proportional to the molecular weight to the surface of a liquid, is therefore equal to
^A ; and by differentiation of this expression with respect to T, it follows that the varia-
tion of the molecular surface energy with respect to temperature is equal to a constant, say
k ; by integration, W. Ramsay and J . Shields' expression follows at once.
Returning to the results of W. Ramsay and J. Shields' experiments, in the one
group — called normal liquids — the value of k was virtually constant, averaging
2'2 — with a positive or negative deviation of 5 per cent, when the constant d=6.
It is assumed that in normal liquids the gaseous and liquid molecules are of the same
degree of complexity. If the liquid molecules were formed by the association of a
number of gaseous molecules so that (i) all the molecules were associated to an
equal extent ; and (ii) the degrees of the association were not altered by changes of
temperature, the liquid would give constant values for the constant k ; but it is
extremely unlikely that mere liquefaction would produce an equal or uniform
association of the molecules, and that the degree of association would not be altered
by a rise of temperature. Hence it is inferred that the so-called normal liquids are
not associated. Among the normal or non-associated liquids are : Carbon disul-
phide, nitrogen peroxide (not the gas), silicon tetrachloride, phosphorus trichloride,
phosphoryl chloride, sulphur dichloride, thionyl chloride, sulphuryl chloride, nickel
tetracarbonyl, carbon tetrachloride, ethyl ether, benzene, hydrocarbons, etc.
In another group— called associated liquids — the value of k was not constant,
but ranged between comparatively wide limits. Accordingly for these liquids
a{Mv)^ varies with the temperature. Since g and v are determined experimentally
in each case, it follows that M varies with the temperature, and that the molecules
must be more complex at low than at high temperatures. Assuming that the mole-
cular weights of non-associated liquids are the same as the molecular weights of the
compounds in the gaseous state, the product of the molecular weight, M, in the
gaseous state, multiplied by a factor i, will give the relative molecular weight of the
liquid molecules ; i of course varies with the temperature, and, for these liquids,
a{iM.v)^=k{T—d). If k' denotes the observed value of the constant for any
THE KINETIC THEORY OF ATOMS AND MOLECULES 857
temperature, and k=2'12, the division of the last equation by (2), gives ti=2*12/A;'.
Such liquids are undoubtedly associated, and have a greater molecular weight than
when in the gaseous state. Among the associated liquids are water, formic, nitric,
and sulphuric acids, bromine, the alcohols, organic acids, etc. The values of the
constant i for acetic acid, methyl alcohol, and ethyl alcohol are respectively 3'73,
3-43, and 2*79. These numbers are probably a little high. Thus, i for water between
0° and 140° changes from 1-707 to 1-289 ; "^for acetic acid between 20° to 280° from
2-13 io 1-30 ; for methyl alcohol between —89-8° and 220°, from 2-65 to 1-75 ; and
for ethyl alcohol, between —89-8° to 230°, from 2-03 to I'OO. The variation of
i^ with temperature was not allowed for, and a new formula making provision for
this variation has been obtained, and the results are better.
a(Mv)i:
JcJT-d)
l+er
. (3)
The revised formula includes the constants d and e, whose numerical values are
dependent upon the nature of the liquid. The constants are :
Methyl alcohol
Ethyl alcohol
Water . .
Acetic acid ....
The formula for computing the degree of complexity i of the molecules of the liquid
now assumes the form :
Critical
k
d
e
temperature
1-489
4-22
0-00104
240-0
2-170
4-8
0-00193
243-1
2-631
19-5
0-00218
358-1
1-910
11-9
0-00163
321-5
.=|-^(l-«r)j .
(4)
The degree of association for water thus becomes :
i
0°
20°
60°
100°
140° C
Water
. 1-7
1-6
1-5
1-4
1-3
Ethyl alcohol
. 2-0
1-7
1-4
1-2
10
Methyl alcohol
. 2-7
2-3
2-1
1-9
1-8
Acetic acid .
. 2-1
1-9
1-7
1-5
1-3
The general conclusion is that W. Ramsay and J. Shields' equsition a(Mv)^=k{r—d),
where A;=2-12 (nearly), applies generally for substances whose chemical nature
does not alter with temperature. If the expression
G{Mv)^=kT be plotted, the slope of the curve repre-
sents the value of k, and the introduction of the
constant d means that the straight line starts not
from the critical point, but from a point at a
distance d from the critical point. The distance
between the dotted and the curved line OA, Fig. 19,
represents the deviation of the observations from
Ramsay and Shields' rule in the vicinity of the
critical temperature.
F. M. Jager has measured the surface tension
and molecular surface energy of about 200 organic
liquids between —80° and 250°, and of about 50
inorganic substances in the molten condition between
300° and 1650°. The surface tension, g, and the
molecular surface energy, fju, of the halides of phosphorus, arsenic, antimony,
and bismuth increase with the molecular weight, while the variation of fi
with temperature, dfi/dT, is more or less normal. W^ith the halides of the five
alkali metals the surface tension of the molten salt decreases (i) with increasing
atomic weight of the halogen from fluorine to iodine, and also (ii) with increasing
atomic weight of the metal. The molecular surface energy varies in an irregular
1
1 ■
/
^
§
1/
j
/
Xy'
Temt
>erarure T
0
5
0
100
Fig 19. — r. Eotvos' Curve for
Benzene.
858 INORGANIC AND THEORETICAL CHEMISTRY
manner, while dfi/dT is in all cases small. The alkali sulphates, nitrates, borates,
molybdates, and tiingstates were also investigated.
The association of the molecules of Uquids.— ^The abnormal vapour densities
of certain liquids at temperatures near their boiling points, led chemists — e.g. A. Neu-
mann 33 — to the view that the molecules of suph liquids may be formed by the
coalescence or association of two or more molecules of the substance as they occur
in the gaseous state, and in 1888, P. de Heen 3* developed a theory of liquids based
upon the assumption that the constituent molecules of certain liquids are aggregates
of the molecules as they occur in the gaseous state, so that he postulated what he
called liquidogenic and gasogenic molecules. A liquid under ordinary conditions is
a solution of gasogenic in the liquidogenic molecules. If a very small volume of
liquid is in equilibrium with a large volume of vapour, the liquid will be saturated
with gasogenic molecules and cannot furnish liquidogenic molecules to the vapour ;
conversely, if the volume of the liquid is large, and the vapour small, the vapour
will consist largely of liquidogenic molecules. The density of a saturated vapour of
a pure substance, like that of a mixture, is therefore supposed to be dependent on
the relative masses of liquid and vapour phases. This is contrary to experience.
Modifications of P. de Heen's theory in which the two forms of molecules are in a
definite state of equilibrium have been employed to explain supposed phenomena
which occur at the critical temperature of a liquid or gas, namely, that (i) the greater
the proportion of liquid confined in the tube, heated to the critical temperature, the
higher the critical temperature ; and (ii) that the critical temperature at which a
meniscus appears on cooling is lower than that obtained on heating. There is, how-
ever, considerable doubt about the accuracy of both conclusions, since the experi-
mental errors are large, and care has not always been taken to use pure materials free
from absorbed air.35 It is generally held that the difference between the liquid and
gaseous states is solely due to the greater propinquity of the molecules in the liquid
state, and not to polymerization of the molecules. Nothing very definite, however,
was known about this until about 1887, when W. Ramsay and S. Young 36 showed
that the density of acetic acid vapour increases as the boiling point is approached ;
and about 1888, E. Beckmann 37 showed, by freezing-point determinations, that
acetic acid and ethyl alcohol possess more complex molecules in concentrated solu-
tion than in the vaporous state ; while naphthalene does not form complex molecules
under the same conditions. J. T. Cundall (1891) also showed that solutions of
nitrogen peroxide behave as if the molecules are more complex in solution than in
the vaporous state.
References.
^ C. V. Boys, Soap-bubbhs and the Forces which would mould them, London, 1890 ; C. R. Darling,
Liquid Drops and Globules, London, 1914 ; H. C. Proctor, Chem. News, 118. 292, 1919.
2 J. C. Maxwell, Theory of Heat, London, 281, 1875; R. S. Willows, and E. Hatschek,
Surface Tension and Surface Energy, London, 1915.
3 Lord Kelvin (W. Thomson), Phil. Mag., (4), 42. 375, 1871 ; Lord Rayleigh, ih., (5), 30.
386, 1890 ; ib., (5), 48. 321, 1899 ; L. Grummach, Ann. Physik, (4), 3. 660, 1900 ; F. Kolacck, If ied.
Ann., 5. 425, 1878; ib., 6. 616, 1879 ; G. Quincke, Pogg. Ann., 134. 356, 1868 ; 135. 621, 1868 ;
G. Hagen, ib., 67. 167, 1846 ; 77. 455, 1849 ; T. Lohnstein, Ann. Physik, (4), 20. 606, 1906 ;
(4), 22. 771, 1907.
* M. PJanck, Vorlesungen iiber Thermodynamik, Leipzig, 175, 1905 ; M. Cantor, Wied. Ann.,
67. 687, 1899 ; D. Konovaloff, ib., 14. 34, 1881 ; G. N. Antonoff, Journ. Chirn. Phys., 5. 372,
1907 ; Phil. Mag., (6), 36. 377, 1918; (6), 38. 417, 1919; Lord Rayleigh, ib., (5), 16. 309, 1883 ;
(5), 30. 285, 1890 ; (5), 33. 209, 468, 1892.
^ D. I. Mendeleeff, Compt. Bend., 50. 52, 1860 ; 51. 97, 1860 ; A. L. Selby, Phil. Mag., (5),
31. 430, 1891 ; M. Prud'horame, Journ. Chim. Phys., 14. 285, 1916 ; 16. 405, 1918.
« M. Cantor, Wied. Ann., 47. 420, 1892 ; H. Pellat, Compt. Bend., 118. 1193, 1894 ; M. L.
Frankenheim, Pogg. Ann., 75. 29, 1848; Journ. prakt. Chem., (2), 23. 401, 1841 ; Die Lehre
von der Kohdsion, Breslau, 1835 ; A. Ferguson, Science Progress, 9. 428, 1915.
' J. D. van der Waals, Zeit. phys. Chem., 13. 657, 1894.
« Lord Kelvin (W. Ihomson), Proc. Boy. Soc, 9. 255, 1858.
THE KINETIC THEORY OF ATOMS AND MOLECULES 859
9 R. D. Kleeman, Phil. Mag., (0), 19. 873, 1910; (0), 21. 83, 1911 ; W. A. Kistiakowsky,
Zeit. Elektrochem., 12. 513, 1906 ; Journ. Russian Phys. Ckem. Soc, 45. 782, 1913 ; P. N. Pavloff,
ib., 48. 1008, 1175, 1916 ; C. Schall, Ber., 14. 555, 1881.
10 E. H. Amagat, Ann. Cfiim. Phys., (6), 29. 551, 1893; J. W. Gihhs^Scientific Papers, London,
1. 219, 1906.
11 W. C. Rontgen and J. Schneider, Wied. Ann., 29. 165, 1886 ; A. Einstein, Ann. Physik,
(4), 4. 513, 1901 ; J. D. van der Waals, On the Continuity of the Liquid and Gaseous States,
London, 362, 1918 ; A. Ritzel, Zeit. phys. Chem., 60. 319, 1907.
12 T. W. Richards and J. H. Mathews, Zeit. phys. Chem., 61. 449, 1908 ; A. Ritzel, ih., 60.
319, 1907 ; I. Traube, Ber. deut. phys. Ges., 11. 231, 1909.
13 J. J. Waterston, Phil. Mag., (4), 14. 279, 1857 ; W. Sutherland, ib., (5), 27. 305, 1889 ;
R. Eotvos, Wied. Ann., 27. 448, 1886.
1* G. Bakker, Zeit. phys. Chem., 10. 5, 1892 ; 12. 670, 1893 ; 18. 519, 1895 ; W. C. McC.
I^wis, ib., 79. 196, 1912 ; Trans. Faraday Soc, 7. 94, 1911 ; E. Mathias, Chaleur de vaporization
des gaz liquefies, Paris, 1890 ; K. Fuchs, Exner's Repert., 26. 345, 1890 ; J. Stefan, Wied. Ann.,
29. 655, 1886.
15 I. Traube, Ann. Physik, (4), 8. 300, 1902 ; Zeit. anorg. Chem., 34. 423, 1903 ; A. P. Mathews,
Journ. Phys. Chem., 17. 154, 180, 605, 1913.
16 P. Walden, Zeit. phys. Chem., 65. 129, 257, 1909 ; 66. 385, 1909 ; E. T. Whittaker, Proc.
Roy. Soc, 81. 21, 1908 ; R. D. Kleeman, Phil. Mag., (6), 18. 491,901, 1909 : (6), 19. 783, 1910 ;
J. E. Mills and D. MacRae, Journ. Am,er. Chem. Soc., 32. 1162, 1910 ; D. L. Hammick, Phil. Mag.,
(6), 38. 240, 1919 ; G. Rudorf, ib., (6), 39. 238, 1920 ; W. Herz, Zeit. Elektrochem., 25. 323, 1919.
1' H. Crompton, Proc. Chem. Soc, 17. 61, 1901 ; C. Dieterici, Ann. Physik, (4), 25. 569, 1909 ;
R. D. Kleeman, Phil. Mag., (6), 18. 78, 1910 ; A. J. Batschinsky, Ann. Physik, (4), 14. 288, 1904 ;
D. Tyrer, Journ. Phys. Chem., 17. 717, 1913 ; J. E. Mills, Journ. Amer. Chem. Soc, 31. 1099,
1909; J. Puschl, Sitzber. Akad. Wien, 75. 75, 1877 ; 82. 1102, 1881 ; W. Jager, ib., 100. 1122,
1891 ; B. Walter, Wied. Ann., 16. 500, 1880 ; J. A. Groshans, ib., 64. 778, 1898 ; C. Antoine,
Ann. Chim. Phys., (6), 26. 426, 1892.
18 L. Henry, Ann. Soc Scient. Bruxelles, 267", 1879 ; W. C. McC. Lewis, Phil. Mag., (6), 21.
268. 1911.
. 1^ J. D. van der Waals, Die Continuitdt des gasjormigen und flussigen Zustandes, Leipzig, 1899.
20 P. de Heen, Ann. Chim. Phys., (6), 5. 83, 1883 ; A. Nadejedine, Rep. Phys., 20. 441, 1884 ;
0. Tumlirz, Sitzber. Akad. Wien. 101. 184, 1892 ; H. Mache, ib.. 111. 382, 1902.
21 P. Walden, Zeit. phys. Chem., 65. 129, 1909 ; 66. 385, 1909 ; Zeit. Elektrochem., 14. 713,
1908 ; S. W. Smith, Journ. Inst. Metals, 17. i, 65, 1917.
2 2 P. N. Pavloff, Zeit. phys. Chem., 65. 1, 545, 1908 ; P. Walden, ib., 65. 129, 257, 1909 ; 66.
385, 1909 ; J. W. Mellor, A. Latimer, and A. D. Holdcroft, Trans. Cer. Soc, 9. 126, 1909.
23 G. Geffcken, Zeit. phys. Chem., 49. 257, 1904 ; A. Ritzel, ib., 60. 319, 1907 ; A. Christoff,
ib., 53. 321, 1905 ; 55. 622, 1906 ; F. W. Skirroff, ib., 41. 139, 1902.
2* W. Sutherland, Phil. Mag., (5), 38. 188, 1894 ; (5), 40. 477, 1895 ; P. Volkmann, Wied. Ann.,
16. 334, 1882 • 17. 384, 1882 ; W. H. Whatmough, Zeit. phys. Chem., 39. 129, 19C1 ; M. Cantor
Ann. Physik, (4), 7. 698, 1902 ; C. E. Linebarger, Amer. Journ. Science, (4), 2. 226, 1896 ;
S. D. Poisson, Nouvelle theorie de Faction capillaire, Paris, 107, 293, 1831 ; H. Rodenbeck, Ueber
Capillaritdtsbestimmungen von Flitssigkeitsgemischen, Bonn, 1879.
26 A. Buligimsky, Pogg. Ann., 134. 150, 1868 ; G. Quincke, ib., 169. 337, 560, 1877 ; P. Volk-
mann, Wied. Ann., 17. 353, 1882 ; 28. 135, 1886 ; 0. Rother, ib., 21. 576, 1884 ; A. Valson, Ccrnipt.
Rend., 74. 103, 1872 ; E. Duclaux, ib., 85. 1068, 1877 ; Ann. Chim. Phys., (5), 2. 256, 1874 ; (5),
13. 76, 1878 ; (5), 16. 1009, 1879 ; N. E. Dorsay, Phil. Mag., (5), 44. 134, 367, 1897 ; E. H. Archi-
bald, Trans. Nova Scotia Inst. Science, 9. 335, 1898 ; W. Ochse, Exner's Repert., 26. 641, 1890 ;
M. Goldstein, Zeit. phys. Chem., 5. 233, 1890; D. I. Mendeleeff, Compt. Rend., 50. 52, 1860;
51. 96, 1860; L. Wilhelmy, Pogg. Ann., 119. 177, 1863; 121. 44, 1864; 122. 1, 1864;
G. Quincke, ib., 105. 1, 1858 ; 134. 356, 1868 ; 135. 621, 1868 ; 138. 141, 1869 ; 139. 1, 1870 ; 160.
337, 1877 ; R. Schiff, Liehig's Ann., 223. 47, 1884 ; Gazz. Chim. Ital, 14. 292, 368, 1884 ; Ber.,
18. 1603, 1885 ; E. Duclaux, Ann. Chim. Phys., (5), 13. 76, 1878 ; J. Hock, Sitzber. Akad. Wien,
108. 1516, 1900 ; R. Feustel, Ann. Physik, (4), 16. 86, 1905 ; R. Feustel, ib., (4), 16. 86, 1905 ;
E. C. Linebarger, Amer. Journ. Science, (3), 44. 83, 1892 ; Ber., 25. 937, 1892 ; F. Bede, Mem.
Acad. Bruxelles, (5), 30. 1, 1861 ; P. Dutoit and L. L. Friedrich, Arch. Science Nat. Geneve,
(4), 9. 105, 1901 ; A. Bartoh, Nuovo Cimento, (3), 6. 141, 1879 ; Atti Accad. Lincei, (3), 7. 340,
1884 ; F. Cantoni, ib., (3), 4. 74, 1880 ; A. Gradenwitz, Ueber eine neue Methodc zur Bcstimmurhg
von Kapillarkonstenten verdilnnter Salzlosungen, Breslau, 1902.
26 N. E. Dorsev, Phil. Mag., (5), 44. 134, 367, 1897 ; C. Forch, Ann. Physik, (4), 17. 744, 1906 ;
(3), 68. 801, 1899 ;' W. H. Whatmough, Zeit. phys. Chem., 39. 129, 1901 ; C. E. Linebarger, Journ.
Amer. Chem. Soc, 21. 41 1, 1899 ; G Pann, Beitrdge zur Feststellung der wahren Oberfldcheiispannung
wasseriger Sulfat-, Nitrat-, und Karbonatlosungen, Konigsberg, 1906 ; E. H. Archibald, Trans.
Nova Scotia Inst. Science, 9. 335, 1898 ; G. Jager, Sitzber. Akad. Wien, 100. 493, 1891 ; 101.
103, 1892 ; H. Sentis, Journ. Phys., (3), 6. 183, 1897.
2' J. W. Gibbs, Trans. Connecticut Acad., 3. 439, 1876 ; Scientific Papers, London, 1. 219,
1906 ; H. FreundUch, Kapillarchemie, Leipzig, 50, 1909 ; H. Freundhch and F. Em§lander,
Zeit. phys. Chem., 49. 317, 1904 ; K. Drucker, ib., 52. 641, 1905 ; S. Milner, Phil. Mag., (6), 13.
96, 1907 ; W. C. McC. Lewis, ib., (6), 15. 499, 1908 ; (6), 16. 466, 1909 ; J. J. Thomson, Applica-
860 INORGANIC AND THEORETICAL CHEMISTRY
tions of Dynamics to Physics and Chemistry, London, 190, 1888 ; O. Sackur, Lehrbuch der Thermo-
chemie und Thermodynamik, Berlin, 293, 1912.
" J. von Zawidsky, Zeit. phys. Chem., 35. 77, 1900 ; 42. 1, 1903 ; C. C. Benson, Journ. Phys.
Chem., 7. 632, 1903 ; S. R. Milner, Phil. Mag., (6), 13. 96, 1907 j H. R. Proctor, Chem. News,
118. 292, 1919.
2» W. C. McC. Ivewis, Phil. Mag., (6), 17. 466, 1909 ; G. N. Antonoff, ib., (6), 36. 377, 1918 ;
(6), 38. 417, 1919 ; S. Arrhenius, Medd. Veten. Nobelinstituf, 2. 7, 1911.
30 I. Traube, Liebig's Ann., 265. 27, 1891 ; P. Walden, Zeit. phys. Chem., 66. 385, 1909.
«i R. Eotvos, Wied. Ann., 27. 448, 1886 ; W. Ramsay and J. Shields, Phil. Trans., 184. A,
647, 1893 ; Zeit. phys. Chem., 12. 433, 1893 ; 15. 106, 1894 ; Journ. Chem. Soc, 63. 1089, 1893 ;
E. C. C. Baly and F. G. Donnan, ib., 81. 907, 1902 ; W. Ramsay, Proc. Roy. Soc, 56. 171, 1894 ;
W. Ramsay and E. Aston, ib., 5Q. 182, 1894; Trans. Roy. Irish Acad., 32. 93, 1902; Jourii.
Chem. Soc., 65. 167, 1894 ; W. E. S. Turner and E. W. Merry, ib., 97. 2069, 1910 ; F. H. Getmann,
Amer. Chem. Journ., 44. 145, 1910 ; G. Carrara and G. Ferrari, Gazz. Chim. Ital, 36. 419, 1906 ;
P. Dutoit and L. Friderich, Compt. Rend., 130. 327, 1900; P. Walden, Zeit. phys. Chem., 75.
555, 1910 ; F. M. Jager, Zeit. anorg. Chem., 101. 1, 1917 ; M. Prud'homme, Journ. Chim. Phys.,
14. 285, 1916 ; 16. 405, 1918 ; P. A. Guye and A. Baud, Archiv. Sciences Phys. Nat. Geneve,
(4), 11. 409, 537, 1901 ; I. Homfray and P. A. Guye, Journ. Chim. Phys., 1. 505, 1903 ; P. A.
Guye and J. Bolle, ib., 3. 40, 1905 ; W. Sutherland, Phil. Mag., (5), 27. 305, 1889 ; W. Ostwald,
Lehrbuch der allgemeinen C^emie, Leipzig, 1. i, 541, 1903 ; J. D. van der Waals, Zeit. phys. Chem..
13. 713, 1894.
32 J. Stefan, Wied. Ann., 29. 655, 1886.
33 A. Neumann, Liebig's Ann., 155. 325, 1870 ; Ber., 10. 2099, 1877 ; 11. 33, 1878 ; 13. 46.8,
1880 ; I. Traube, Ann. Physik, (4), 8. 289, 1902.
3* P. de Heen, Bull. Acad. Roy. Belgique, 24. 96, 1892 ; A. Battelli, Nuovo Cimento, (3), 33.
22, 1892; G. Zambiasi, Atti Accad. Lincei, (5), 1. 423, 1892; J. B. Hannay, Proc. Roy. Soc,
30. 484, 1880 ; B. Galitzine, Wied. Ann., 50. 521, 1893.
35 M. W. Travers and F. L. Usher, Proc Roy. Soc, 78. 247, 1906 ; S. Young, Stoichiometry,
London, 1918.
3« W. Ramsay and S. Young, Phil. Mag., (5), 24. 196, 1887.
37 E. Beckmann, Zeit. phys. Chem., 2. 728, 1888 ; A. P. Pari^ek and 0. Sulc, Ber., 26. 1408,
1893 ; J. Cundall, Journ. Chem. Soc, 59. 1076, 1891.
§ 23. The Association or Polymerization of Liquids
The custom of comparing all liquids in an indiscriminate fashion, in the hope of establish-
ing general relationships has been attended with a certain degree of success, but exceptions
of a very puzzling order frequently arise. One great cause of such exceptions is to be found
in the fact that monomolecular and associated liquids have been indiscriminately compared
one with another, when strictly speaking they are not truly i comparable. — H. CTtOMPTON
(1898).
There is much evidence indicating that the normal molecules of certain liquids,
vapours, and even solids may coalesce or associate into more complex aggregates
when the vapour condenses to a liquid, or even when the temperature is changed
without a change in the state of aggregation, as when n molecules of a substance, A,
coalesce to form one molecule, and vice versa : nA^An. So long as our knowledge
of molecular weights was deduced from the study of vapour densities, the conclusions
were strictly applicable to matter in the gaseous state only. A large number of
attempts have been made to get an insight into the molecular condition of liquids
and solids, and in the case of liquids, the problem now approaches within a measur-
able distance of a successful solution ; indeed, liquids are now classed as associated
or non-associated according as the molecular weight can be represented by iM or
M, where M denotes the formula -weight deduced from the vapour density, and i
the factor of association.
The evidence for polymerization is based upon some irregularity in the variation
of many of the physical properties of the liquids with temperature — e.g. heat of
vaporization ; external work of evaporation ; vapour pressure curves ; molecular
volumes ; molecular refraction ; densities ; viscosities ; etc. Several methods are
based on the behaviour of liquids either at or near their critical points, for it will be
evident that when matter is undergoing a change of state, if the molecules simul-
taneously alter their degree of association, the variation should be rendered apparent
in a marked degree. D. Tyrer lays down the condition that any exact equation
THE KINETIC THEORY OF ATOMS AND MOLECULES 861
which might be employed to calculate the association factors of liquids should con-
form with the law of mixtures, for with partially associated liquids we are dealing
with a mixture and not with an individual in the chemical sense. In the case of
water, for example, we are possibly dealing with a mixture containing not one mole-
cule but several, H2O, (H20)2, (H20)3, ... If the first three molecules have the
respective molecular weights, M^, M^, M^, . • . and %, n^, ^3, . . . of the respective
molecules be present, the mean molecular weight M of the associated liquid, (H20)t,
will be
niMi+W2M2-f%^3+ • • •
iM==
%H-W2+%+
and the various physical properties X^, X2, X^, + . . . should be functions of the
right member of this equation. This applies for all mixtures whether of associated
or normal liquids which have no chemical action on one another.
In 1894, P. A. Guye made a collection of evidence which furnished a number of
criteria indicating the association of liquids, but much more has accumulated since
then. D. Tyrer has tried a number and found them to fail when tested by the
above criterion. Ten of these methods are here indicated :
(1) W. Ramsay and J. Shields' work on surface energy. Various attempts have
been made to modify W. Ramsay and J. Shields' equation, thus, A. Batschinsky 1
proposed to substitute Tc=lQ'3l{r]T^y''IDc^'\ where rj denotes the viscosity at
the absolute temperature T ; and Dc, the critical density — ^the density at 0° may be
substituted for Dc if the constant be altered. P. Walden also says that a{Mv)
=K{Tc—T—d) gives as good results as the equation of W. Ramsay and J. Shields.
(2) About 1890, P. A. Guye 2 found it necessary to double the ordinary molecular
weights of methyl alcohol, acetic acid, and water in order to make these liquids con-
form with his rule that the quotient obtained by dividing the absolute critical
temperature Tc, by the critical pressure pc, is equal to the molecular refraction r
multiplied by a constant which is 1*85 for most liquids, but 1*1 for associated
liquids.
(3) According to S. Young and G. L. Thomas,^ the ratio of the actual density of
a liquid to the density at the critical point is 3-85 ; and with liquids assumed to be
associated, the ratio appears to have a greater value — thus with some alcohols and
acetic acid, the ratio ranges from 4*0 to 5"0.
(4) If the arithmetical mean of the density of a liquid and gas is not a linear
function of the temperature, as indicated in L. Cailletet and E. Mathias' rule,^ it is
assumed that the liquid is associated — e.g. water.
(5) If the heat of vaporization rises to a maximum with rise of temperature,
and then diminishes, it is assumed that complex molecules are being converted into
simpler ones during the descent of the curve. W. Ramsay and S. Young ^ found this
to be the case with ethyl alcohol and acetic acid. Normally, the curve shows that the
heat of vaporization decreases regularly up to the critical temperature when it
becomes zero.
(o) The greater value for the heat of vaporization for unit increase in volume in
the case of alcohols and water corresponds with the consumption of a greater amount
of internal work for the expansion against external pressure, owing to the dissocia-
tion of the complex molecules. For example, F. Trouton's rule, and G. G. Longi-
nescu's rule.^ G. G. Longinescu found that for non-associated liquids, n={TbllOOD)",
where Tb denotes the absolute boiling point ; D, the specific gravity at 0° ; and n,
the number of atoms in the molecule. In W. A. Kistiakowsky's modification of
F. Trouton's rule, a^MITf,=l'14:, where a^ denotes the capillary rise in a tube of one
mm. radius, at the absolute boiling temperature T^. For associated liquids, the
constant is smaller — e.g. for methyl alcohol it is 0*482, and for acetic acid, 0*576.
(7) The vapour pressure curves of normal liquids do not cut one another at any
point in their course, but the vapour pressure curves of associated liquids often cut
across those of normal liquids. 7
862 INORGANIC AND THEORETICAL CHEMISTRY
(8) The ratio of the volume of a saturated vapour at some chosen pressure to
that at the critical pressure points to the association of organic acids and alcohols. ^
(9) J. D. van der Waals' vapour pressure equation,^ log ^c— log p=f{Tc—T)IT,
where pc a-nd Tc denote the critical pressure and temperature respectively, and j)
denotes another pressure at the temperature T, gives a constant approximately 3 to 4
with most liquids, but the numerical value of /rises to 3'2-34 for water ; 3*36-
349 for acetic acid ; 3'58^'02 for methyl alcohol ; and 3-49-3'77 for ethyl alcohol.
(10) A. E. Dunstan and E. B. Thole lo have shown that the quotient of the
viscosity by the molecular volume is nearly 60x10"^ for normal liquids, but for
water, the fatty alcohols, and acetic acid, the fraction has more than twice its normal
value.
Several other methods have been described^e.^r. I. Traube's method of molecular
volumes, the empirical formulae of D. Tyrer, M. M. Garver, E. T. Whittaker,ii and
R. D. Kleeman, etc. Indeed, nearly every physical property which has been
accurately measured — specific heats, osmotic pressures, etc. — has been related with
the molecular weights.
, Refebences.
1 M. M. Garver, Journ. Phys. Chem., 16. 454, 669, 1912 ; 19. 500, 1915 ; D. Tyrer, ib., 19.
81, 1915 ; Phil. Mag., (6), 20. 522, 1910 ; Zeit. phys. Chem., 80.50, 1912 ; A. Batschiiisky, ih.,
75. 665, 1911 ; 82. 86, 1913 ; P. Walden, ih., 65. 129, 1908; H. Crompton, Science Prog., 7. 175,
1898.
2 P. A. Guye, Arch. Sciences Phys. Nat. Gentve, (3), 31. 38, 1894 ; Bull. Soc. Chim., (3), 13.
34, 1895 ; Ann. Chim. Phys., (6), 21. 206, 1890 ; (6), 26. 97, 1892 ; Compt. Rend., 110. 141, 1890 ;
W. W. RandeU, Amer. Chem. Journ., 18. 462, 1895 ; M. Altschul, Zeit. phys. Chem., 11. 577,
1893 ; R. Nasini, ib., 16. 248, 1895 ; D. Tyrer, ib., 80. 50, 1912.
» S. Young and G. L. Thomas, Journ. Chem. Soc, 63. 1251, 1893 ; Phil. Mag., (5), 34. 507,
1892.
* L. Cailletet and E. Mathias, Compt. Bend., 102, 1202, 1886.
6 W. Ramsay and S. Young, Phil. Mag., (5), 24. 196, 1887 ; Proc. Phys. Soc., 7. 303, 1885.
« G. G. Longinescu, Ann. Sc. Univ. Jassy, 1. 359, 1901 ; 3. 126, 1903 ; Journ. Chim. Phys.,
6. 552, 1908 ; W. A. Kistiakowsky, Zeit. Elektrochem., 12. 513, 1906.
' H. V. Regnault, Mem. Acad., 26. 700, 1862; S. Young, Stoichiometry, London, 358,1908.
8 S. Young, Journ. Chem. Soc, 63. 1257, 1893.
» T. Estreieher, Phil. Mag., (5), 40. 454, 1895.
10 A. E. Dunstan and F. B. Thole, Proc Chem. Soc, 23. 19, 1907 ; Zeit. phys. Chem., 6). 732,
1905 ; T. E. Thorpe and J. W. Rodger, Phil. Trans., 185. 397, 1895 ; E. C. Bingham and J. P.
Harrison, Zeit. phys. Chem., 56. 1, 1909.
11 E T. Whittaker, Proc Boy. Soc, 81. 21, 1908; R. D. Kleeman, PM. Mag., (6), 18. 6, 39,
1909 ; (6), 20. 664, 1910.
§ 24. Thermal Effects attending the Expansion and Compression of Gases
I use the word attraction in a general way for any endeavour of what kind so ever, made
by bodies to approach to each other- — whether that endeavour arise from the action of the
bodies themselves, or whether it may arise from the action of the aether or of any medium
whatsoever, whether corporeal or incorporeal, anyhow impelling bodies placed therein
towards one another.- — Isaac Newton.
In an essay Of the cold produced hy evaporating fluids, and of some other means of
producing cold, W. Cullen (1755) ^ seems to have been the first to notice that the
temperature of air is decreased hy rarefaction, and increased by compression ; and
J. Dalton (1802) made an attempt to measure the change of temperature which occurs
when air is compressed or rarefied. In a general way, it has been proved that if a
gas, whose molecules exert no attraction on one another, be confined in a suitable
vessel, and compressed, the mechanical work employed in compressing the gas is
equivalent to the product of the pressure into the change in volume. This energy
is transformed into an equivalent amount of heat which raises the temperature of
THE KINETIC THEORY OF ATOMS AND MOLECULES
863
-
«
\
\
i
^
•termal
W.,i
rr
^
H
the gas. On the other hand, if the gas of itself expands against atmospheric pressure
from a volume v to a volume v^, the gas will be cooled because the gas itself has done
a certain amount of work against atmospheric pressure f, equivalent to the product
of the atmospheric pressure into the change in volume — ^;^2;. ??(vi— v), or more
accurately, f log (vo/vi). The phenomenon is illustrated by the time-honoured
experiment in which a piece of tinder in a glass cylinder containing a little ether is
ignited by suddenly compressing the piston in the cylinder.
The adiabatic expansion or compression of gases. — ^If the expansion of a gas
against atmospheric pressure be performed slowly enough, the gas will remain at a
constant temperature, and the expansion is said to be isothermal; the energy
required to overcome the external pressure is absorbed as heat from its surroundings.
On the other hand, if the walls of the containing vessel be made of some insulating
material which prevents the passage of heat inwards or outwards, the energy required
for the expansion is absorbed from the kinetic energy of the gas molecules themselves,
and the temperature of the gas decreases. Such an operation is said to be adiabatic,
from a, not ; hafiaiv^tv, to pass through, or transmit. An approximation to an
adiabatic change is obtained when gases are suddenly expanded or compressed,
because there is not then time for the heat to dissipate.
Boyle's law describes the relation between the pressure
and volume of a gas when the operation is performed
isothermal ly ; this law will not be valid for adiabatic
changes. No perfectly adiabatic substance is available for
measuring the relations between the adiabatic changes of
volume and the pressure of a gas, but the law has been
derived from several lines of reasoning, and the observed
results approximate to the theoretical values. Obviously,
the actual change of temperature which occurs during
the adiabatic process must depend in some way on the
specific heat of the gas concerned. If the thermal
capacity of the gas be small, the change of temperature
will be greater than if the thermal capacity be large. If y denotes the ratio of the
two specific heats of a gas, CpjC^, then, if Jc be a constant.
The dotted curve. Fig. 19, represents Boyle's law (isothermal) curve 2?^=constant,
the other the adiabatic curve ^i;'y=constant . If the isothermal curve passes through
the point 0 when the volume of the gas is unity and the pressure ^, it passes above the
adiabatic for values of v greater than unity, and under it for values of v less than
unity. By substituting the gas law, ipv^=RT, in these equations, two other equa-
tions can be obtained :
7—1
which express the relation between the volume, pressure, and temperature of gases
undergoing adiabatic changes.
These expressions- — •pv'y = constant, etc. — can be deduced from the laws of thermo-
dynamics, somewhat as follows : Let the temperature of unit mass of gas at a constant pressure
be raised a small amount dT owing to the absorption of an amount of heat dQ, which in turn
is equivalent to CpdT when Cp is the specific heat of the gas at constant pressure ; let the
volume of the gas at the same time be augmented dv. The final condition of the gas can be
regarded as the joint effect of two operations : (i) The temperature may be supposed to
increase by an amovmt dT while the volume remains constant, so that heat equivalent to
CvdT is absorbed ; (ii) The temperature remains constant while the volume increases by
an amount dv. Let H denote what has been called the latent heat of expansion ; that is,
the amount of heat absorbed per unit change of volume without change of temperature ;
accordingly, the quantity of heat absorbed when the volume changes by a small amount
dv, at a constant temperature, will be H.dv. Consequently, dQ=Hdv-\-CvdT ; or CpdT
Volume y
Fig. 20.- — Isothermal and
Adiabatic pv-Curves.
©'-'4:^ (?)
864 INORGANIC AND THEORETICAL CHEMISTRY
= Hdv-^CvdT. From Charles' law, it follows that {v+dv)/{T+dT) =viT ; or, dv = {v/T)dT ;
and, after substituting this value of dv in CpdT=H.dv-\-CvdT, dividing through by d2\ it
follows that Cp — Cv=Hv/T ; or H = {Cp — Cv)T/v. Again, if a gas changes its volume dv
under adiabatic conditions without any exchange of heat between it and its surroundings,
the temperature would change by an amoTint dT, and dQ must be zero. Hence, H.dv
-{-CvdT=0. Substitute the above value of S in this equation ; put y for Cp/Cv; and it
follows that {Y—l)dvjv-\-dT/T=0. This on integration gives the relation 7?^'^= constant,
and the other expressions follow.
The cooling (or heating) effect obtained by expanding (or compressing) gases
adiabatically can be computed from these equations. Thus, air expanded adiabati-
cally from 20° and 50 atm. to 1 atm. pressure will be cooled to —177° ; and air at
—60° and 50° atm. pressure will be cooled to 204° on expanding adiabatically to
one atmosphere pressure. With air initially at 0°, and a final pressure of one atmo-
sphere (y= 1*41),
Initial pressure . . 50 100 200 300 400 500 atm.
Final temperature . . -187-0° -201'5° -214-5° -221-0° -225-1° -229-2°
Hence, it is possible to cool gases considerably by adiabatic expansion. By utilizing
this principle, L. P. Cailletet (1877) 2 liquefied small quantities of air, oxygen,
nitrogen, methane, and carbon monoxide. G. Claude (1909) has employed the
principle in the continuous production of liquid air.
Examples." — (1) A litre of air at 0° expands adiabatically to two litres. Find the fall
of temperature when y for air is 1-4. Here Tg X2»'4 = 273 ; 2"*-: 1-32 ; hence, ^2 = 207° K
or -66°.
(2) If the ignition temperature of electrolytic gas be 585°, what adiabatic compression
would cause the mixture to ignite if the gas were originally at 15° and atmospheric pressure ?
Assume that the ignition temperature is not altered by variations of pressure. Ansr. 43
atm.
The Joule-Thomson or the Joule-Kelvin effect. — No heat is developed when an
ideal gas expands into a vacuum. This was established by some experiments by
J. L. Gay Lussac, described in his Essai pour determiner les variations de temperature
qu'eprouvet les gaz en changeant de densite (1807) ; and by J. P. Joule, in his memoir.
On the changes of temperature produced by the rarefaction and condensation of air
(1845). Compressed air was allowed to expand into an evacuated vessel, and the
result, as J. P. Joule expressed it, was as follows : " No change of temperature
occurs when air is allowed to expand in such a way as not to develop mechanical
power" ; and generally, when a gas expands without doing external work, and
without taking in or giving out heat, its temperature does not change— a statement
sometimes called Joule's law. The first increment of gas into a vacuum will of course
produce a pressure, and each successive increment supplements the pressure produced
by the earlier portions. Since a gas expands indefinitely, each increment may be
regarded as filling the whole space or volume v, and consequently the work done is
simply vjdp=pv. This is also equal to the kinetic energy of the gas. Consequently,
the work done in forcing the later portions of gas into the evacuated space against
the existing pressure is wholly transformed into kinetic energy. No change of
temperature occurs, because the work done hy the gas is equal to the work done on
the gas. The case is analogous with the transmission of energy by a rod or belt
whose state of strain is steady — the rod or belt transmits the impressed energy with-
out loss ; so also if gas be forced into a cylinder when the temperature and pressure
are constant, the stress introduced by the entering gas is relieved by the escape of
an equal amount of gas. If the gas is in a steady state, the work done by the
escaping gas will be equal to the work done on the gas. Otherwise expressed, the
work required to transform unit mass of any fluid from a place where there is a uni-
form pressure pi to another place where there is a uniform pressure p2, is equal
to P2'^2~Pi^i^ where the subscripts refer to the respective states of the gas if the
density of the gas is strictly proportional to the pressure in the two states, P2V2
THE KINETIC THEORY OF ATOMS AND MOLECULES 865
~Pi^i=0, and no work is done. If, however, the external pressure be diminished by
internal attractions depending on the distance apart of the molecules, so that piVi
is less than 2?2^2j tlie external work done on the gas is no longer equal to the external
work done by the gas — because during expansion part of the energy is spent in doing
internal work separating the molecules. The assumption made in Joule's law
is that no work is performed against intermolecular attractions. However, inter-
molecular attractions are evidenced with most gases. The experiments of Gay
Lussac and Joule 3 were not sufficiently sensitive to detect the small change of
temperature which occurs when such gases expand in vacuo, so that although no
external work is done by the gas, internal work is done against intermolecular
attraction. The molecules are torn apart, so to speak, against the (feeble) attractive
force drawing them together. This involves an expenditure of energy — work
must be done — and the gas is cooled.
The work W done against intermolecular attractive forces will depend on the
distance of the molecules apart. Suppose the attractive forces be such that they
vary inversely as the fourth power of the distance apart of the molecules, then,
when a gas expands from a volume Vi to a volume V2, W=a{pi—p2)} where a is a
constant.
If the molecular attraction / varies inversely as the fourth power of the mean distance
rj of the molecules apart, f=c/ri*, where c is a constant ; and if the gas expands until the
mean distance of the molecules apart becomes r^, the work done against molecular attraction
will be :
/ f.dr=c I — : Internal work= -( — — — |= — |— |
since r is linear, and the volume v of the gas will therefore vary as the cube of r, so that
Vi=br^^ and V2=br2^, where 6 is a constant. Again, since the temperature is nearly
constant, ^ji'i =^2^2 =" constant. Collecting the various constant terms under the symbol
a, it follows that the work done against molecular attraction when the gas expands from a
volume v^ to a volume Vg is aiPi—Pz)-
Consequently, the work done in overcoming the intermolecular attractive forces
will be proportional to the difference between the initial and final pressures of
the gas (temperature constant). This result is in harmony with observations. The
cooling ejSect actually produced when a gas expands by simple outflow in vacuo is
due to the absorption of heat equivalent to this work.
In later and more delicate experiments. Lord Kelvin (W. Thomson) and J. P.
Joule — 1852-62— forced a steady stream of gas under a pressure slowly along a
tube A, Fig. 21, in the direction of the
arrows, through small orifice, 0, where it
expanded against the pressure pi. To
avoid eddies in the gas, a porous plug was — "
actually used. For the sake of simplicity, '^ ' ^
suppose the tube AB has unit sectional Fig. 21.— J. P. Joule and W, Thomson's
area, and that it is made of some material Experiment,
which does not conduct heat away from the
gas. Two phenomena occur : (1) the gas is slightly heated by friction us it passes
through the orifice 0 ; and (2) the gas is cooled as it passes through 0 by doing work
against a pressure p^- Suppose a piston A, Fig. 21, moves from left to right so as
to drive a volume of air, Vg, at a pressure p^, into the compartment BC. The work
done on the gas is obviously p^v^. Similarly, the work done by the gas as it pushes
the piston from, say, B to 0 through a distance v-^, will be pii\. Hence, if the gas
obeys Boyle's law, we shall have PiVi=poV2, and there will be no loss in the internal
kinetic energy of the gas through overcoming internal attractions, and no variation
of temperature of the gas on the sides AB and BC. If, however, work be done
against molecular attraction during the expansion of the gas, the work of expansion
on the side BC will exceed the work of compression on the side AB. The work
VOL. I. 3 K
E
866 INORGANIC AND THEORETICAL CHEMISTRY
performed when the gas expands against intermolecular attractive forces diminishes
the kinetic energy of the gas, and this ceases to be sensible as heat ; accordingly,
the temperature of the gas is lowered owing to the loss of the kinetic energy of
the molecules of the gas itself. Hence, the gas on the side BO will be cooled
below the temperature of the gas on the side AB. The change of temperature
which occurs when a compressed gas expands adiabatically through a small
orifice is called the Joule-Thomson effect. In J. P. Joule and W. Thomson's
experiments, the temperature of carbon dioxide, nitrogen, oxygen, and air fell about
1° ; or, more exactly, the observed fall of temperature per atmosphere difference
of pressure was
Air. Carbon dioxide. Oxygen^ Nitrogen. Hydrogen.
0-208*' 1-005° 0-263° 0-249° —0-039°
SO that the phenomenon with hydrogen is reversed, there is a rise of temperature of
the gas on the side BG. This corresponds with Regnault's observation that the
product pv increases with hydrogen above a certain temperature ; if, however,
the experiment be conducted at a lower temperature, hydrogen gas behaves like
other gases and is cooled ; and at higher temperatures, other gases behave like
hydrogen.
According to J. Rose Innes,^ the measurements of Joule and Thomson can be
represented by a formula of the type: Fall of temperature=aT-i+6, where a
and h are constants such that for air, a=141*5, 6=0'697 ; for carbon dioxide,
a=2165-0, 6=4-98; and for hydrogen, a=64-l, 6=— 0*331. In the case of
hydrogen, the change of temperature will be zero when jr=194° K., or —79° C. This
means that the Joule-Thomson effect with hydrogen will change sign from heating
to cooling in the neighbourhood of —80°, and this is called the inversion tempera-
ture. The inversion temperature of helium is —240°.
The porous plug experiment shows that the fall of temperature which occurs
when the pressure falls from p to pi varies very nearly in the inverse proportion
to the square of the absolute temperature, and is approximately 0*25° per
atmosphere ; or, more exactly.
Fall of temperature =AI-=- J {p—pi)
where T is the temperature of the gas at the initial pressure p, and pi the final
pressure of the gas. The constant A is 4-0-276°, say J° per atm. for air ; +1-388°
for carbon dioxide ; and —0-049° for hydrogen. The magnitude of this constant,
and accordingly also the cooling effect, is greater the more the gas deviates from the
ideal gas laws — presumably because more work is done against intermolecular
attractive forces.
Examples.. — (1) If carbon dioxide, at 4 atm. pressure and 0°, on passing through a
porous plug falls to a pressure of one atmosphere, show that the temperature falls about
4-2.
(2) If air at 0° is driven through an orifice and at the same time falls 3-6 atm. in pressure,
show that the fall in temperature is 0-994°, and if the fall in pressure is 10 atm., the fall in
temperature is 2-76°.
The greater the pressure, the smaller the corresponding cooling effect per
atmosphere difference of pressure. Thus, E. Vogel & found for oxygen at 0°,
and an initial pressure p, calculated for a pressure difference p^Pi of one
atmosphere :
p ... 20 60 100 120 140 160 atm.
Cooling effect . 0-260° 0-225° 0-191" 0174° 0157° 0-139°
At about 300 atm. pressure and ordinary temperatures the cooling effect is nil.
THE KINETIC THEORY OF ATOMS AND MOLECULES 867
The observed results are represented by the following modifications of Thomson's
equation :
/273\2
Cooling effect =(0*268— 0-00086;>)(^—2)i)(-^j
The lower the temperature, the greater the cooling efEect. Thus, W. P. Bradley
and C. F. Hale found that when ^=204 atm. and pi=l atm..
Temperature
. 0°
-20°
-40°
-60°
-80°
-90'
Cooling effect
. 44-6°
52-1°
61-r
72-5°
88-2°
99-2'
If externa] work W be performed on unit mass of gas compressing it isothermally,
the intermolecular attractions will do a quantity of internal work w which will
appear as heat. Hence, during an isothermal compression the total heat Q^
taken from the gas, will be Q=W-\-w ; and the total energy of the gas has been
diminished by an amount w. Consequently, if the compressed gas be allowed to
expand back to its original volume, without receiving energy from some outside
source, internal work equivalent to w will have to be performed at the expense of
the kinetic energy of the gas, and the temperature of the gas will accordingly fall.
If w is equivalent to the heat abstracted from a gram of gas and which produces
a fall of temperature ^°, then if the heat be taken from half the gas, the fall of
temperature will be 2^°, and if from an nth. part of the gas, there will be a fall
of nd\
If the specific heat of air at constant volume be 0*177, and if, for convenience,
it be assumed that the specific heat is constant, it will require 210 X0'177=37 cals.
to lower the temperature of a gram of air from 20° to —190° ; further, if the latent
heat of vaporization be 50 cals. per gram, and a gram of the air at —190° be liquefied,
it will require 37 -{-50=87 cals. to cool a gram of air from 20°, and liquefy it at
—190°. According to Joule and Thomson's experiment, the expansion of a gram
of compressed air will cool it J° per atmosphere ; if the gram of air expands from a
pressure of 160 atm., it will therefore be cooled 40°. This corresponds with 7 Cals.
of internal work per gram ; and accordingly, since 87 cals. of heat are required in
order to cool a gram of air from 20° and to liquefy it, it follows that under ideal
conditions only ^-th to -j-th of the gram of air can be liquefied by expansion from
a pressure of 160 atm. when the rejected air has the same temperature as the
compressed air. If m-i and m^ denote respectively the masses of the initial and the
rejected air, then mi—m^ will denote the mass of air which has been liquefied
during the expansion. Let W denote the external work done during isothermal
compression, and w the work due to intermolecular attractions ; then, if Qj and Q2,
respectively denote the heat equivalents of W and w,
mi— ^2 w
This means that the ratio of the mass of liquefied air to the mass of rejected air is
equal to the ratio of the work due to intermolecular attraction to the external work
of isothermal compression. Since the cooling due to expansion of air from a pressure
of 160 atm. is equivalent to w=7 cals. of internal work per gram, and the work W
required for the isothermal compression of a gram of air to 160 atm. is equivalent
to Tf=9S cals., it follows that ivIW=y4- This means that under ideal conditions,
one-fourteenth of the work performed by the isothermal compression of air at 20°
to a pressure of 130 atm., and its subsequent adiabatic expansion to the original
pressure, is due to intermolecular attractions. In any system of liquefying gases
by utilizing the slight cooling efiect produced when internal work is done against
intermolecular attractions, the test of the efiiciency of the process is how near
,^ . ^ . Liquid obtained
Maximum efficiency = — -^— — — ~
Work done
868 INORGANIC AND THEOEETICAL CHEMISTRY
approaches the theoretical ; for the air costs virtually nothing. Special refrigerants
may enable a large percentage of compressed air to be liquefied ; reducing the
pressure for the expanded air below one atmosphere will give a bigger cooling
effect, etc. ; but all this requires the application of more work, and it is for the
chemical engineer to find if the different factors increase or diminish the efficiency
fraction.
As previously indicated, it is sometimes convenient to use an idealized gas as a
limiting case or standard of comparison in the theoretical study of molecular forces,
etc. Such a gas— called a perfect or ideal gas — (i) obeys Boyle's and Charles'
laws for all pressures and temperatures ; and (ii) it suffers no change of temperature
when it is allowed to expand into an evacuated vessel. In this sense, perfection
is not an objective quality of any particular gas, but it rather denotes a favourite
and familiar fiction whose sole justification is that it facilitates the general in-
vestigation of the properties of gases.
References.
^ W. Cullen, Essays and Observations, Physical and Literary, Edinburgh, 2. 145, 1765.
« L. P. Cailletet, Ann. Chim. Phys., (5), 5. 138, 1878 ; Compt. Jfend., 85. 1210, 1270, 1879.
3 J. L. Gay Lussac, Mem.. d'Arcueil, 1, 1807 ; Gilbert's Ann., 30. 249, 1808 ; J. P. Joule, Phil.
Mag., (3), 26. 369, 1845 : Lord Kelvin (W. Thomson) and J. P. Joule, Phil. Trans., 143. 357,
1853 ; 144. 321, 1854 ; 152. 579, 1862.
* J. Rose Innes, Phil. Mag., (5), 45. 227, 1898 ; (5), 50. 251, 1900 ; (6), 2. 130, 1901 ; (6),
6. 353, 1903 ; (6), 15. 301, 1908 ; A. W. Porter, Phil Mag., (6), 11. 554, 1906.
5 E. Vogel, Ueber die Temperaturverdnderung von Luft und Sauerstoff beim Stromen durch
eine Droselstelle bet 10° C. und Drucken bis zu 150 Atm., Miinchen, 1910 ; W. P. Bradley and
C. F. Hale, Phys. Rev., 20. 258, 1909.
§ 26. The Liquefaction of Gases
If the earth should all of a sudden find itself placed in very cold regions, the water
which now forms our rivers and seas, and probably the greater number of liquids which
we know would be transformed into solid mountains and into very hard rocks. The air,
according to this hypothesis, or at least a portion of the aeriform substances which compose
it, would doubtless cease to exist in the condition of invisible fluids, and through the
absence of a sufficient degree of heat, it would revert to the liquid state, and this change
would produce new liquids of which we have no idea.- — A. L. I/AVoisier (1789).
The conversion of liquids into vapours, and the reverse condensation of vapours
into liquids during distillation must have attracted attention before the Christian
era, and have suggested the possibility of the condensation of aeriform fluids in
general. M. A. Lucanus, in his satirical Vera historia, written in the second century,
said that the inhabitants of the moon drink air squeezed or compressed into goblets,
for this operation produces a liquid like dew. There are also some poetic references
to liquid air in Virgil (c. 50 B.C.). This, of course, does not mean that the liquid in
question was known to the Greek and Roman poets. i Some time before the composi-
tion of air was known, H. Boerhaave, in his Elementa chemiw (Lugduni Batavorum,
1732), described an attempt to condense air to the liquid state by artificial cold,
although he succeeded in producing no other result than the condensation of the
moisture in the air. He therefore concluded :
We may fairly assert that the fluidity of air, in all the large compass from the most
rarefied to the most compressed, remains without alteration ; and that therefore it is neither
capable of being solidified by the int€;nscst cold, nor the greatest degree of compression.
John Dalton saw as clearly as A. L. Lavoisier the probable result of subjecting
gases to a great enough cold and pressure. He said :
THE KINETIC THEORY OF ATOMS AND MOLECULES 869
There can scarcely be a doubt entertained respecting the reducibility of all elastic
fluids of whatever kinds mto liquids ; and we ought not to despair of effecting it in low
temperatures, and by strong pressures exerted upon the unmixed gases.
Among the early experiments there is M. van Marum's 2 liquefaction of ammonia
under a pressure of about 3 atm. ; L. B. G. de Morveau, and A. F. de Fourcroy and
L. N. Vauqueiin's liquefaction of the same gas by cooling with a freezing mixture to
about —40°, and their nugatory attempt to liquefy hydrogen chloride, hydrogen
sulphide, and sulphur dioxide by similar means. G. Monge and L. Clouet reported
the liquefaction of sulphur dioxide by simple cooling. In some cases, the drying
of the gases must have been so imperfect that the condensation of moisture or of
a solution of the gas in water was mistaken for the liquefaction of gas itself.
T. Northmore's work ^ on the liquefaction of gases by compression is perhaps
the most important ef the earlier experiments. He succeeded in liquefying chlorine,
hydrogen chloride, and sulphur dioxide. When the attempt was made with carbon
dioxide, the receiver unexpectedly burst with violence. In 1823, M. Faraday *
liquefied a number of gases by the joint effect of pressure and cold. The gases
were generated in one leg of a hermetically sealed glass V-tube, and condensed by
cooling with the other leg of the V-tube in a freezing mixture. The increasing
pressure developed by the gas enabled sulphur dioxide, hydrogen sulphide, carbon
dioxide, nitrogen monoxide, cyanogen, and ammonia to be condensed ; and
H. Davy likewise succeeded in condensing hydrogen chloride in this manner.
D. Colladon ^ attempted to liquefy air by confining it in a stout glass tube sealed
with mercury so arranged that more mercury could be forced into the tube by
hydraulic pressure. He worked at —30° and produced a pressure of 400 atm.,
but he did not succeed in condensing the air. M. Thilorier, in 1834, generated
carbon dioxide in a wrought-iron vessel and used the increasing pressure as the gas
accumulated in the vessel for its condensation in an adjacent iron vessel cooled by
a freezing mixture. M. Thilorier's apparatus was therefore a kind of Faraday's
tube of large capacity, and made of cast iron. One of the cylinders burst before a
class, and its fragments were scattered about with tremendous force ; it cut off both
legs of the unfortunate operator, M. Hervy, and the injury was followed by death.
This showed that cast iron was not a safe metal for the chambers of an apparatus
for condensing gases under pressures, M. Thilorier succeeded in liquefying and
solidifying carbon dioxide, and he also prepared a freezing mixture of solid carbon
dioxide and ether which enabled a temperature of —110° to be attained. Modifica-
tions of M. Thilorier's apparatus were employed by R. Addams, J. K. Mitchell,
M. Faraday, J. 0. Natterer, etc.^ In all these systems the gases were liquefied by
cold and pressure acting jointly or alone. A few gases — like hydrogen, oxygen,
nitrogen, air, etc. — resisted all attempts to liquefaction in this way, and the re-
fractory gases were called permanent gases to distinguish them from those more
condensable. The term lost its main significance when L. P. Cailletet and R. Pictet
succeeded in liquefying oxygen and hydrogen about 1877, and C. von Linde succeeded
in manufacturing liquid air on an industrial scale.
The success of the attempts to liquefy gases has been largely dependent upon
the methods for producing great cold. About two centuries ago, 1724, G. D.
Fahrenheit congratulated himself that no one could produce a lower temperature
than he had done by mixing together snow and salt ; and he accordingly made
this temperature the zero of his thermometer — Fahrenheit's thermometer. In
his wildest dreams, Fahrenheit is not likely to have imagined that temperatures
400° F. below the point he evidently thought a very creditable achievement, would,
in later years, be comparatively easily attained. When, a few years afterwards,
Fahrenheit succeeded in reducing the temperature a few degrees below his own
zero, H. Boerhaave said that the result was incredible, and asked : " What mortal
man could ever have thought of it ? " The different methods which have been
used for producing low temperatures include :
(1) Methods involving the use of freezing mixtures. — If ice or snow be mixed
870
INORGANIC AND THEORETICAL CHEMISTRY
with salt and water above the eutectic temperature, the water will tend to dissolve
salt, and the ice will melt and make the solution capable of dissolving more salt.
If the liquid be thermally insulated, the latent heat required for the fusion of the
ice is abstracted from the liquid, and the temperature falls until the solution has
the eutectic composition. The temperatures obtainable by the use of mixtures of
ice and salts — the so-called freezing mixtures — range as low as —55° — the eutectic
temperature of calcium chloride and water. The following represent some results
actually obtained with mixtures of one part of snow with
iSodium carbonate (cryst.) — ith part
Potassium chloride — Jrd part .
Sodium chloride — Jrd part
Calcium chloride (cryst.) — 2 parts
Fall of temperature
from 0° to
-2°
. -12°
. -18°
. -42°
Expansion
U quid Air — ^^^
(2) Cooling by the adidbatic expansion of cold comjpressed gases. — C. W. Siemens
(1857), E. Solvay (1885), and F. Windhausen (1892) 7 obtained patents for producing
very low temperatures by the expansion of air in a suitable cylinder, and using the
cooled expanded air to cool the incoming air by a kind of recuperation of the cold.
In L. P. Cailletet's apparatus, gas is confined in a stout glass tube by mercury and
the mercury is pumped by a hydraulic press into a suitable reservoir so as to com-
press the gas. L. P. Cailletet designed
B his apparatus so that the pressure con-
fining the gas could be quickly relieved,
ill order that successive experiments
could be made with the same mass of
gas. By chance, on relieving the
pressure under which some gaseous
acetylene was confined, L. P. Cailletet
noticed a thick mist — un hrouillard
epais — developed in the tube, and he
at first supposed it was produced by
the condensation of moisture or im-
FiG. 22.— Claude's Process for the Liquefaction of purities in the gas. On testing this
Air (Diagrammatic). assumption with purified acetylene and
also nitrous oxide, the sudden expan-
sion of the compressed gas still gave the mist, and he then attributed the phenomenon
to liquefaction produced by the intense cooling of the expanding gas. L. P. Cailletet
then er^ployed greater pressures with the so-called permanent gases. On Dec.
2nd, 1877, he wrote to H. St. C. DeviJle :
I hasten to inform you, and you first without losing a moment, that I have liquefied
this day both carbon monoxide and oxygen. I am perhaps wrong in saying liquefied,
because at the temperature I obtained by evaporating sulphur dioxide, that is at —29°
under 200 atm. pressure, I did not see any liquid, but a dust so dense that I was able to
infer the presence of a vapour very close to its point of liquefaction.
He added that no sign of the liquid dust was detected with hydrogen treated in a
similar manner, but in a later experiment the mist was obtained working from
300 atm. pressure at —28°. Cailletet's apparatus was used by J. Ogier^ for liquefying
silicomethane ; by L. Ilosvay de N. Ilosva, for liquefying carbonyl sulphide ; by
J. Ansdell in studying liquid acetylene, and hydrogen chloride ; by C. Vincent and
J. Chappius for methyl and ethyl chlorides, the methylamines, etc. ; by P.
Hautefeuille and J. Chappius, L. Troost, and A. Ladenburg for liquid ozone. G.
Claude has employed the principle for the continuous liquefaction of air.
The purified air at about 40 atm. pressure is driven through an inner tube A, Fig. 22,
to an expansion apparatus, B^; and the cooled and expanded air circulates upwards about
THE KINETIC THEORY OF ATOMS AND MOLECULES 871
the tube A. The cooled air passes from the liquefier, about the inlet tube A, and is thence
returned via A to the expansion apparatus. The compressed air in the liquefier is thus
progressively cooled by the expanded gas from the pump A, until liquefaction is attained.
If the temperature of the gas passing to the pump B is too low, there is but a very slight
cooling of the gas, and this is aggravated by the increased specific heat of the gas at a
low temperature. Consequently, the temperature of the gas entering the pump B is not
allowed to fall too low. The liquefied gas is run off at the cock C. Whenever necessary,
petroleum ether is used as a lubricant for low temperatures. The advantages claimed
for this process are the rapid liquefaction of the air at comparatively low pressures.
(3) Methods involving the rapid evaporation of volatile liquids. — In 1755,
W. Cullen froze water by its own rapid evaporation, and in 1862, Carre exhibited at
the International Exhibition, a machine for manufacturing ice in which the cold
was produced by the rapid evaporation of liquid ammonia. Ice is made by this
process to-day. In 1835, M. Thilorier showed that the rapid evaporation of liquid
or solid carbon dioxide mixed with ether gave a temperature of —100°, and
by this means, M. Faraday obtained temperatures as low as —110°. In 1878,
J. 0. Natterer obtained a temperature of —140° by the rapid evaporation of a
mixture of liquid carbon dioxide and nitrous oxide.
A. A. B. Bussy ^ also liquefied sulphur dioxide by cooling, and, on allowing the
liquid to evaporate rapidly, produced temperatures which enabled him to liquefy
chlorine and ammonia, and to solidify cyanogen. These operations introduce the
work of K. Pictet.io In order to lessen the risk of compressing gases at enormous
pressures, Pictet cooled a gas, B, below its critical temperature by the rapid evapora-
tion of another liquefied gas. A, and used the liquefied gas B to cool a third gas, C,
and so on. This is called the cascade method o£ Uquefying gases. In one series
of experiments, liquid sulphur dioxide was allowed to evaporate around a system
of tubes containing carbon dioxide, which then liquefied under a feeble pressure ;
the liquid carbon dioxide was allowed to evaporate in a similar way about a
tube containing oxygen. When the valve closing the tube containing the oxygen
was opened, Pictet said that he noticed. oxygen escaped in the form of a transparent
jet surrounded by a white cloud which he took to be solid oxygen. As a result,
R. Pictet telegraphed to the French Academy of Sciences on Dec. 22nd, 1877 :
Oxygene liquefie aujourd'hui sous 320 atmospheres et 140*' defroid par acides sulphureux
et carbonique accoupMs.
This telegram and L. P. Cailletet's letter to J. B. A. Dumas were read at the
meeting of the French Academy on Dec. 24th, 1877. Some of R. Pictet's de-
scriptions of liquid oxygen and also of liquid hydrogen, which he claimed to have
prepared, have not been confirmed by later work. At the same meeting, J. C.
Jamin pointed out that it was still necessary to assemble into a real liquid the
impalpable mist which had been momentarily obtained by Pictet and Cailletet ;
and some time afterwards, S. von Wroblewsky pointed out that the subject required
developing so that liquid oxygen could be poured as readily as liquid ethylene.
He said :
It is my conviction that the thing will be successfully carried out only when we return
to R. Pictet's method, and by cycles of various liquefied gases make a cascade of tempera-
tures whose last step will produce the stream of liquid oxygen.
S. von Wroblewsky and K. Olschewsky n cooled the condensing tube of an apparatus
like that of L. P. Cailletet by means of ethylene evaporating under reduced pressure
and cooled by solid carbon dioxide. As a result, they were able to telegraph to the
French Academy on April 7th, 1883 :
Oxygene liqu^fi^, completeiiient liquide, incolore comme I'acide carbonique. Vous
recevrez ime note dans quelques jours.
Even here comparatively small amounts of liquid were obtained, and it was
872
INORGANIC AND THEORETICAL CHEMISTRY
not until 1895 that C. von Linde was able to liquefy the so-called permanent
gases — air and oxygen on a commercial scale. Linde's apparatus was based on
quite a different principle. H. K. Onnes 12 at Leyden used the cascade process with
methyl chloride, down to —90° ; ethylene down to —160° ; and oxygen down to
—217°. This critical temperature of hydrogen and helium are so much below
the lowest temperatures obtained by the evaporation of gases of higher boiling
point that the cascade method is inapplicable for their condensation.
(4) Cooling hy the Jouh-Thomson effect. — When a gas, cooled by passing through
a small orifice, is made to circulate around the tube leading the compressed gas
to the orifice, the gas passing to the orifice is cooled, and on passing through the
orifice is cooled still more. By continuing the cycle, it follows that the tempera-
ture can be reduced indefinitely low, or the gas liquefies. For example, air at —20°
is cooled to —36° by expanding in the orifice from a pressure of 50 atm. to one
atmosphere ; if this cooled air is compressed and again expanded from —36°, the
temperature drops to —54°, then to —101°, —136°, and finally —190°, when the
expanding air condenses. The theory was known for almost half a century before
it was realized in industrial practice. The so-called self -intensive or cumulative
systems for cooling gases were elaborated
by C. von Linde,i3 J. Dewar, W. Hampson,
and C. E. Tripler between 1894-5.
4ir enters
at 200 atm,
pressure
/J ir issues
at 20 atm.
pressure
The idea will be understood after an ex-
amination of Fig. 23. The air to be liquefied
• — freed from carbon dioxide, moisture, organic
matter, etc.- — enters the inner tube of con-
centric or annular pipes, A^ under' a pressure
of about 200 atm. This tube is hundreds of
yards long and coiled spirally to economize
space. By regulating the valve O, the com-
pressed air is allowed to suddenly expand in
the chamber Z) to a pressure of about 50 atm.
The air thus chilled passes back through the
tube B which surrounds the tube A conveying
the incoming air. The latter is thus cooled
still more. The gas, at 50 atm. pressure, passes
along to the pumps, where it is returned with
more air to the inner tube. In this manner, the
Liquid /fir incoming air at 200 atm. pressure is cooled more
and more as it issues from the jet O. Finally,
when the temperature is reduced low enough.
Fig. 23.— Linde's Apparatus for the Lique- drops of liquid air issue from the jet, and
faction of Air (Diagrammatic). collects in the receiver. The tubes must all be
packed in a non-conducting medium — wool,
feathers, etc. — to protect them from the external heat. From a preceding discussion,
the work required to compress a gas from atmospheric pressure to 200 atm. pressure is
proportional to log 200, while the work required to compress the gas from 50 to 200 atm.
is proportional to log 2. It is therefore found to be more economical and efficient to allow
the gas to expand from 200 to 50 atm. pressure rather than from 200 to one atm. pressure,
when the expanded gas is returned to the compressor. Several improvements on Linde's
form of liquefier have been devised, but the main principles are well illustrated by the
original apparatus.
In 1898, J. Dewar liquefied hydrogen by a similar method, but the gas was
first cooled below its inversion temperature, —80°, in liquid air before subjecting
it to the Joule-Thomson process. By evaporating liquid hydrogen under reduced
pressure, a temperature of —259°, or 14° K., was obtained. The inversion tempera-
ture of helium is —240°, or 33° K., and H. K. Onnes, in 1908, succeeded in liquefying
helium by cooling it below its inversion temperature in liquid hydrogen, and then
cooling it still further by the Joule-Thomson process. Helium boils at 4*29° K.
under atmospheric pressures ; and, by evaporating liquid helium under reduced
pressure, H. K. Onnes obtained a temperature 1*48° K., or —271*6°, that is, within
less than two degrees of absolute zero.
THE KINETIC THEORY OF ATOMS AND MOLECULES 873
Preserving liquid air. — 'TJiere is a far greater difference between the temperature of
liquid air (about —190°) and ordinary atmospheric air, than between the temperature of
ice and boihng water. The preservation of liquid air is thus a far more difficult problem
than would be involved in preventing cold water boiling away
while surrounded by a steam jacket at 200''. James Dewar (1892) i*
solved the problem by keeping the liquid air in the double glass
flasks with an evacuated space between the inner and outer walls.
Vessels similar in principle are said to have been used by L. J. G.
VioUe in 1882, and by A. d'Arsonval in 1887. Glass is a poor con-
ductor, and a vacuum is a non-conductor. Hence, the liquid in the
inner vessel can receive heat only from above, and by radiation.
J. Dewar also silvered the glass walls of the evacuated space so as
to reduce the effects of radiant heat. Still air is a very bad conductor
of heat, so that the open end of the vessel is plugged lightly with
cotton wool in order to reduce the ingress of heat from outside to a
minimum. In this way, liquid air is transported by rail, etc., with
a surprisingly little loss. En passant, a similar principle is utilized
in the so-called thermos flasks, which will not only keep the contained liquids cool, but
also retard the cooling of hot liquids ; and J. Dewar claimed that he utilized the principle
of the vaevium vessel as an insulator in calorimetric experiments in 1874.
24. — De war's
Flasks.
References.
1 M. A. Lucanus, Vera historia, 2. 89 ; Virgil, Georgics, 1. 404 ; ^Eneid, 6. 202.
2 M. van Marum, Gilbert's Ann., 1. 145, 1799 ; A. F. de Fourcroy and L. N. Vauquelin,
Ann. Chim. Phys., (1), 29. 281, 1799 ; L. B. G. de Morveau, ih., (1), 29. 290, 297, 1799.
3 T. Northmore, Nicholson's Journ., 12. 368, 1805 ; 13. 233, 1806.
« M. Faraday, Phil. Trans., 113. 189, 1823.
5 R. Pictet, Ann. Chim. Phys., (5), 13. 288, 1878 ; M. Thilorier, ih., (2), 60. 427, 1835.
^ R. Addams, Brit. Assoc. Rep., 70, 1838 ; J. K. Mitchell, Amer. Journ. Science, (1), 35.
346, 1839 ; G. Aime, Ann. Chim. Phys., (3), 8. 275, 1843 ; M. Faraday, Phil. Trans., 135. 155,
1845 ; J. 0. Natterer, Journ. prakt. Chem., (1), 31. 375, 1844 ; Pogg. Ann., 62. 132, 1844 ; Sitzber.
Akad. Wien, 5. 351, 1850 ; 6. 557, 1851 ; 12. 199, 1854.
' C. W. Siemens, Proc. Inst. Civ. Eng., 68. 176, 1882 ; Brit. Pat. No., 2064, 1857 ; E. Solvay,
Compt. Rend., 121. 1141, 1895; Brit. Pat. No., 13466, 1885; F. Windhausen, ib., 14851,
1892 ; L. P. Cailletet, Ann. Chim. Phys., (5), 15. 138, 1878 ; Comjit. Rend., 85. 1210, 1270, 1879.
« J. Ogier, Compt. Rend., 88. 236, 1879 ; L. Troost, ib., 126. 1781, 1898 ; P. Hautefeuille and
J. Chappius, ib., 94. 1249, 1882; C. Vincent and J. Chappius, ib., 100. 1216, 1885; 103. 379,
1886 ; Journ. Phys., (2), 5. 58, 1886 ; L. Ilosvay deN. Ilosva, Bull. Soc. Chim., (2), 37. 294, 1882 ;
J. Ansdell, Proc. Roy. Soc, 29. 209, 1879 ; Chem. News, 41. 75, 1880 ; A. Ladenburg, Ber., 31.
2508, 1898 ; G. Claude, L'air Hquide, Paris, 1909.
9 A. A. B. Bussy, Ann. Chim. Phys., (2), 26. 63, 1824.
i» R. Pictet, Compt. Rend., 85. 1214, 1220, 1877; 86. 37, 106, 1878; Arch. Science Nat.
Geneve, 61. 16, 1878.
" S. von Wroblewsky and K. Olschewsky, Compt. Rend., 97. 1553, 1883 ; Wied. Ann., 20,
243, 1883 ; S. von Wroblewsky, ib., 20. 860, 1883 ; 25. 371, 1885 ; 26. 134, 1885 ; K. Olschewsky.
ib., 31. 58, 1887 ; Phil. Mag., (5), 39. 188, 1895 ; J. Dewar, ib., (5), 39. 298, 1895 ; (5), 18. 210,
1884 ; (5), 34. 205, 326, 1892 ; (3), 36. 328, 1893 ; Proc. Roy. Soc, 63. 256, 1898.
12 H. K. Onnes, Comm. Phys. Lab. Leiden, 54, 1899 ; 87, 1903.
13 C. von Linde, Ber., 32. 925, 1899; Wied. Ann., 57. 329, 1896; Zeit. Kdlteind., 4. 23,
1897; German Pat., D.R.P. (June 5) 88824, 1895; M. Schrotter, Zeit. Ver. deut. Ing., 39. 1157,
1895; C. E. Tripler, Eng. Neivs, 39. 246, 1898; H. Lorenz, Zeit. Kdlteind., ^.23, 1897; R. Pictet,
ib., 7. 1, 1903; R. A. Hehl, Flussige Luft, Halle, J 901 ; W. Hampson, Brit. Pat. No. (May 23),
10165, 1895 — this patent antedates that of C. von Linde by thirteen days, but W. Hampson in
the provisional specification alludes only to the usual cycle of expansion of cooled and compressed
air, and referred to expansion by simple outflow only in the final specification ; T. O'C. Sloane,
Liquid Air, New York, 1899; J. Dewar, Proc. Roy. Inst., 15. 133, 1895; Phil. Mag., (5), 18.
210, 1884. For the patent literature, see 0. Kausch, Zeit. kompr. verfl. Gase, 5. 172, 187, 1902 ;
6. 33, 1903.
1* J. Dewar, Proc. Roy. Inst., 14. 1, 1893; A. d'Arsonval, Cwnpt. Rend. Soc. Biol, (8), 5.
136, 142, 1888.
Fia . 25. — Composition of the Liquid and
Vapour from Liquid Air (E. C. C.
Baly).
874 INOKGANIC AND THEOEETICAL CHEMISTKY
§ 26. The Manufacture of Oxygen and Nitrogen from Liquid Air
When a thing is possible according to theory and only practical difficulties oppose its
realization, it is infinitely probable that those difficulties are not insurmountable. —
G. Claude.
When liquid air evaporates, the nitrogen, boiling at —195*5°, is more volatile
than the oxygen, boiling at —182 "5°, and escapes first so that the gas which comes
from the liquid during the earlier stages of the evaporation contains so little oxygen
that it will extinguish a lighted taper'; as
evaporation continues the liquid becomes richer
and richer in oxygen until finally almost pure
oxygen separates. If oxygen gas be bubbled
through liquid air (— 193*5°), the bubbles of
gas escaping contain 93 per cent, of nitrogen.
The oxygen is condensed from the rising
bubbles, and the more volatile nitrogen takes
its place. For equilibrium, there is a definite
relation between the composition of the liquid
mixture and of the rising vapour. This has
been investigated by E. C. C. Baly (1900) ,i and
the results are illustrated by the curve, Fig. 25,
which shows for each proportion of oxygen in
the mixed liquid, the corresponding proportion
in the vapour which is necessary for equili-
brium. Thus, if the liquid contains 30 per
cent, of oxygen (and 70 of nitrogen), the vapour
will contain 10 per cent, of oxygen (and 90 of nitrogen). When oxygen is bubbled
through liquid air containing 21 per cent, of oxygen (and 79 of nitrogen), then, for
equilibrium, oxygen will condense and nitrogen evaporate until the vapour contains
7 per cent, of oxygen (and 93 of nitrogen). E. C. C. Baly's measurements of the
percentage composition of liquid air and of the vapour in equilibrium at difl^erent
temperatures are graphed in Fig. 26. A horizontal line across the curves at any
assigned temperature shows the composition of liquid and vapour when in equili-
brium. The composition of liquid and vapour in
the rectifying columns or scrubbers of Linde's or
Claude's apparatus approximate to the values
indicated by these curves. These important
principles must be clear before the modern
method of separating oxygen and nitrogen from
liquid air can be understood.
Linde's process. — J. Dewar noticed that when
liquid air boils, the more volatile constituent —
nitrogen — is given ofi preferentially during the
Fio. 26— The Percentage Composi- ^^st stages of the evaporation, while the residual
tion of Liquid Air in Equilibrium liquid becomes progressively richer and richer in
with its Vapour at Different oxygen. In 1893, J. .H. Parkinson 2 patented a
Temperatures. proposal to utilize this fact for the commercial pro-
duction of oxygen, but his apparatus was not
satisfactory. It was not until the development of C. von Linde's process 3 in 1895
that it became practicable to manufacture oxygen from liquid air.
A diagrammatic sketch of C. von Linde's apparatus (1895) is indicated in Fig. 27.
Purified air is compressed to about 200 atm., and driven along a pipe which divides
at Af Fig. 27, into two streams and then passes down the interior tubes of a double set of
annular or concentric pipes similar to the worm tube, Fig. 23. The two inner tubes finally
unite into one single pipe, B. The air then passes through a spiral, S, via the regulating
valve R, and finally streams at C into the collecting vessel. The action is here similar to
I'
|84
^
:/
/>
y
/
/
nO*"
y
/
>1
f
/
/
/
A
k'
/
y
\^^
/
Pi
y ce
nt 0
Yo.,
/gen
0 2
0 3
« K
10 5
A £
A 7
0 8
0 c
•0 KM
THE KINETIC THEORY OF ATOMS AND MOLECULES
875
Oxygen, at ___
fS'atm.pressune "^
,-_ Cdmpressed €ur
f;:=^20datm<
J (A pressure
^ Nitrogen, at
^'^'iS'atm.pressuTe
that described in the process for the liquefaction of air. Fig. 23. After a time, the air is
lifiuefied in the collecting vessel, about the spiral S. The more volatile nitrogen boils
off more rapidly than the oxygen. Hence, a gas rich in nitrogen passes up one of the two
annular outer pipes as indicated on the left of Fig. 27. The liquid rich in oxygen is kept
at a constant level by means of the valve, and thus the rate at which the liquid air in the
collecting vessel is allowed to boil is also regulated. The oxygen passes from this tube on
the right of Fig. 27 along the outer annular pipe, and finally emerges from the apparatus,
whence it is pumped into cylinders, etc., for use. If the valves are all properly regulated,
the inrushing air is cooled by the counter currents of oxygen and nitrogen. The two latter
gases pass along the tubes as indicated in the diagram. The tubes, etc., are all well insulated
with non-conducting materials- — ^feathers, wool, etc.
In the 1895 apparatus for extracting oxygen from the air, the product was not
much more than 60 per cent, purity, and it was not until C. von Linde employed
the rectifying tower in 1902, that a purer gas was obtained. In C. von Linde's
rectifying column, there is a downward stream of liquid air which starts at about
80° K. ; and an upward stream of gas which is rich in nitrogen, and which starts
at about 91 "5° K. The percolation of the reverse streams through the tower changes
the descending current into oxygen of 98-99 per cent, purity, and the ascending
current into nitrogen of 93 per cent, purity. The
reason why the nitrogen is not richer than 93 per cent,
purity will be clear from Fig. 26. The ascending
vapours are always richer in oxygen than corresponds
with the descending liquid, there is a progressive
condensation of oxygen from the vapour, and an
evaporation of nitrogen from the liquid, to enable
that equilibrium to be established. By this pro-
cess, oxygen can be obtained as pure as is com-
mercially desired, but the escaping nitrogen contains
over 7 per cent, of oxygen. The original apparatus
has been much improved in recent years, principally
with the object of preparing a purer nitrogen for use
in the cyanamide process ; but, as G. Claude has
said : " We must salute in this apparatus the arche-
type of all the later machines, for it was the first to
demonstrate that the manufacture of oxygen from
liquid air is commercially possible." The rectifica-
tion of the nitrogen is effected by condensing the 93
per cent, product, and rectifying it in another tower.
A product with two per cent, oxygen is thus ob-
tained. The supplementary rectification by means of a liquid rich in nitrogen
was patented by R. J. Levy and A. Helbrouner in 1903.^
It must not be assumed that two gases which exert no chemical or physical
action on one another, can be separated from a mixture without the expenditure
of energy. Each gas in the mixture exerts its own partial pressure, and the total
work required to separate the components of a mixture of gases which exert no
chemical or physical action on one another is equal to the sum of the work of
isothermal compression of each of the constituent gases from its partial pressure
in the mixture to its final pressure. Lord Rayleigh ^ dealt with the converse
problem in 1875 by showing that mechanical work is performed by the simple
admixture of two gases without the exercise of chemical affinity.
G. Claude's counter-current process of rectification. — ^An improvement on
Linde's process, by G. Claude (1903), enables practically pure oxygen and nitrogen
to be obtained in a simple and effective manner. It is based on the fractional
condensation of compressed air. The liquid just formed is drained back through
a condenser so that it is scrubbed in contact with the gas which is being condensed.
The liquid collects in two portions — one contains about 50 per cent, of oxygen,
and the other is almost wholly nitrogen. The two liquids are separately rectified.
liquid /Iw
Fia. 27. — -Lindo's Apparatus for
the Separation of Oxygen and
Nitrogen from Liquid Air
(Diagrammatic).
876
INORGANIC AND THEORETICAL CHEMISTRY
A diagrammatic sketch of Claude's apparatus is shown in Fig. 28. The cooled and ]jurified
air enters the lower part of the apparatus at a pressure of about 5 atm. and rises through
a series of vertical pipes P surrounded by liquid oxygen, where it is partially ]ic[uefied.
The liquid containing about 47 per cent, oxygen and 53 per cent, of nitrogen drains into
the lower vessel A. The vapour which has survived condensation enters B, and then
descends through a ring of pipes C arranged concentrically about the set previously de-
scribed. Here all is liquefied. The liquid which ultimately collects in this vessel /) is
very rich in nitrogen. The pressure of the vapour
gg^ in the central receptacle forces the liquid nitrogen
Nitrogen ^o enter the summit of the rectifying colunm E,
^ and the liquid, containing 47 per cent, of oxygen,
is likewise forced to enter the rectifying column at ,
^clifying F lower down. The pressure and rates of flow are ^
~/i»/./«.« regulated by the cocks RR\ The liquid nitrogen "
is 3° or 4° lower in temperature than the liquid ^
rich in oxygen. Nitrogen evaporates from the I
down-streaming liquid, and oxygen condenses from
the up -streaming gases. The heat supplied by
the condensation of oxygen helps on the evapora-
tion of nitrogen. Consequently, the descending
liquid gets progressively richer and richer in
oxygeji, and the ascending gases richer in nitro-
gen. The liquid oxygen drains into the recep- ,
tacle G, and is there evaporated by the latent i
heat of the gases condensing in the tubes.
Finally, oxygen containing from 2 to 4 per cent,
of nitrogen passes from the oxygen exit, and
nitrogen containing 0*2 to 1 per cent, of oxygen
Cb^^or ssed ®s°*P®^ ^* *h® *op ^^ *^® rectifying column.
Aif^.enters
These processes enable oxygen and nitrogen
to be extracted from atmospheric air com-
T^ «o ^ ^, J , * ^ i^ .1- paratively cheaply — one ton of coal for driving
Fig. 28.— G. Claude s Apparatus for the fu • i. ■ -j^x •?
Separation of Oxygen from Uquid Air ^^^ compression apparatus is said to furnish
(Diagrammatic). one ton of oxygen and four tons of nitrogen.
As C. von Linde has said, " the heat is elimi-
nated from the air exclusively by the expenditure of internal work." The work
required for the liquefaction of these gases is solely spent in separating the
molecules of the gases from one another against their intermolecular attractions,
and in compressing the gases from their partial pressure in the original mixture
to their final pressure.
LiquLd_
CtKygen
Liquid
Nltroger^
(80%N) \
References.
1 E. C. C. Baly, PUl. Mag., (5), 49. 517, 1900.
2 J. H. Parkinson, Brit. Pat. No. 4411, 1892.
3 C. von Linde, Wied. Ann., 57. 329, 1896; Zeit. Kdlteind., 4. 23, 1897; H. Lorenz, ih., 4.
44, 1897; 10. 29, 1903; R. Pictet, ib., 7. 1, 1903: M. Schrotter, Zeit. Ver. deut. Ing , 39. 1157.
1895.
* R. J. Levy and A. Helbrouner, Brit. Pat. No. 5649, 1903 ; C. von Linde, ib., 11221, 1903 ;
G. Claude, ib., 28682, 1903; Compt. Bend., 141. 762, 823, 1905; R. C. A. Banfield, Schweiz.
Chem. Ztg., 1. 9, 1919.
6 Lord Rayleigh, Phil. Mag., (4), 49. 311, 1875.
CHAPTER XIV
OZONE AND HYDROGEN PEROXIDE
§ 1. The Discovery of Ozone and o! Hydrogen Peroxide
In natural philosophy, no observations are trivial, no truths insignificant ; that which
to us is barren is often so for this reason only, that we do not sufficiently know nor
sufficiently examine it.' — T. Bebgmann (1779).
In 1785, eleven years after the discovery of oxygen by J. Priestley, M. van Marum i
said that he noticed a peculiar smell in the vicinity of electrical machines in motion,
but he does not appear to have made any attempt to find the cause of the smell
beyond saying that " it seems clearly to be the smell of electrical matter," and
noting that the " electrical matter " has the power of acting directly on mercury.
A smell, presumably similar, has been noticed from ancient times to accompany
thunderbolts, and Homer has made several references to this odour in his Odyssey
(12. 417 ; 14. 307), and in his Iliad (8. 135 ; 14. 4:15).2 For a time the sulphurous
smell was thought to be a popular error in which the blue lurid tint of lightning
was connected with the appearance of burning sulphur, and the odour was then
imagined. Now, however, there seems no reason to doubt that the peculiar odour
which is sometimes perceptible during thunderstorms, and which has been likened
to sulphur, is identical with the odour of ozone.
In 1782, T. Cavallo ^ referred to electrified air as the aura electrica ; he noticed
its purifying action on decaying animal and vegetable matter ; and he recommended
its use as a disinfectant, H. Davy, in his Lectures on Agricultural Chemistry (London,
6, 1826), says that
In 1826 Dr. John Davy recognized the existence of this principle in the atmosphere
and published a formula for the preparation of chemical tests to be used in its detection
resembling that afterwards adopted by Schonbein.
In 1801, W. Cruickshank also mentioned that a similar odour is produced near
one of the electrodes (anode) when acidulated water is electrolyzed. In 1839, C. F.
Schonbein's attention was arrested by the similarity in the odour developed during
the electrolysis of water and during the working of an electrical machine. The
odour emitted by an electrical machine had been explained by assuming that the
sensation was due solely to a peculiar action of electricity on the olfactory nerves,
and not to the presence of a material substance. In a memoir, Recherches sur la
nature de Vodeur qui se manifeste dans certaines actions chimiques (1840), C. F. Schon-
bein * claimed that the smell must be due to the formation of a definite sub-
stance to which he gave the name ozone — from o^w, I smell. He said :
After making many fruitless experiments in order to find the relations between the
smell which is developed when ordinary electricity discharges from the points of a conductor
in air, and when water is decomposed by a voltaic current, I have finally arrived, not at a
complete solution of the problem, but at a point where le viritdble cause de Vodeur electrirjve
can be recognized more or less clearly.
According to C. F. Schonbein, ozone is a distinct form of matter with
an identity of its own ; ozone is one and the same body from whatever source
it be derived ; for he found that the same substance was produced when an electrical
877
878 INORGANIC AND THEORETICAL CHEMISTRY
machine is working ; when moist air is passed over oxidizing phosphorus ; and
when water is electrolyzed. Hence, T. Andrews (1856) could say :
I
For the first recognition of ozone and description of its properties, we are indebted to
the sagacity of SchGnbein, to whom the entire merit of the discovery unquestionably
belongs.
Numerous investigations on ozone have proved that the gas is a condensed form of
oxygen which can be symbolized by the formula O3 when ordinary oxygen is
symbolized O2.
Hydrogen peroxide was first described by L. J. Thenard ^ in 1818 in a paper
entitled Observations sur la comhinaisons nouvelles entre Foxy gene et divers acides.
The discovery was made while investigating the action of sulphuric, nitric, arsenic,
phosphoric, and acetic acids on barium peroxide. He first supposed that the liquids
which he obtained were produits suroxygknes of the different acids, but he later
showed that the liquid, eau oxygenee, contains a super-oxide of hydrogen, which
is endowed with energetic oxidizing properties, and decomposes when heated into
water and oxygen. After analyzing the liquid, he concluded :
L'eau la plus oxygenee est un. bi-oxyde d'hydrog^ne qui contient, relativement a la meme
quantite d'hydrog^ne, deux fois autant d'oxyg^ne que l'eau ordinaire, et que, toutes les
foia que l'eau oxygenee ne contient pas cette quantite d'oxygene, elle peut etre regardee
comme un melange d'eau pure et de bi-oxyde d'hydrogene.
References.
^ M. van Marum, Beschreihung einer EleJctrisiermaschine, Leipzig, 1786.
2 F. Mohr, Pogg. Ann., 91. 619, 1854.
* T. Cavallo, Complete Treatise on Electricity, London, 1782 ; W. Cruickshank, Nichclson's
Journ., 4. 254, 1801 ; Gilbert's Ann., 7. 107, 1801.
* C. F. Schonbein, Compt. Rend.., 10. 706, 1840 ; Journ. prakt. Chem., (1), 34. 492, 1845 ;
(1), 42. 383, 1847 ; (1), 51. 321, 1850 ; (1), 52. 135, 183, 1851 ; (1), 53. 248, 501, 1851 ; (1), 54.
7, 65, 1851 ; (1), 56. 343, 1852 ; (1), 61. 193, 1853 ; (1), 65. 96, 1855 ; (1), 66. 272, 1855 ; (1),
67. 496, 1856 ; (1), 75. 73, 101, 1858 ; (1), 77. 257, 1859 ; (1), 79. 65, 1859 ; (1), 80. 257, 18<)0 ;
(1), 81. 1, 257, 1860 ; (1), 83. 86, 95, 1860 ; (1), 84. 193, 406, 1861 ; (1), 86. 65, 1862 ; (1), 89.
7, 323, 1863 ; (1), 95. 385, 469, 1865 ; (1), 98. 257, 280, 1866 ; (1), 99. 11, 1866 ; (1), 100. 469,
1867 ; (1), 101. 321, 1867 ; (1), 102. 145, 1867 ; (1), 105. 198, 1869 ; Pogg. Ann., 50. 616, 1840 ;
59. 240, 1843 ; 63. 620, 1844 ; 65. 69, 161, 173, 190, 196, 1845 ; 66. 291, 1845 ; 67. 78, 225, 240,
1846 ; 68. 42, 1846 ; 71. 517, 1847 ; 72. 450, 1847 ; 75. 361, 1848 ; 78. 162, 1849 ; 100. 1, 1857 ;
Pogg. Ann. Erghd., 2. 224, 1848 ; PM. Mag., (4), 11. 137, 1856 ; (4), 21. 88, 1861 ; Ueher die
langsame Verbrennung der Korper in atmospherische Luft, Basel, 1845 ; Memoire sur V ozone.
Bale, 1849 ; Ueber die Erzeugung des Ozons auf chemischen Weg, Basel, 1844 ; Liebig's Ann., 72.
222, 1849; 89. 257, 1854; C. Engler, Historischkritische Studien fiber das Ozon, Halle, 1880;
G, W. A. Kahlbaum and E. Scheer, Christian Friedrich Schonbein, Leipzig, 1900-1.
5 J. L. Thenard, Ann. Chim. Phys., (2), 8. 306, 1818 ; (2), 9. 51, 94, 314, 441, 1818 ; (2), 10.
114, 335, 1819 ; (2), 11. 86, 208, 1819 ; (2), 50. 80, 1882 ; Les classiques de la science, 3, 1914.
§ 2. The Modes of Formation and Preparation of Ozone
When oxygen is heated to an elevated temperature, three endothermal reactions
probably occur: (i) 302=203— 68-2 Cals.; (ii) 02=20-QCals. ; (iii)03=30-Qi Cals.
Hess' principle shows that 2Qi— 302+203=3Q, and when the system is in equili-
brium 302^203 ; 02^20 ; and 03^30. In confirmation it has been shown experi-
mentally that at about 2400°, a great part of the oxygen is dissociated into atoms or
associated into molecules of ozone. The formation of ozone, O3, from oxygen, O2, is
attended by an absorption of energy nearly equivalent to 302=203— 68'2 Cals. Con-
sequently, energy is required for the formation of ozone, and this energy must be
borrowed from a foreign source, or it may be obtained whenever oxygen is developed
at a low temperature by a strongly exothermic reaction. Hence, the various methods
of preparing ozone have been arranged in two groups : physical and chemical—
OZONE AND HYDROGEN PEROXIDE 879
in the former, energy is added to the oxygen directly ; in the latter, indirectly. It
is, however, not easy to draw a strict line of demarcation.
(1) The formation of ozone by the action of heat. — The reputed formation of ozone
when hydrogen is burned, and when air or oxygen is passed over glowing platinum,
are probably mal-observations ; for example, in L. Troost and P. Hautefeuille's
oft-quoted experiment (1877), ozone was said to be formed by passing oxygen
through a porcelain tube heated above 1400°, but J. K. Clement (1904) i could find
no ozone under these conditions, not even when a Nernst's filament was heated to
over 2000° in oxygen. True, an odour resembling ozone could be detected, and
starch and potassium iodide test-paper is coloured blue. These phenomena are a
result of the formation of nitrogen oxide, not ozone ; this is shown by the fact that
the so-called tetra-base test-paper is coloured pale yellow characteristic of the
nitrogen oxides ; had ozone been present, the paper would have been coloured
violet — hydrogen peroxide has no efEect.
Ozone seems to be fairly stable at ordinary temperatures, although it gradually
decomposes on standing. Low temperatures favour the accumulation of ozone
in a system. It also appears to be fairly stable at high temperatures, while at
intermediate temperatures it decomposes very rapidly. As a matter of fact, ozone
is readily formed at high temperatures. There is a balanced reaction between
ozone and oxygen such that the higher the temperature, the greater the proportion
of ozone in equilibrium with the oxygen. W. Nernst (1913) estimates that 015
per cent, of ozone by weight can exist in equilibrium with oxygen at 1296°, 1'52
per cent, at 2048°, and 16'50 per cent, at 4500°. Hence, if ozone be formed at
a high temperature, the hot gas must be cooled more quickly than the ozone can
decompose. Ozone decomposes very much more quickly than nitric oxide, so that
if both be formed at a high temperature, the latter alone is able to survive if the
heated gases are not cooled with very great rapidity. This rapid cooling has been
accomplished by F. Eischer and E. Brahmer (1906) by rapidly chilling the heated
oxygen in various ways — e.g. by burning hydrogen or other substances beneath
the surface of liquid air or liquid oxygen ; or by plunging a glowing platinum wire
or glowing Nernst's filament under the surface of liquid air ; or by blowing air or
oxygen against a heated Nernst's filament. After hydrogen has burned beneath
the surface of liquid air for two or three minutes, the outflowing gas smells like
ozone, but it does not give the ozone reaction with the tetra-base test-paper.
The liquid in the tube contains frozen nitric oxides but no hydrogen peroxide.
After most of the clear liquid has evaporated, the residue gives all "the reactions
characteristic of ozone. This shows that nitric oxide and ozone are produced
under the conditions of the experiment. Similar results were obtained by burning
carbon monoxide, acetylene, hydrogen sulphide, sulphur, or charcoal, but were
complicated, of course, by other products of combustion. J. K. Clement's failure
to detect ozone in the products of combustion does not prove that ozone is not
formed by the heat of the flame, for ozone as well as nitrogen oxide may have been
formed, and the latter alone may have survived on cooling. Indeed, the blackening
of a silver foil in the hydrogen, or oxy-carbon monoxide blast flame, is taken by
W. Manchot (1909) to indicate the presence of ozone, since neither hydrogen peroxide
nor nitrogen oxide gives this reaction. To summarize, the effect of heating air to
a high temperature :
Product, Product.
Shorfc . . . O3 { ^^^^ ^°°^^^ .... NU
Fast cooling .... NO-fOj
Time of heating — ;— |
Long. . . 0,+N0{ B.^te<,„,i„?
Fast cooling . . . . Og
- - NO
When a platinum wire at 1700° is plunged beneath the surface of liquid air, ozone
but no nitric oxide, is formed, presumably because the temperature required for the
production of appreciable amounts of the latter is higher than for ozone. The
formation of ozone is not due to the light emitted by the glowing filaments because,
880 INOKGANIC AND THEORETICAL CHEMISTRY
if the glower be placed in an evacuated quartz tube plunged in liquid air, no appre-
ciable amount of ozone is formed in the liquid air under the conditions of the
experiment so long as the quartz is cool. Nitric oxide, NO, is of course oxidized
to nitrogen peroxide, NO2, in the presence of air or ozone. _
(2) 'The for motion of ozone hy the action of ultraviolet and radioactive radiations. — ^M
In 1894, P. Lenard 2 showed that the cathode rays which penetrated an aluminium
window in a vacuum tube, ozonized the air through which they passed ; but it was
not made clear whether the ozone was produced by the cathode rays, or indirectly
by the ultraviolet light produced by cathode rays in air. P. Lenard detected no
other chemical effects by these rays. Six years later, P. Lenard showed that oxygen
is ozonized by ultraviolet light of wave-length between 0*00014 mm. and 0*00019 mm.,
i.e. by rays of great refrangibility to which air is almost opaque. Air is more opaque
than rock salt, fluorspar, or quartz to the most chemically active ultraviolet rays.
Hence, air-spaces in the path of the rays should be avoided in designing apparatus
for ozonizing oxygen by ultraviolet rays. Ozone is produced by the action of
radiations of short wave-lengths on oxygen. Cathode rays and ultraviolet light
rays, acting on air or oxygen, produce ozone, and this the more the lower the
temperature. If liquid or solid oxygen be exposed to ultraviolet light, ozone can
be detected in the oxygen obtained by subsequent evaporation. The mercury
vapour or uriol lamp is commonly employed for producing ultraviolet radiations ;
the lamp is made of quartz, not glass, because quartz is transparent and glass is
opaque to these radiations. E. Goldstein (1903) passed a discharge through a
quartz Geissler's tube, and found that the air in the vicinity was strongly ozonized.
If the pressure of the gas in the Geissler's tube be too large, no ozonization occurred.
It was assumed that rays of ultraviolet light of small wave-length penetrated the
quartz, and ozonized atmospheric oxygen. In H. N. Potter's patented process
for ozonizing air, a current of air is conducted spirally about a quartz mercury
vapour lamp, within a sheathing of ordinary glass. The mercury lamp gives off
a copious stream of ultraviolet light.
There is no definite equilibrium ratio between the rates of formation and decom-
position of ozone when air or oxygen is exposed to ultraviolet light ; the amount
formed increases with fall of temperature, and decreases with fall of pressure. The
decomposition of ozone by ultraviolet light is very slow, but is strongly accelerated
by small quantities of hydrogen which reacts thus: H2+03=H204-02. No signs
of the reaction 203+H2=H202+202 are observed when hydrogen is in excess ;
the main reaction is H2+03=H20+02, and when ozone is in excess, the main
reaction is 203=302-
According to E. Regener (1906) ,3 ozone has a maximum absorptive power for
ultraviolet light of wave-length 258/x/x, and light of wave-length 200-300)Lt/Lt converts
ozone into ordinary oxygen ; on the other hand, V. Schumann showed that
oxygen absorbs waves below 193/x/x, and the conversion of oxygen into ozone by
ultraviolet light is largely the result of exposure to light of wave-length less than
200/i,/x. There is therefore an equilibrium value for the reaction 302=203, which
decreases with rise of temperature, E. Regener found in one set of experiments
3*4 per cent, of ozone at 20°, 3*15 at 40°, 2*7 at 54° ; but the actual proportion of
the two gases is also determined by the intensity of the ultraviolet light — vide Fig. 6.
The solar spectrum ceases abruptly at 293/x/x, from which it is inferred that light
of shorter wave-length is absorbed by the atmosphere ; and further, since oxygen
itself does not appreciably absorb light of greater wave-length than 200jLt/>t, it follows
that the ozone, formed by the absorption of light of wave-length below 200)Lt/x, is
partially destroyed by light of wave-length approximately 293/a/x. This is con-
firmed by the spectroscopic observation that ozone gas has two well-defined absorp-
tion bands in the red part of the spectrum, and that the residual transmitted light
is markedly blue. These observations are said to favour the hypothesis that the
ozone in the atmosphere may have been formed by the action of the ultraviolet
rays from the sun, on the oxygen in the upper regions of the atmosphere, and that
OZONE AND HYDROGEN PEROXIDE
881
ozone is formed in the upper atmosphere in sujficiently large quantities to account
for the normal blue colour of the sky ; this has been rendered further probable by
actual determinations of the amounts of ozone in the upper atmosphere. At a
mean altitude of 15*2 kilometres, for instance, J. N. Pring (1914) found a mean
of 2'1 xlO~^ volumes of ozone per volume of air.
* The action of the radiations from radium or other radioactive substances * on
air or oxygen furnishes some ozone. If radium be enclosed in a tube with oxygen,
ozone is formed, but not if the radioactive substance is in a separate glass vessel
which in turn is placed in a tube of oxygen. The radiations from radium, polonium,
etc., can produce ozone. S. C. Lind (1911)^ showed that the amount of ozone
formed by the action of a-rays on oxygen is such that one molecule of ozone is
formed for two pairs of gaseous ions. F. Kriiger and M. Moller found that in
the passage of electrons of high velocity through gaseous oxygen, one pair of ions
is concerned in the formation of each molecule of ozone.
(3) The formation of ozone hy electrolysis. — In 1840, C. F. Schonbein ^ observed
the presence of ozone in the gases evolved during the electrolysis of acidulated
water and salt solutions. Ozone can generally be detected in the oxygen gas obtained
during the electrolysis of acidulated water with non-oxidizable electrodes— e.^. gold,
platinum, etc. The yield of ozone in the 'electrolysis of acidulated water is very
small, but is increased by reducing the temperature and by increasing the current
density — i.e. by increasing the intensity of the current or reducing the surface
of the anode. By the electrolysis of sulphuric acid
of a specific gravity between 1*075 and 1*1, with an
anode made by imbedding platinum foil in glass and
grinding away the edge so that a line of platinum
0"1 mm. broad and 11*5 mm. long, is exposed, F.
Fischer and K. Bendixsohn (1909) '^ prepared oxy-
gen containing 17 to 28 per cent, of ozone. This
form of anode prevents long contact between the
platinum surface and the gas. The curve, Fig. 1,
shows the relation between the yield of ozone by
weight and the concentration of the sulphuric acid.
With a current density of 65 amps, per sq. cm., the
maximum yield is obtained with an acid of specific
gravity I'l. An acid of specific gravity 1'22 has a maximum electrical con-
ductivity, so that the best conducting acid does not give the maximum yield of
ozone. Solutions of phosphoric, chromic, nitric, perchloric, or hydrofluoric acid ;
ammonium or potassium sulphate, potassium hydrogen carbonate ; or of sodium
or potassium hydroxide do not give such good yields as sulphuric acid. Platinum-
iridium electrodes give the best results — lead or lead peroxide electrodes are rapidly
destroyed. The cell should be immersed in water cooled to 0°, and the anode
should be internally cooled by a freezing mixture to about —14°. The method of
making ozone by the electrolysis of sulphuric acid is dearer than the electrical
discharge process ; but it is an advantage that the gas is free from nitrogen, and
that the hydrogen obtained as a by-product may be utilized.
(4) Theformaiion of ozone by the action of an electrical discharge. — The electrical
discharge through air as dielectric produces a variety of effects : luminous, thermal,
chemical, mechanical, and magnetic. In the production of ozone by the electric
discharge, chemical action is alone wanted, and accordingly the conditions should
be such as to keep the waste of energy expended in producing other effects as low
as possible.
The relations between the current C and voltage F in a gaseous discharge are somewhat
complex.* If the current from a positively charged point, passing through the air to the
earth, be gradually increased, the voltage rises rapidly and very small currents pass as a
non-luminous or invisible discharge which produces no chemical effects ; finally, Fig. 2.
there is a slight discontinuity in the voltage-current curve, and the discharge becomes
VOL. I. 3 L
20
18
16
14
12
10
8
6
100 105 ilO irS 1-20'
Fig. 1.— Yield of Ozone with
Sulphuric Acid of Different
Concentrations.
/
\
"
/
\
j
\
1
\
\
'^'^-
^
\
§
\
<^
Specific Gravity of^Acid^
INORGANIC AND THEORETICAL CHEMISTRY
50000
40000
30000
20000
10000
Vofts
luminoua— glow or silent discharge. This is a high tension discharge of little energy. Very
little electricity leaks during the so-called invisible and glow discharges. For an
air-gap a few centimetres in length, there are very high voltages and small currents
of a few amperes ; as the voltage increases, the conductivity of the air increases,
and the voltage rapidly falls, the current increases, and the so-called brush discharge,
more luminous than the glow discharge, is developed. The appearance and nature of the
brush discharge varies considerably with the conditions and the shape of the electrodes ;
it develops very little light and heat, and it acts almost exclusively on the oxygen- — very
little nitrogen is oxidized. The conditions favourable for the formation of the brush
discharge are important since the production of ozone by the electrical discharge is largely
the result of its work. The brush discharge is sometimes called silent, or dark discharge.
These terms are misnomers ; the brush discharge is neither silent nor dark, for it is attended
by a peculiar sound, and it is coloured bluish-white. The German equivalent is Bueschelent-
ladung ; the French equivalents are Veffluve electrique and Vaigrette electrique according as
the discharge shows a narrowed stem or is expanded fan-wise. The transition from one
form of high tension discharge to another may be accompanied by sparking, where the
discharge is characterized by a loud snapping noise and a yellow colour. The spark and
brush (Sscharges may be mixed. Sparking is very detrimental to the formation of ozone.
At —194°, E. Briner and E. Durand (1907) * found that 99 per cent, of oxygen is converted
into ozone with the silent discharge, and only 1 per cent, with the spark discharge ; and
even this was considered to be produced by the silent discharge which occurred simul-
taneously with the sparking.
There is another discontinuity in the voltage -amperage curve of electrical discharge
as the brush discharge changes into the electrical flame- — the precursor of the high-tension
arc. When the arc has been established, the path of the discharge is strongly luminous,
and the ratio O/F is higher than before ; but instead of the voltage increasing with current,
it now decreases owing to the fact that the resistance
of the air-gap decreases faster than the increase of
current. The high-tension arc in air at ordinary
pressure corresponds with the glow-discharge in a
gas at a low pressure. The electric arc is active in
producing nitric oxide in air. The temperature is
different in different parts of the arc and depends on
the current and voltage ; it approximates 2200° or
2500° in the positive colimin. As the current in-
creases still further, the temperature rises, and
another discontinuity occurs as the discharge passes
into the low-tension or lighting arc. The low-tension
arc is used in steel and carbide furnaces ; the tem-
perature is very high ; and the electrode material
is largely vaporized. The temperature with carbon
electrodes is about 3500°. With a negatively
charged point, the results are similar but rather
less complex. If the electric discharge takes place between parallel conductivity plates,
with one or both covered by a solid dielectric, the phenomenon is rather more complex.
The brush discharge then changes its character, but it still retains its valuable property
of converting oxygen into ozone. As a matter of fact, a series of electric sparks in oxygen
will form ozone, and in air a mixture of nitrogen oxide and ozone ^^ — all depends on the rate
of cooling as indicated previously.
The brush discharge is most favourable for the production of ozone.^i In the
arc and spark discharges much heat and light are developed, and energetic chemical
action occurs ; ozone is formed, but the nitrogen of the air also reacts with the oxygen
under this stimulus. The fact that nitrogen oxides are often produced when ozone
is made from atmospheric oxygen rather confused the minds of the early investigators
as to the real nature of ozone, for the two products were not always clearly discrimi-
nated by the tests employed — usually, the bluing of starch and potassium iodide
test-paper.
E. Warburg 12 found that under different conditions, the electrical discharge
produces from 93 to 1000 times as much ozone as would have been obtained by
electrolysis. One equivalent of hydrogen reduces 24 grms. of ozone, and there-
fore the equivalent of ozone is taken to be 24. E. Warburg and G. Leithauser
obtained between 0'003 and 0*1 grm. of ozone per coulomb, so that between 240
and 8(X)0 coulombs are required to produce 24 grms. of ozone ; these numbers are
not at all equivalent to the electrochemical equivalent, 96,540 coulombs. The
energy required is greater than the heat of the reaction. The highest yield of
0*2 Amperes
Fig. 2. — Voltage -Current Curve of
Electrical Discharges.
OZONE AND HYDROGEN PEROXIDE
883
ozone, 70 grms. per kilowatt hour, is equivalent to 589 Cals. per gram-molecule —
nearly 20 times the energy equivalent to the heat of the reaction. Hence it was
inferred that the formation of ozone by the silent discharge is not a direct electro-
lytic action, but is rather a secondary efiect of the ultraviolet and cathode rays
generated in some profusion by this discharge. Quartz plates are virtually trans-
parent to ultraviolet light, so that if the discharge passes inside a quartz vessel
surrounded by oxygen, ozone is formed ; while if under similar conditions, the
discharge passes inside a glass vessel, which is almost opaque to ultraviolet rays,
there is little or no ozonization of the oxygen. E. Warburg assumes that the
electrons, generated by the ultraviolet and cathode rays which have a velocity as
high as that required for the production of luminosity, are effective in forming
ozone either directly by impact with oxygen molecules, or indirectly by the inter-
mediate production of short aether waves.
The amount of ozone obtained per coulomb of electricity follows no known law,
and it is therefore necessary to find the yield of ozone under different conditions
empirically. The silent discharge has a deozonizing as well as an ozonizing effect
on oxygen. The speed of the ozonization is proportional to the amount of oxygen
present, and the speed of the deozonization is proportional to the amount of ozone
present. In other words, the reaction follows the law of opposing reactions. If
the discharge be passed an infinite time, a certain definite limiting concentration
of ozone will be reached when the rate of decomposition is equal to the rate of forma-
tion of the ozone : 302=f^203.
E. Warburg 13 obtained the results shown in Table I. with an apparatus containing
oxygen, and fitted with a small pointed platinum wire (0*05 mm. diameter) connected
to the negative pole of an electrostatic machine, and discharging on to another
platinum wire (0*05 mm. diameter) connected to earth. The results show that the
maximum concentration of ozone decreases as the temperature rises, owing to an
Table I.- — ^Maximum Concentration of Ozone.
Temperature.
Ozone-*per cent,
by volume.
A constant pro-
portional to the
rate of formation.
A constant pro-
portionai to the
rate of decomposition.
-71
0
17
60
93
5-74
4-19
3-53
2-22
1-23
0-0232
0-0219
0-0225
0-0214
0-0277
0-380
0-503
0-616
0-939
1-420
increase in the speed of decomposition of ozone, and not to a marked reduction
in the speed of formation of ozone. The spontaneous decomposition of ozone was
negligibly small.
E. Goldstein's i* experiment shows how a low temperature favours the formation
of ozone. He introduced oxygen into an evacuated Geissler's tube until the pressure
registered a few centimetres. The tube was partially immersed in liquid air, and
electrical discharge passed through the tube. In about half a minute, the pressure
sank to about 01 mm. Oxygen was again introduced until the pressure reached
a few centimetres, and the process repeated again and again. In this way,
E. Goldstein claimed 100 per cent, conversion of oxygen into dark blue liquid ozone.
In general, the greater the pressure of the gas, the greater the yield. In
E. Warburg's experiments, the yield at a pressure ^—between 460 and 780 mm. —
was (0*32360+0'00089^) times the yield at a pressure of 760 mm. — temperature
between 17° and 23°. According to E. Warburg, the decrease in the yield with a
rise of temperature is largely due to the decrease in the density of the oxygen.
E. Warburg's experiments further showed that the maximum concentration
with a positive discharge is one-third the value obtained with a negative discharge
884 INORGANIC AND THEORETICAL CHEMISTRY
owing to the greater ozonizing effect of the latter, while the speed of deozonization
and the temperature effect is nearly the same with both a positive and a negative
discharge. E. Bichat and A. Guntz (1888) ^^ also testified to the greater efficacy of
the negative discharge in the formation of ozone, but A. Vosmaer (1916) maintains
that this is an error probably due to a reversal of the poles during the working of
the electrostatic machine, so that what was thought to be the negative pole was
really positive.
The yield of ozone decreases with increasing current up to a certain point ; thus
with oxygen of 98*5 per cent, purity :
Volts .... 7230 8800 12500
Amperes . . . 0*00000883 0*0000175 0 0000523
Ozone (grms. per coulomb) 0*0950 0*0908 0*0485
If the current be increased still more, the yield reaches a minimum, and then
increases with the current. Thus, with oxygen of 96 per cent, purity :
Volts . ... 6080 7000 9610 12510
Amperes . . . 0*00000146 0*0000219 0*0000524 0*0001307
Ozone (grms. per coulomb) 0*0423 0*0375 0*0307 0 0422
The effect of variations of temperature and pressure is complicated, because not only
is the substance itself altered, but the reagent which brings about the reaction is
also modified. According to A. Chassy ^^ the amount of ozone formed is proportional
to the voltage^ but there is a doubt about this since it is very difficult to- vary the
voltage in electrical discharges through gases without at the same time varying the
energy. The formation of ozone, after all, is a question of the transformation of
electrical into chemical energy. The quantity of available electrical energy may be
regarded as a product of the voltage and amperage, and since the voltage for a given
apparatus does not change very much,, the yield of ozone must depend on the
amperage, and bear no special relation to the voltage.i^ The voltage can be regarded
only as a force which can neither be under nor over a certain limiting value consistent
with the apparatus. If the voltage be too high there is a danger of sparking through
the air, or of cracking or piercing a solid dielectric. Hence, the voltage is run as
high as is consistent with safety in order to keep the product — voltage X amperage —
high. The yield is then large, not because the voltage is high but because a large
amount of electrical energy is available. Most experiments on high-tension currents
have been made with alternating currents. A higher frequency than 100 is favour-
able to the brush discharge. M. W. Franklin i^ says that the yield of ozone is pro-
portional to the periodicity of the current. According to E. Warburg, the yield
with an alternating current is not so good as with a direct current, but this has not
been definitely accepted.
According to E. Warburg and G. Leithauser, the presence of waler vapour in the
oxygen to be ozonized reduced the yield nearly proportional to the pressure of the
water vapour.i^ With a negative discharge, a reduction of 7 mm. in the pressure
of the water vapour reduced the yield to 94 per cent, of its value for dry oxygen.
The effect of moisture is greater with air, and greater still with a positive discharge.
The retardation with moisture is greater with oxygen than with air. E. P. Perman
and R. H. Greaves also find water vapour accelerates the decomposition of ozone.
Even when extreme precautions are taken to dry the gas thoroughly, W. A.
Shenstone and T. A. Cundall still found that oxygen is ozonized by the silent
discharge, and they also obtained a greater yield with moist than with the thoroughly
dried gas. H. B. Baker then stated that " ozone is formed as rapidly in oxygen
dried with phosphorus pentoxide as it is in the same tube when the oxygen is dried
by sulphuric acid." W. A. Shenstone (1897) then showed that a high percentage
of ozone is formed by the action of the silent discharge on oxygen saturated with
water vapour and the ozone produced is remarkably stable ; while on partially
drying the gas, the percentage of ozone produced is considerably reduced, and the
OZONE AND HYDROGEN PEROXIDE
885
gas is singularly unstable." W. A. Shenstone's result has not been confirmed, and
it is probably a mal-observation due to the presence of nitrogen peroxide in the
dry gas, derived from the action of the discharge on slight trace of nitrogen in the
oxygen gas. E. Warburg, and D. L. Chapman and H. E. Jones found that the
velocity of decomposition of ozone at 100° is virtually the same whether water
vapour be present or absent.
According to E. Warburg (1904), no nitrogen oxide is formed when the oxygen
is mixed with 7 per cent, of nitrogen, but this is doubtful since with air, nitrogen
oxides are formed. A spark discharge produces oxides of nitrogen alone ; these
Ozonized Oxygen
Fig. 3. — L. von Babo's Tube Ozoniaeur.
Oxygen
oxides prevent the formation of ozone. A little nitrogen or hydrogen favours the
production of ozone,^^ but if sparking occurs, the hydrogen unites with the oxygen —
explosively^f sufficient be present. According to P. Hautefeuille and J. Chappius,
a little silicon tetrafluoride or hydrogen fluoride does not affect the yield of ozone ;
but a trace of chlorine or nitrogen oxide hinders the ozonization of oxygen.^i
The brush discharge may be produced from a static electric machine, from a
battery of cells or a dynamo and an induction coil ; or from an alternating current
from a dynamo of high periodicity, and transformed up to several thousand volts.
The latter is the means employed in the commercial production of ozone, while the
Ozonized
Oxygen
Fig. 4." — Preparation of Ozone with Siemens' Tube.
induction coil furnishes a useful ozonizing discharge in common use for demonstra-
tions and other purposes. L. von Babo's ozonizer modified by A. Houzeau,22
is one of the oldest, and simplest. It is illustrated in Fig. 3. The discharge takes
place between two platinum wires, one of which, B, Fig. 3, is fitted axially in the
narrow tube — this wire is about one or two mm. thick and 40 cm. long ; tlie other
wire is rather thin, and is wound around the outside of the narrow tube. The
ends of these wires, ^'B', are put in communication with the two poles of an induction
coil (G, Fig. 4). A slow current of oxygen is passed into tube at one end ; ozonized
oxygen escapes at the other end. The ozonizer illustrated at A B, Fig. 4, is
886
INORGANIC AND THEORETICAL CHEMISTRY
virtually the one devised by W. von Siemens (1857) .23 Siemens' ozonizer (or
ozonator, or electrizer), as it is called, consists of two concentric tubes. The inner
tube is coated on its inner surface with tinfoil in metallic contact with the terminal A ;
and the outer tube is coated on its surface with tinfoil in metallic contact with the
terminal B. This forms the so-called Siemens' induction tube. The two terminals
are connected with an induction coil. A slow stream of oxygen is led from the
gasholder G through the calcium chloride drying tube D, and then through the
annular space between the concentric tubes, and is there exposed to the action of
the silent discharge of electricity, operated by the induction coil G, and battery E.
Ozone is decomposed by cork and indiarubber. In consequence, these materials
should not be used for any part of the ozonizer in contact with the gas. The gas
issuing from the ozone tube, or ozonizer, is charged with 3 to 8 per cent, of ozone.
If the oxygen contains traces of chloride, the gas should be washed in dilute alkali
before drying. In some forms of ozonizer there are three concentric tubes, and
cold water circulates in the inner tube while the discharge is passing so as to prevent
a rise of temperature, and thus increase the yield of ozone.
The ozonizer devised by B. C. Brodie in 1872— Brodie's ozonizer 24_is usually
called Berthelofs ozomzer, though the latter designed
it some years after B. C. Brodie. Here the tin-
foil coatings are replaced with sulphuric acid as
illustrated in Fig. 5. It consists of two concentric
thin glass tubes with an annular space from 1 to
4 mm. There are inlet and egress tubes as shown.
The inner tube is about 30 mm. diameter and
30-35 cm. long. This apparatus is suspended in
a cylinder of water acidulated with sulphuric acid
(1 : 10), and the inner tube likewise contains the
same liquid. Platinum wires dip in the liquids.
The wires A and B are connected with the induc-
tion coil and a slow current of oxygen is sent
through the apparatus — ozonized oxygen escapes.
The quantity of ozone per coulomb increases with
the potential.25 According to F. Russ, the yield
of ozone with an ozonizer of vitreous quartz is but
half that obtained with a glass ozonizer. The
yield of ozone per kilowatt hour is greater when
one electrode is not covered with an insulator
Fig. 5. — B. C. Brodie's Ozone Tube, (glass) . The parts of the tube which do not dip
in the dilute acid are covered with shellac varnish
in order to prevent external sparking. The gas circulates in the annular space
between the inner and outer tubes. If the gas travels through the apparatus
too quickly some escapes the action of the discharge ; and if too slowly, the
ozone may be decomposed. About one bubble of gas per second gives good
results. The cooler the vessel, the better the yield ; at —23°, oxygen containing
21 per cent, of ozone can be obtained.
It will be observed that in W. von Siemens' or in M. Berthelot's ozonizer there
are three dielectrics in the path of the discharge — two layers of glass, and one of
gas ; in L. von Babo's ozonizer there is one layer of glass, and one of gas. Neglecting
the gas to be ozonized which must necessarily be present in all ozonizers, there
are two types of ozonizers in use, for the ozonization of air or oxygen on a large scale :
those which have dielectrics in the path of the discharge — e.g. E. W. von Siemens and
J. G. Halske's, Abraham and Marmier's, and Linder's ozonizers 26 — and those which
have no dielectric — e.g. A. Schneller's, A. Vosmaer's, and H. Tindal's ozonizers. 27 In
the so-called ozonair system, a series of mica plates, covered on both sides with a metal
alloy, are mounted side by side in a box. The plates serve as electrodes. Air
passes into the box between the plates. It is claimed that the gauze promotes the
OZONE AND HYDROGEN PEROXIDE 887
formation of a sparkless brush discharge, and that the open arrangement of the
plates suffices to keep them cool without the aid of water cooling.
Ozone is said to be produced by the violent mechanical disturbance of air 28 —
say, when grinding wheels are being tested for bursting speed. This may be an
effect of heat or of electrification or both. Ozone is also said to be formed during
the evaporation of water. This statement may be regarded as not proven ; nor
is the evidence satisfactory as to the formation of ozone when water or a salt
solution — e.g. sea water — ^is splashed about,^^ although these statements have been
cited to explain the greater ozone content of air in the vicinity of the sea, water-
works, irrigation plants, and waterfalls ; and for the bleaching of linen, etc.,
spread on lawns. The electrification of air by the splashing of liquids is,
however, a well-known phenomenon investigated by P. Lenardj^o Lord Kel-
vin, etc.
(5) The formation of ozone in chemical reactions. — Ozone is usually present in
the oxygen obtained by low temperature exothermal reactions ; but not if the
temperature of the reaction be high, because the ozone, if it be formed, is at once
transformed into oxygen. Potassium chlorate gives ozone-free oxygen, but if
the chlorate be mixed with catalytic agents which lower the temperature of the
reaction, ozone may be found in the resulting oxygen (q.v.). Crystallized iodic
or periodic acid decomposes between 130°-135° giving off strongly ozonized oxygen ;
and, according to C. F. Rammelsberg,3i aqueous solutions of periodic acid or
periodates smell of ozone ; but A. R. Leeds has shown that what C. F. Rammelsberg
thought to be ozone is really chlorine or nitrous acid, present as an impurity in the
periodate. Ozonized oxygen is also formed when many oxidizing agents are heated
alone or mixed with acids. For instance, ozonized oxygen is obtained from silver
peroxide (C. F. Schonbein, 1855), lead peroxide (C. F. Schonbein, 1855), and mercuric
oxide (C. T. Kingzett, 1872). As previously indicated, A. R. Leeds contended that
C. F. Schonbein mistook chlorine for ozone in assuming that ozonized oxygen
was formed by heating these oxides — both gases affect a solution of starch and
potassium iodide in the same way; but 0. Brunck 32 showed that manganese
dioxide, chromium, nickel, cobalt, and gold sesquioxides, silver mono- and di-
oxide, mercuric oxide, and chromium and uranium trioxides give ozonized oxygen
if heated in an atmosphere of oxygen, but not if heated in an atmosphere free
from oxygen. Ozonized oxygen is formed by the decomposition of hydrogen
peroxide (A. Riche, 1860), not only with sulphuric acid, but also with dilute sul-
phurous acid, with finely divided metals, and with all substances 33 which stimulate
the decomposition of this compound. F. Raschig also detected ozone in gas
obtained by dissolving nitrogen peroxide in sulphuric acid in the lead-chamber
process. 34 Ozonized oxygen is also obtained by the action of sulphuric acid on
barium peroxide (A. Houzeau, 1881), sodium peroxide (C. Arnold and C. Mentzel,
1902), persulphuric acid and the persulphates (A. von Baeyer and V. Villiger, 1901,
and P. Malaquin, 1911), percarbonic acid and the percarbonates (C. Arnold and
C. Mentzel, 1902), perborates (S. Tanatar, 1898), per-monophosphates (J. Schmidlin
and P. Massini, 1910), permanganates (G. Bertazzi, 1855), bichromates (C. Weltzien,
1867), peruranic acid (P. G. Melikoff and L. Pissarjewsky, 1897), fluovanadic com-
pounds (P. G. Melikoff and P. Kasanezky, 1901), barium ferrate (A. Baschieri, 1906),
per-pyrosulphates (W. Traube, 1909), acetone peroxide (A. von Baeyer and
V. Villiger, 1900), performic acid (J. d'Ans and A. Kneip, 1915), etc.35
H. Moissan 36 (1891) found that when drops of water are allowed to fall into a
vessel containing fluorine gas, the water is decomposed, and hydrogen fluoride and
deep blue vapours of ozonized oxygen are produced. At 0°, oxygen containing up
to 21 per cent, of ozone by weight was obtained. The lower the temperature of the
reaction, the greater the yield of ozone. 0. Ruff and F. W. Tschirch 37 obtained
ozonized oxygen by the action of osmium octaflupride on soda lye. L. Graf en-
berg 38 obtained ozonized oxygen by the electrolysis of hydrofluoric acid — ^with
the 40 per cent, acid, a maximum yield of 5*2 per cent, of ozone was obtained.
888 INORGANIC AND THEORETICAL CHEMISTRY
E. B. R. Prideaux also electrolyzed saturated solutions of alkali fluorides, and
obtained a yield never exceeding 1 per cent, of ozone.
(6) The formation of ozone during slow oxidations. — Ozone is formed during the
slow oxidation of many substances ; C. F. Schonbein 39 detected it in the atmosphere
of a flask containing a couple of sticks of clean phosphorus, and J. C. G. de Marignac
prepared ozonized air by aspirating atmospheric air through a flask or tube contain-
ing a few pieces of clean phosphorus partly submerged in water. According to
C. F. Schonbein, one part of phosphorus converts yrj^oo^h part of oxygen into ozone.
The resulting gas should be washed to free it from phosphoric oxide. The action
is very slow at the freezing point of water, but between 15° and 20°, the action is
fairly quick, and at 38° but little ozone is formed ; according to A. R. Leeds, the
optimum temperature is 24°, but even then, the yield is very small — say about
2 mgrm. of ozone per litre of air. According to R. Engel, reducing the pressure
favours the formation of ozone ; thus, the action does not occur with appreciable
velocity below 6°, but under reduced pressure, the action may occur at 9°. Since
ozone is decomposed in contact with phosphorus, a rapid current of air is desirable.
According to A. R. Leeds and R. Bottger, the addition of sulphuric acid and
potassium permanganate or potassium dichromate to the water is said to increase
the yield of ozone ; but C. Arnold and C. Mentzel found no advantage in the use of
chromic acid.
Instead of air, a mixture of oxygen and carbon dioxide can be used. J. C. G. de
Marignac found that a mixture of oxygen and hydrogen sometimes exploded. Minute
traces of ammonia, sulphur dioxide, nitrogen peroxide, alcohol, ether, ethylene,
and ethereal oils retard the action. The liquid in the flask 40 contains phosphorous
and phosphoric acids, hydrogen peroxide, and ammonium nitrite. According to
C. T. Kingzett, the ratio of ozone to hydrogen peroxide formed in the reaction
is 1 : 2'4 ; and according to A. R. Leeds, 1:1. The evidence is not quite satis-
factory since H. McLeod failed to find hydrogen peroxide, but his tests were not
very delicate.
P. Villard ^i has suggested that the oxidation of oxygen to ozone in this reaction
is produced by radiations of short-wave length generated during the phosphorescence
of the phosphorus. The phosphorescence of phosphorus is probably an ozonizing
action ; similarly, according to L. Bloch, when sulphur is heated to about 200°,
the luminescence is accompanied by the formation of ozone. According to J. H.
van't HofE *2 and W. P. Jorissen, only half an atom of oxygen is available for the
oxidation of oxygen to ozone per atom of phosphorus oxidized. Ozone is not
formed when a substance is present capable of giving rise to the formation of hydro-
gen— e.g. hydrogen peroxide is formed by the oxidation of water when zinc, lead,
etc., is substituted for the phosphorus. During the oxidation of phosphorus the
ambient air becomes electrically conducting.
Many other substances also furnish ozone or activate oxygen during their oxida-
tion. Ozone is said to be formed during the combustion of ether as well as during
the combustion of hydrogen compounds generally. At any rate, the potassium
iodide test indicates the formation of ozone (or hydrogen peroxide) when a spiral
of hot platinum is placed above the surface of a little ether in the bottom of a beaker.
The ether burns on the surface of the platinum, and the platinum remains incan-
descent as long as any ether remains in the beaker. The alleged formation of ozone
during the evaporation of alcohol, ether, ethereal oils, hydrocarbons, etc.,^^ i^^ay be
an effect of an analogous phenomenon — autoxidation. By exposing p-diacetyl-
diamino-stilbene-o-disulphonic acid, or many of its salts, in glass vessels to sunlight,
the colour changes from bright yellow to reddish-brown, and the reverse change
takes place in darkness. The glass cuts ofE the rays below 350/x wave-length.
The presence of oxygen is necessary, and ozone is simultaneously formed. This
may be a kind of autoxidation phenomenon.
According to C. F. Schonbein,^* ozone is formed when turpentine, benzene,
petroleum, aldehyde, coal tar, many hydrocarbons, mineral and essential oils —
OZONE AND HYDROGEN PEROXIDE 889
eucalyptus, lavender, cinnamon, etc. — are oxidized. Acids and resinous matter
are formed at the same time. There are ozonators on the market which contain
an essential oil — say cinnamon oil — which slowly evaporates ; they do not give
ozone at all. C. Engler ^5 has shown that it is not ozone which is produced by these
organic compounds, but rather an unstable peroxide, or ozonide, which can act as
a powerful oxidizing agent. For instance, when turpentine is agitated with a large
volume of air, oxygen is absorbed, and part of the substance is oxidized, and it is
this oxide which has led to the assumption that ozone is formed and dissolved by
the hydrocarbon — for it can colour starch and potassium iodide test-paper blue,
decolorize indigo blue, colour tincture of guaiacum blue, etc. When a 5 to 7 per cent,
benzene solution of dimethylfulvene, CgH^o, is agitated with air 46 for a few days, a
precipitate of a peroxide, C8H10O4, is formed which detonates at 120°, and which
gives the usual reactions for peroxides only when a little ether, alcohol, ethyl acetate,
or chloroform is present.
References,
1 L. Troost and P. Hautefcuille, CompL Rend., 84. 946, 1877 ; J. K. Clement, Ueher die
Bildung des Ozons bei hohen Temperaturen, Gottingen, 1904 ; 0. Loew, Ber., 22. 3325, 1889 ;
Dingier' s Journ., 206. 421, 1872; Zeit. Chem., 13. 65, 269, 1869; R. Bottger, Jahrb. phys. Ver.
Frankfurt a. M., 14, 1875 ; J. D. Boeke, Ohem. News, 22. 57, 1870 ; R. H. Ridout, tb,, 41. 98,
1880 ; A. R. Leeds, ib., 49. 237. 1884 ; J. T. Cundall, ib., 61. 19, 1890 ; J. Schnauss^ Arch.
Pharm., 192. 193, 1870 ; K. Than, Journ. prakt. Chem., (2), 1. 415, 1870 ; H. Struve, Zeit. anal
Chem., 10. 292, 1871 ; W. Radulowitsch, Ber., 7. 1454, 1874; P. Rumine, ib., 5. 123, 1872;
W. Nernst, Zeit. Elektrochem., 9. 891, 1903 ; E. Warburg, Ann. Phi/sik, (4), 9. 1286, 1902 ;
(4), 13. 1080, 1904 ; S. Jahn, Zeit. phys. Chem., 48. 260, 1906 ; F. Fischer and E. Brahmer, Ber.,
39. 940, 1906 ; H. Marx, ib., 39. 2557, 3631, 1906 ; 40. 443, 1111, 1907.
2 P. Lenard, Ann. Physik, (3), 51. 225, 1894 ; (4), 1. 486, 1900 ; W. Hallwachs, ib., (4), 30.
602, 1909 ; E. Warburg, ib., (4), 13. 464, 1904 ; E. Regener, ib., (4), 20. 1033, 1906 ; E. Warburg
and E. Regener, Sitzber. Akad. Berlin, 1228, 1904 ; E. Goldstein, Ber., 36. 3042, 1903 ; F. Fischer
and F. Brahmer, ib., 39. 490, 1906 ; F. Fischer, ib., 42. 2228, 1909 ; Phys. Zeit., 10. 453, 1909 ;
H. Thiele, Zeit. angew. Chem., 22. 2472, 1909 ; A. Pfliiger. Phys. Zeit, 5. 414, 1904 ; A. Laden-
burg, ib., 5. 525, 1904 ; H. N. Potter, 11. S. A. Pat. No. 54, 965, 1907.
3 E. Regener, Ann. Physik, (4), 20. 1033, 1906 ; E. Warburg and E. Regener, Sitzber. Akad.
Berlin, 1228, 1904 ; E. Warburg, ib., 216, 1912 ; D. L. Chapman, S. Chadwick, and J. E. Ram.s
bottom, Jourv. Chem. Soc. 91. 942, 1907 ; F. M. G. Johnson and D. Mcintosh, Journ. Amer.
Chem. Soc, 31. 1146, 1909 ; H. Thiele, Zeit. angew. Chem., 22. 2472, 1909 ; E. van Aubel, Compt.
Rend., 149. 983, 1909 ; 150. 96, 1910 ; Phys. Zeit., 11. 63, 1910 ; F. Kriiger and M. Moller, ib.,
13. 1010, 1912 ; D. Berthelot and H. Gaudechon, Compt. Rend., 150. 96, 1169, 1910 ; J. Cour-
mont, Th. Nogier, and A. Rochaix, ib., 149. 160, 1909 ; W. G. Chlopin, Zeit. anorg. Chem., 71.
198, 1911 ; H. Kuhne, Zeit. Elektrochem., 12. 409, 1906 ; F. Weigert, Zeit. phys. Chem., 80. 78,
1912; C. Baskerville, Chem. News, 95. 255, 1907; J. N. Pring, Science Prog., 9. US, 1915;
F. Weigert and H. Bohm, Zeit. phys. Chem., 90. 189, 1915 ; V. Schumann, Smithsonian Contrib.
Knouledge, 29, 1903.
* M. and P. Curie, Compt. Rend., 149. 366, 1909 ; 150. 52, 1910 ; W. Ramsay, Journ. Chem.
Soc, 91. 931, 1907 ; R. Nasini and M. G. Levy, Atti Accad. Lincei, (3), 17. 616, 1908 ; S. C. Lind,
Monatsh., 33. 295, 1912 ; Le Radium, 9. 104, 1912 : 10. 174, 1913 ; 11. 108, 1914 ; W. Marckwald,
Ber., 36. 2662, 1903 ; F. Giesel, ib., 41. 1059, 1908 ; M. Curie and A. Debieme, Compt. Rend.,
150. 386, 1910 ; Le Radium, 7. 38, 1909 ; E. B. Andersen, Phys. Zeit., 10. 54, 1909 ; H. Stobbe
and H. MaUison, Ber., 46. 1226, 1913.
6 S. C. Lind, Sitzber. Akad. Wien, 120. 1709, 1911 ; Monatsh., 32. 295, 1911 ; Atner. Chem.
Journ., 47. 397, 1911 ; Le Radium, 9. 104, 1911 ; 11. 108, 1914; Trans. Amer. ElectrocJiem. Soc,
24. 339, 1913; Zeit. phys. Chem., 84. 759, 1913; Journ. Phys. Chem., 16. 564, 1912; Journ.
Amer. Chem. Soc, 41. 531, 551, 1919: P. Kriiger, NernsVs Festschrift, 240, 1911 ; Phys. Zeit.,
13. 1040, 1912 ; F. Kriiger and M. Moller, ib., 13. 729, 1912.
« C. F. Schonbein, Po^gr. Ann., 50. 616, 1840 ; A. W. Hoffmann, ib.,\Z2. 607, 1867 ; J. C. G. de
Marignac, Compt. Rend., 20. SOS, \M5 ; M. Berthelot, i&., 86. 71, 1878; P. Hautefeuille and
J. Chappius, ib., 91. 522, 875, 1880 ; J. L. Soret, ih., 56. 390, 1863 ; S. Edme, ib., 59. 291. 1864 :
G. Baumert, Ann. Chim. Phys., (3), 89. 38, 1853; T. Andrews, Phil. Tran.s., 146. 1, 1856;
J. Tyndall, ih., 152. 84, 1862 ; B. C. Brodie, Journ. Chem. Soc, 17. 293, 1864 ; H. McLeod, ib.,
49. 591, 1886; A. Rundspaden, Liebig's Ann., 151. 306, 1869; L. Carius, ih., 174. 1, 1874;
F. Richarz, Weid. Ann., 2A. 183, 1885 ; G. Targetti, Nuovo Cimento, (4), 10. 360, 1899.
' F. Fischer and K. Massenez, Zeit. anorg. Chem., 52. 202, 229, 1906 ; F. Fischer and
K. Bendixsohn, ib., 61. 13, 183, 1909 ; R. Kremann, ib., 36. 403,1903 ; L. Grafenberg, ib., 36.
355, 1903.
8 M. Topler, Ann. Physik, (4), 7. 477, 1902 ; G. Brion, Zeit. Elektrochem., 13. 761, 1907 ;
890 INORGANIC AND THEORETICAL CHEMISTRY
14. 245, 1908; W. Cramp and B. Hoyle, Electrochem. Ind., 7. 74, 1909 ; A. J. AUmand, The
Principles of Applied Electrochemistry ^ London, 183, 1912 ; W. G. Cady and H. D. Arnold, Phys.
ZeiL, 8. 890, 1907 ; H. Ayrton, The Electric Arc, London, 1902 ; H. de la Coux, Uozone et ses
application^ indmtrielles, Paris, 1904 ; E. Rasch Electric Arc Phenomena, New York, 1913.
» E. Briner and E. Durand, Compt. Rend., 145. 248, 1907.
10 E. Frhmj and E. Becquerel, Ann. Chim. Phys., (3), 35. 62, 1852 ; A. Houzeau, Compt.
Rend., 70. 1286, 1870 ; A. Behr, Ber., 20. 439, 1887 ; F. Fischer, Zeit. Elektrochem., 17. 535,
1911 ; F. Fischer and 0. Ringe, Ber., 41. 945, 1908 ; E. H. Reiser and L. McMaster, Amer. Chcm.
Journ., 39. 96, 1903 ; T. M. Lowry, Journ. Chem. Soc, 101. 1152, 1912.
11 P. Askenasy, Einfuhrung in die technische Elcktrochemie, Braunschweig, 212, 1910;
E. Warburg and G. Leithauser, Ann. Physik, (4), 28. 1, 1908.
12 E. Warburg. Sitzber. Akad. Berlin, 1011, 1903; E. Warburg and G. Leithauser, Ann.
Physik, (4), 20. 742, 1906 ; E. Warburg, ib., (4), 17. 7, 1905 ; (4), 13. 464, 1904 ; A. W. Gray, ib..
(4), 13.477, 1904; E. Regener, ib., (4), 20. 1033, 1906; W. Nemst, Ber. deut. elektrochem. Ges.,
38, 1894 ; E. Goldstein, Ber., 36. 3042, 1903.
1* E. Warburg, Ann. Physik, (4), 9. 781, 1900.
1* E. Goldstein, Ber., 36. 3042. 1903 ; P. Hautefeuille and J. Chappius, Compt. Rend., 91.
230, 1880 ; E. Briner and E. Durand, ib., 145. 248, 1907 ; A. Ladenburg, Ber., 34. 3849,
1901 ; A. Beill, Monatsh., 14. 71, 1893 ; V. Rothmund and A. Burgstaller, ib., 34. 565, 1913 ;
A. Mailfert, Bidl. Soc. Chim., (2), 34. 674, 1880.
15 E. Bichat and A. Guntz, Compt. Rend., 107. 344, 1888; Ann. Chim. Phys., (6), 19. 131,
1890 ; A. Vosmaer, Ozone, London, 45, 1916.
16 A. Chassy, Compt. Rend., 134. 1298, 1902.
1' A. W. Gray, Elektrotech. Zeit., 518, 1905 ; D. H. Kabakjian, Phys. Rev., 31. 117, 1910.
i« M. W. Franklin, Trans. Amer. Inst. Elect. Eng., 31. 985, 1912 ; M. Otto, Ann. Chim. Phys., (7).
13. 77, 1898 ; N. A. Puschin and M. Kauchtscheff, Journ. Russian Phys. Chem. Soc, 46. 576,
1914 ; G. Lechner, Zeit. Elektrochem., 17. 412, 1911.
i» E. Warburg and G. Leithauser, Ann. Physik, (4), 20. 751, 1906 ; E. Warburg, ih., (4), 13.
470, 1904 ; J. J. Thomson and R. Threlfall, Proc. Roy. Soc, 40. 340, 1885 ; D. J. Bohe, Ber.,
6. 439, 1873 ; W. A. Shenstone and M. Priest, ib., 63. 938, 1893 ; W. A. Shenstone and W. T.
Evans, ib., 73. 246, 1898 ; H. B. Baker, ib., 65. 611, 1894 ; W. A. Shenstone, ib., 71. 471, 1897 ;
D. L. Chapman and H. E. Jones, ib., 97. 2463, 1910 ; 99. 1811, 1911 ; E. P. Perman and R. H.
Greaves, Proc Roy. Soc, 80. A, 353, 1908 ; E. Fischer and H. Marx, Ber., 29. 3631, 1906.
20 M. Berthelot, Compt. Rend., 88. 50, 1879 ; P. Hautefeuille and J. Chappius, ib., 91. 762,
1880 ; W. A. Shenstone and W. T. Evans, Journ. Chem. Soc, 73. 246, 1898.
21 T. Andrews, Phil. Trans., 150. 127, 1860; W. A. Shenstone and W. T. Evans, Journ. Chem.
Soc, 73. 246, 1898.
22 L. von Babo, Liebig's Ann. Suppl, 2. 265, 1863 ; L. von Babo and A. Claus, ib., 2. 297,
1863 ; Liebig's Ann., 140. 348, 1866 ; A. Houzeau, Compt. Rend., 70. 1286, 1870 ; 74. 256, 1872 ;
A. W. Wright, Amer. Journ. Science, (3), 4. 26, 1872 ; Chem. News, 26. 213, 1872.
23 W. von Siemens, Pogg. Ann., 102. 120, 1857.
2* B. C. Brodie, Phil. Trans., 162. 435, 1872 ; H. Kobbe, Kurzes Lehrbuch der anorganischen
Chemie, Braunschweig, 106, 1877 ; M. Berthelot, Ann. Chim. Phys., (5), 10. 165, 1876 ; (5), 12.
453, 1877.
25 A. W. Gray, Phys. Rev., (1), 19. 362, 1904 ; F. Russ, Zeit. Elektrochem., 12. 409, 1906.
26 E. W. von Siemens and J. G.Halske, German Pat., D.R.P. 134,929, 1900 ; Zeit. Elektrochem.,
17. 535. 1911 ; H. Abraham and L. Marmier, ib., 6. 273, 434, 1900 ; O. Linder, U. S. Pat. No.
951, 443, 1910.
2' A. Schneller, Electrotech. Zeit., 592, 1891 ; U. S. Pat. No. 587770, 1897 ; A. Vosmaer,
ib., 6.36.304, 1898; 709427, 709379, 1902; 754261, 1904; 919403, 1909; H. Tindal, Ber.,
27. 675, 1894 ; 28. 1071, 1895 ; Zeit. Elektrochem., 6. 282, 1900 ; Ozonair Co., Chem. News, 102.
207, 1910 ; C. H. Jones, Chem. Met. Eng., 22. 805, 1920.
28 S. Pierre, Compt. Rend., 58. 420, 1864.
2» E. von Gorup-Besanez, Liebig's Ann., 161, 232, 1872; G. Bellucci, Ber., 8. 905, 1875; 9.
581, 1876 ; Gazz. Chim. Ital., 5. 88, 1875 ; H. Scoutetten, Compt. Rend., 42. 941, 1856 ; A. Poey,
ib., 57. 348, 1863 ; A. Morin, ib., 57. 720, 1863 ; D. Huizinga, Journ. prakt. Chem., (i). 102. 201,
1867 ; M. Monte, Ber., 8. 509, 1875 ; P. F. A. Ascherson, Ber. Akad. Munchen, 77, 274, 1877.
'0 P. Lenard, Wied. Ann., 46. 584, 1892 ; Lord Kelvin, Proc Roy. Soc, 47. 335, 1894.
»i C. F. Rammelsberg, Ber., 1. 73, 1868; Pogg. Ann., 134. 534, 1868 : i). T. Kingzett, Chem.
Netvs, 25. 242, 1872 ; A. R. Leeds, ib., 40. 257, 1879 ; 42. 304, 1880 ; H. Croft, ib., 25. 87, 1872 ;
Phil. Mag., (4), 43. .547, 1872 ; G. Bellucci, Ber., 8. 905, 1875.
32 0. Brunck, Zeit. anorg. Chem., 10. 222, 1895.
3' A. Riche, Bull. Soc Chim., (1), 2. 178, 1860; F. Ferraboschi, Proc Chem. Soc, 25. 179,
1909 ; Chem. Ztg., 1909 ; S. Cohne, Chem. News, 34. 4, 1876.
8* F. Raschig, Zeit. angew, Chcm., 18. 1293, 1905 ; 20. 708, 1907.
35 A. Riche, Bull. Soc Chim.., (1), 2. 178, I860 ; M. PnKrhomme,«6.,(3), 29. 260, .300, 1903 ;
A. Houzeau, Ann. Chim. Phys., (3), 62. 129, 1861 ; M. Otto, ih., (7), 13. 78, 1898 ; P. Malaquin,
Journ Pharm. Chim., (7), 3. 329, 1911 ; C. Arnold and i). Mentzcl, Ber., 35. 1324, 2902, 1902 ;
A. von Baeyer and V. Villiger, ib., 33. 1569, 1900 ; 34. 855, 1901 ; E. Bamberger, ib., 33. 1959,
1900 ; W. traube, ib., 46. 2513, 1913 ; J. Schmidlin and P. Massini, ib., 43. 1162, 1910 ; C. von
OZONE AND HYDROGEN PEROXIDE 891
Qirsewald and x\- Wolotkin, ih., 42. 808, 1909 ; P. G. Melikoff and L. Pissarjewsky, ih., 30. 2907,
1897 ; P. G. Melikoff and P. Kasanezky, Zeit. anorg. Chem., 28. 242, 1901 ; W. ManchQt, {6.,
39. 1352, 1906 ; J. d'Ans and A. Kneip, ib., 48. 1136, 1915 ; G. Bertazzl, Ntiovo Cimenio, 2. 291,
1855 ; R. Bottger, Journ. prakt. Chem., (1), 88. 377, 1862 ; B. Franke, ib., (2), 36. 166, 1887 ;
A. Baschieri, Oazz. Chim. Hal, 36. ii., 282, 1906; C. Weltzien, Liebig's Ann., 142. 107, 1867 ;
P. Malaquin, Journ. Pharm. Chim., (7), 3. 329, 1911 ; Ber., 34. 855, 1901 ; A. R. Leeds, Chem.
News, 39. 18, 1879 ; C. C. Frye, ib., 73. 122, 1896 ; L. W. Winkler, Zeit. anorg. Chem., 1. 84,
1892 ; T. E. Thorpe and F. J. Hambly, Journ. Chem. Soc, 53. 175, 1888 ; S. Tanatar, Ber., 33.
305, 1900 ; 36. 1893, 1903 ; 42. 1516, 1909.
38 H. Moissan, Ann. Chim. Phi^9., (6), 24. 224, 1891 ; Compt. Bend., 129. 670, 1899 ; Bull.
Soc. Chim., (3), 23. 259, 1900.
" 0. Ruff and F. W. Tschirch, Ber., 46. 929, 1913.
'8 L. Grafenbere:, Zeit. anorg. Chem., 36. 355, 1903; W. R. Grove, Proc. Roy. Soc, 17. 256,
1869 ; E. B. R. Prideaux, Traris. Faraday Soc, 2. 34, 1906.
39 C. F. Schonbein, Pogg. Ann., 65. 69, 1845; J. C. G. de Marignac, Compt. Bend., 20. 808,
1845 ; Ann. Chim. Phys., (3), 14. 252, 18l5 ; A. R. Leeds, Chem. News, 41. 163, 1880 ; 42. 17,
880 ; 43. 97, 1881 ; R. Bottger, Ber., 12. 2187, 1870 ; C. Arnold and C. Mentzel, ih., 35. 2902,
1902 ; R. Engel, Bvll. Soc. Chim., (2), 44. 426, 1885.
4» M. Berthelot, Compt. Rend., 84. 61, 1877 ; C. T. Kingzett, Chem. News, 40. 96, 1879 ; 42.
34, 242, 1880 ; 43. 127, 1881 ; A. R. Leeds, ib., 40. 70, 1879 ; 43. 97, 1881 ; H. McLeod, ib., 41.
163, 1880 ; 42. 17, 1880 ; Journ. Chem. Soc, 37. 118, 1880.
" P. Villard, Compt. Rend., 130. 125, 1900 ; L. Bloch, ib., 148. 782, 1909.
42 J. H. van't Hoff, Zeit. phys. Chem., 16. 411, 1895 ; W. G. Jorissen, ib., 23. 667, 1897.
" G. Bellucci, Ber., 12. 1700, 1879; R. Bottger, Polyt. Notizbl, 35. 95, 1880; J. Sehiel, Ber.,
12. 507, 1879.
** C. F. Schonbein, Ann. Chim. Phys., (3), 52. 221, 1858; H. Fudakowsky, Ber., 6. 106, 1873 ;
E. Schaer, ib., 6. 406, 1873 ; G. Bellucci, ib., 12, 1699, 1879.
*^ C. Engler and J. Weissberg, Ber., 31. 3046, 1898; Kritische Studien fiber die Vorgav^e der
Autoxydation,, Braunschweig, 1904.
4« C. Engler and W. Frankenstein, Ber., 34. 2933, 1901.
§ 3. The Occurrence of Ozone and Hydrogen Peroxide
According to A. Houzeau (1867), country air contains about one volume of
ozone per 700,000 volumes of air ; but a maximum of one part in ten millions would
be nearer the mark. The maximum amount of ozone in the atmosphere is said to
occur during the spring months, and to diminish gradually, reaching a minimum
in winter. The air over the sea is usually, but not always, richer in ozone than
air over land.i Ozone is absent in the air of towns and dwelling-houses, over
marshes, and wherever organic matter is present. It is really extraordinary the
number of determinations which have been made in order to find if the amount of
ozone predominates in spring, summer, autumn, winter, or at any particular part
of the day. Attempts have been made to show the efEect of the electrical and
hygrometric state of the atmosphere ; the force and direction of the wind ; the
intensity of the sunlight ; the geographical and geological formation of particular
districts ; etc. A. Houzeau himself is responsible for 4000 observations ; and
many others have been published. In some cases, the results are contradictory ;
and most of them should be discarded because much of what was formerly alleged
to be ozone may not be ozone at all. This is due to the imperfection of the tests
employed. For example, T. Andrews found that the oxidizing matter in the air
was destroyed by heating the air to 260°. This would not be the case if the oxidizing
matter was chlorine, nitrous acid or sulphur oxides ; but Andrew's test does not
discriminate between ozone and hydrogen peroxide. The old starch and potassium
iodide test-papers did not discriminate between ozone, hydrogen peroxide, nitrogen
oxide, or other oxidizing agents, hence there is some uncertainty about many of
the reports of the occurrence of ozone, and more particularly those referring to
the proportion of ozone in the atmosphere. For example, C. F. Schonbein gave
about 0'0043 mgrm. of ozone per 100 litres of air ; J. Pless and V. Pierre, 0'008
mgrm. ; C. W. Zenger, 0-002 to 0"01 mgrm. ; A. Houzeau, 0-0029 mgrm. ; M. de
Thierry, O'OOSS to 0-0094 mgrm. ; and K. Lespieau, 0*00052 mgrm.2
892 INORGANIC AND THEORETICAL CHEMISTRY
The more recent determinations by E. H. Kaiser and L. McMaster, and by
W. Hayhurst and J. N. Pring,8 have satisfactorily established the presence of ozone
in atmospheric air. The atmosphere of the Alps at an elevation of 6*5 kilometres
contains 1*2 volumes of ozone per million parts of air ; and at an altitude of 20
kilometres, 5*4 volumes per million. No hydrogen peroxide or nitrogen peroxide
was detected. There is also a quite adequate explanation of the formation of
atmospheric ozone. The dark electrical discharges from clouds, etc., and lightning,
and the action of ultraviolet radiations from the sun must all ozonize atmospheric
oxygen. W. N. Hartley ^ showed that from spectroscopic observations ozone
must be a regular constituent of the upper regions of the atmosphere where it is
present in larger proportions than nearer sea-level. The ozone formed in the upper
regions is decomposed by oxidizable substances in the lower regions, C. Fabry and
H. Buisson emphasize the fact that although the assumption that the high ozone
content of the upper atmosphere is not proved, the deduction is very probable
from the blue colour of the sky, and the abrupt break in the solar spectrum for rays
of wave-length 293/>t/A corresponding with the break in the absorption spectrum of
ozone. Gr. D. Liveing and J. Dewar question whether the blueness of the sky can
be justly attributed to ozone because the absorption spectrum of oxygen exhibits
certain bands identical with those of the solar spectrum, which K. Angstrom ^
found to be equally strong whether the atmosphere be wet or dry ; and that day-
light possesses a blue tint when observed through a 18-metre layer of oxygen
compressed at 90 atm.
The presence of ozone in solution in certain spring waters has been established
by R. Nasini and M. G. Levy, and C. Porlezza.^ The spring at Fiuggi and the aqua
forte della Bagnori of Monte Amiata, gave the qualitative reactions of ozone with
starch and potassium iodide, and litmus-potassium iodide test papers ; guaiacum
test paper ; acid-free gold chloride, manganous chloride, C. Arnold's tetramethyl
base test paper; and with silver foil. At 21-50°, a litre of the water contained
0'135 c.c. of ozone in solution. It is suggested that the ozone may be due to the
action of radioactive rocks, or to the autoxidation of ferrous carbonate, or to sulphur
bacteria. A. Schrotter ^ (I860) reported 002 per cent, of ozone — identified by
the smell — occluded in the blue fluorspar of Wolsendorf and Joachimsthal, but
A. Houzeau and 0. Loew considered this inference to be wrong because the smell
attributed to ozone is retained after fluorspar has been heated to 300°. C. F. Schaf-
hautel attributes the smell to the presence of inclusions containing free hypo-
chlorous acid ; and J. Meyer supposes the mineral contains free fluorine as an
inclusion, and this, with traces of moisture from the air, forms ozone. Reports
of the presence of ozone in blood (A. Schmidt, 1862), in milk (C. Arnold, 1881), and
in respired air (A. Struve, 1871) are probably wrong, and they are based only on
old and fallacious tests. Some physiologists assert that the oxygen given ofi by
green plants in light contains ozone,^ but G. Bellueci contradicts this statement.
The occurrence of hydrogen peroxide. — There is a similar uncertainty
about the alleged occurrence of hydrogen peroxide in the atmosphere and rain,
snow and in dew, as in the case with early reports on the occurrence of ozone in air.
C. F. Schonbein, E. Schone,^ and others claim hydrogen peroxide to be present in
the atmosphere and in rain water ; A. Houzeau lo and L. I. de N. Ilsova say that
no hydrogen peroxide is present, and the latter adds that what was thought to be
hydrogen peroxide is really an oxide of nitrogen. It is too true that much of the
published work does not clearly discriminate hydrogen peroxide from other oxidizing
substances. In any case the amount of hydrogen peroxide must be very small —
between 0'04 and 1*00 mgrm. per litre of rain water. In a year's observation
at Moscow, E. Schone reported 110 mgrm. of hydrogen peroxide in the rain and snow
which fell per sq. metre. According to A. Bach,ii when the chlorophyll of plants
acts on carbon dioxide and water in sunlight, formaldehyde, CH2O, and percarbonic
acid, H2CO4, are formed, 3H2C03=2H2C04+CH20 ; and the percarbonic acid breaks
down into carbon dioxide and hydrogen peroxide, H2C04=C024-H202 ; and the
OZONE AND HYDROGEN PEROXIDE 893
latter in turn forms water and oxygen gas. A. Bach recommends the following
reagent for detecting peroxides in plants : 5 c.c. of a solution of 0*03 grm. potassium
dichromate and 5 drops of aniline with sufficient water to make a litre of solution.
Treat 5 c.c. of the solution to be tested with one drop of a 5 per cent, solution of
oxalic acid and the given reagent. A reddish-violet colour will be obtained if
peroxides are present. Of 25 plants examined, 18 gave the reaction for peroxides.
J. Cho (1896) obtained the coloration only where leaves were damaged. The
existence of hydrogen peroxide in vegetable juices has been reported by many
observers,i2 but Gr. Bellucci could not confirm this by the chromic acid reaction,
and he attributes the results obtained to the presence of oxydases in vegetable sap.
C. Wurster made similar remarks with respect to the alleged presence of hydrogen
peroxide in animal fluids. is
References.
I H. E. Schelenz, Arch. Pharm., (3), 27. 224, 1899 ; A. Houzeau, Ann. Chim. Phys., (3), 62.
129, 1861 ; (3), 67. 466, 1863 ; (4), 7. 84, 1865 ; (4), 27. 14, 1872 ; CompU Rend., 40. 947, 1855 ;
43. 34, 1856 ; 45. 873, 1857 ; 46. 89, J 858 ; 50. 829, 1860 ; 57. 798, 1864; 60. 788, 1865; 61.
1113, 186G ; 62. 426, 1866 ; 66. 314, 491, 1868 ; 70. 369, 1286, 1870 ; 74. 242, 256, 712, 1872.
« C. F. Schonbein, Journ. prakt. Chem., (1), 56. 349, 1852; Phil. Mag., (4), 4. 545, 1862;
Liehig's Ann., 89. 257, 1854 ; Pogg. Ann., 72. 463, 1847 ; 75. 366, 1848 ; Ueber die Erzeugung
des Ozons auj chemifichen Wege, Basel, 1844 ; J. Pless and V. Pierre, Sitzber. Akad. Wien, 22.
211, 1856; C. W. Zenger, ib., 24. 78, 1857; A. Houzeau, Ann. Chim. Phys., (4), 27. 5, 1872;
Oompt. Bend., 74. 712, 1872 ; M. de Thierry, ib., 124. 406, 1897 ; R. Lespieau, Bull. Soc. Chim.,
(3), 35. 616, 1906 ; T. Andrews, Phil. Trans., 146. 1, 1856.
3 E. H. Kaiser and L. McMaster, Amer. Chem. Journ., 39. 96, 1908 ; W. Hayhurst and J. N.
Pring, Journ. Chem. Soc, 97. 868, 1910 ; Proc. Boy. Soc, 90. A, 204, 1914.
* W. N. Hartley, Journ. Chem. Soc, 39. 67, 111, 1881 ; E. Schone, Zeit. anorg. Chem., 6.
333, 1894 ; P. Lenard, Ann. Physik, (3), 51. 232, 1894 ; (4), 1. 503, 1900 ; Arkiv. Math. Astron.
Fysik, 1. 395, 1904.
' 6 E. Mayer, ib., (4), 12. 849, 1903 ; C. Fabry and H. Buisson, Compt. Bend., 156, 782, 1913 ;
J. Chappius, ib., 91. 985, 1880 ; 94. 858, 1882 ; G. D. Liveing and J. Dewar, Phil. Mag., (6), 26.
286, 1888 ; K. Angstrom, Arkiv. Math. Astron. Fysik, 1. 347, 395, 1882.
« R. Nasini and M. G. Levy, Gazz. Chim. Ital., 38. i, 190, 1908 ; C. Porlezza, ib., 43. i, 176,
1913 ; Atti Accad. Lincei, (5), 21. ii, 740, 1912.
' A. Schrotter, Pogg. Ann., 111. 561, 1860 ; A. Houzeau, Bull. Soc Chim., (2), 2. 14, 1864 ;
0. Loew, Ber., 14. 1144, 1881 ; C. F. Schafhautl, Liebig's Ann., 46. 344, 1843 ; Journ. prakt.
Chem., (1), 76. 129, 1869 ; J. Meyer, ib., (2), 72. 278, 1905 ; H. Moissan and H. Becquerel, Compt.
Bend., 111. 669, 1890.
« E. von Gorup-Besanez, Liebig's Ann., 161. 232, 1872 ; R. Wolf, Compt. Bend., 40. 419,
1855; H. Scoutetten, ib., 42. 941, 1856; 43. 93, 216, 1856; G. de Luca, ib., 43. 865, 1856 ;
C. Arnold, Arch. Pharm., (3), 19. 41, 1881 ; A. Struve, Zeit. anal. Chem., 10. 292, 1871 ; A. Schmidt,
Ueber Ozon im Blut, Dorpat, 1862 ; P. C. Kosmann, Compt. Bend., 55. 731, 1862 ; F. S. Cloez, ib.,
43. 38, 462, 1856 ; A. Poey, ib., 57. 544, 1863 ; G. BeUucci, ib., 78. 352, 1874 ; Ber., 12. 1699,
1879.
» C. F. Schonbein, Bepert. Pharm., 13. 364, 1869 ; W. Schmitt, Jowrw. prakt. Chem., (1), 107.
60. 1869 ; F. Goppelsroder, ib., (2), 4. 139, 389, 1871 ; H. Struve, ib., (1), 107. 603, 1869 ; Compt.
Bend., 68. 1551, 1869 ; E. Schone, Ber., 7. 1693, 1874 ; 10. 482, 564, 1028, 1877 ; 12. 346, 1879 ;
13. 1503, 1880 ; 26. 3011, 1893 ; 27. 1233, 1894.
10 A. Houzeau, Compt. Bend., 66. 314, 1868 ; 70. 619, 1870 ; L. I. de N. Ilsova, BuU. Soc
Chim., (3), 2. 377, 666, 734, 1839 ; Ber., 27. 920, 1894.
II A. Bach, Compt. Bend., 116. 1145, 1893 ; 118. 286, 1218, 1894 ; Ber., 27. 340, 1894 ; A. Bach
and R. Chodat, ib., 35. 2466, 1902 ; J. Cho, Bull. Coll. Agric Tokyo, 2. 225, 1896.
12 P. de Clermont, Compt. Bend., 80. 1591, 1875 ; A. Bechamp, ib., 94. 1601, 1882 ; P. Bert
and P. Regnard, ib., 94. 1383. 1882 ; E. Griessmayer, Ber., 9. 835, 1876 ; G. Bellucci, Gazz. Chim.
Ital, 8. 392, 1878 ; Ber., 12. 136, 1879.
i» C. Wurster, Ber., 19. 3206, 1886 ; 20. 2934, 1887.
§ 4. The Physical Properties of Ozone
Ozonized air has a strong characteristic smell, which reminds some people of
sulphur dioxide, others of garlic, and others of chlorine. The amount of ozone
894 INORGANIC AND THEORETICAL CHEMISTRY
which can be detected by the olfactory sense is extremely small ; one part in a
million parts of air can be readily perceived. Indeed, ozone can be detected by
smell before starch and potassium iodide paper turns blue. If air highly charged
with ozone be breathed for any length of time, it produces headache ; but in minute
quantities the odour is pleasing and refreshing. In large quantities, ozonized air
acts as an irritant poison causing headache and coughing, and finally infiammatiou
and death ; in small quantities, ozonized air has been recommended medicinally
for pulmonary complaints. According to D. Labbe and S. M. Oudin, air containinj^
one part of ozone in 20,000 may be breathed half an hour without ill effects. L. E.
Hill and M. Flack ^ say that a concentration of one part in a million irritates the
respiratory tract ; exposure for two hours to a concentration of 15 to 20 per million
is not without risk of life. In concentrations even less than one per million, it reduces
respiratory metabolism, and rapidly causes a fall of body temperature. Its bene-
ficial effect, as popularly believed, is a myth. The irritation of the olfactory
nerves may relieve the monotony of close air, and in concentrations of more than
1 per million for brief periods may be of therapeutic value by acting in appropriate
cases as a sort of " blister " to the respiratory tract.
Ozone at ordinary temperatures is a gas with a pale blue colour. Oxygen with
10 per cent, of ozone has a blue tinge when viewed through a tube a metre long.2
Liquid ozone is dark indigo-blue. C. F. Schonbein 3 wrote to M. Faraday (1852)
to the effect that he had made experiments supporting the hypothesis that the
colour of oxy-compounds is due to the contained oxygen, or to a peculiar condition
of that body. To this M. Faraday replied :
Your letter quite excites me and I trust you will establisli undeniably your point. It
would be a great thing to trace the state of combined oxigen by the colour of its compound,
not only because it would show that the oxigen had a special state, which could in the com-
pound produce a special result — but also because it would, as you say, make the optical
effect come within the category of scientific appliances and serve the purpose of a philo-
sophic induction and means of research, whereas it is now simply a thing to be looked at.
Believing that there is nothing superfluous, or deficient, or accidental, or indifferent, in
nature I agree with you in believing that colour is essentially connected with the physical
condition and nature of the body possessing it, and you will be doing a very great service
to philosophy if you give us a hint, however small it may seem at first, in the development,
or as I may even say in the perception of this connexion.
A litre of ozone weighs 2*14:45 grms. at n.p.t. ; and a gram of ozone under the
same conditions occupies 468"3 c.c. The specific gravity is 1*5 (oxygen unity),
or 162 (air unity). P. Hautefeuille and J. Chappius (1880) liquefied ozone by
gradually compressing the gas in Cailletet's apparatus, at —23°, if the compression
be sudden, or without cooling, the ozone is converted into oxygen with a yellow
flash. P. Hautefeuille and L. Chappius liquefied ozone under a pressure of 125 atm.
at —100° ; K. Olschewsky * under atmospheric pressure at — 181°. By passing
ozonized oxygen through a tube cooled by immersion in boiling liquid oxygen, or
by ozonizing oxygen in a tube kept at this temperature, a solution of ozone in
liquid oxygen is obtained. By allowing the liquid to boil, most of the oxygen is
removed, and a deep indigo-blue, almost black, liquid remains, which is opaque
in layers of 2 mm. thick. By allowing the blue liquid to vaporize, A. Ladenburg
(1898) obtained a gas containing about 86 per cent, of ozone. The liquid is par-
ticularly liable to explode when it reaches the boiling point of ozone, or when it
is brought in contact with oxidizable substances. According to H. Erdmann,
pure liquid ozone is not explosive, and he ascribes the explosibility to the presence
of highly concentrated ozone gas ; he says liquid ozone is harmless, and he obtained
Leidenfrost's phenomenon with a drop of the liquid on a porcelain plate. The
indigo-blue vapour of ozone in a test tube explodes if a little turpentine is introduced.
G. D. Liveing and J. Dewar attributed the explosion of liquid ozone at —181° to the
presence of the vapour.
The boiling point of ozone under atmospheric pressure is —106° (K. Olzschewsky) ;
OZONE AND HYDROGEN PEROXIDE 895
—116° (L. Troost) ; or —125° (A. Ladenburg). The ratio of the two specific heats
GpjCv for ozone is 1*29 as determined ^ by extrapolation from the value 1"396 for
oxygen, and the observed values of mixtures of oxygen and ozone. G. N. Lewis and
M. Randall ^ say that although the heat capacity of ozone has not been accurately
determined, no great error will be involved by assuming that -the relation which
holds good for the triatomic gases, CO2 and SO2, is applicable, and therefore Cp
=7-0+0'0O7ir— 0-00000186T2. Subtracting theheatcapacityfor 1 J gram-molecules
of oxygen from Op+6*50-|-0'0010T leaves for the heat Q of the reaction 1J02=03,
at constant temperature : ^=^0— 2*75r+0-0028J2_o-00000062jr3.
According to M. Berth elot (1876), the heat of formation of a gram-molecule of
ozone, O3, is —296 Cals. ; — 36'3 to —36-65 Gals., according to H. G. L. van der
Meulen (1883) ; and —34-5 Cals., according to S. Jahn (1908). The reported
heats of the reaction thus range from 29 to 36 Cals. ; if 34,600 cals. be the best
representative value,the heat of the reaction at Twill be^=34600— 2-75T-l-0'0028T2
— 0-00000062T3 ; and the increase in free energy 34600+2-75T log T— 0-C028T2
+0*00000031 jT^-f-ZT, where the integration constant I can be evaluated very crudely
from F. Fischer and F. Brahmer's work, K—pi/p'^'^, where pi denotes the partial
pressure of ozone, and p that of oxygen ; at 2300°, /!l=0*01. Hence, the increase
in free energy at 2300° is —RT log ^+21000 cals. When this is substituted in
the preceding equation Z=— 22*4, and at 298° K. the increase in free energy is
32,400 cals.
R. Luther and H. J. K. Inglis studied the potential of the ozone electrode by
means of a platinum electrode surrounded by ozone, but they were unable to de-
termine definitely the nature of the electrode reaction ; and with an iridium electrode,
R. Luther obtained values differing by 02 volt from those with the platinum
electrode. S. Jahn found the free energy of the decomposition of ozone, de-
termined from measurements of the potential of the cell O3 j electrolyte | H2,
which is 1-90 volts at 0° ; and of the cell O2 | electrolyte | H2, which is 1'25 volts.
In the former case, the free energy of the reaction is 203+2H2=202+H20
+(4Jxl-90) joules; and in the latter case, 02+2H2=2H20 +(4^x1-25) joules,
when F denotes the electrochemical equivalent, 96,540 coulombs. By subtraction,
eliminating HgO, there remains 2O3— 3O2+(4jFx0'65) joules: or O3=|O2+30
Cals. nearly.
Ozone has a high absorptive and emissive power for ultra-red heat radiations ;
according to J. Tyndall,^ its absorption power is 136 times as great as oxygen.
K. Angstrom found that there is a sharp ultra-red absorption band at 4'8jLt ; a feeble
one at 5*8/x ; an uncertain one at QIjjl ; and a strong — perhaps double — band at
9*1 to 10'0/x. The first and last occur in the solar spectrum. J. Chappius and
E. Schone found that the absorption spectrum of ozone shows thirteen bands and
lines : (1) A narrow band of wave-length 628"5/x,/x in the red ; (2) a very large band
from 609*5 to 593*5 in the orange ; (3) a very large band, 577 to 560, and (4) one at
547 to 544*5 in the yellow ; (5) a large one between 535 to 527 ; (6) one at 508*5 to 502,
and (7) one at 492*5 to 491 in the green ; (8) one at 484*5 to 479, and (9) one at 470
to 468 ; (10) one at 464*5 to 460, and (11) at 444 in the blue. There is also one at
(12) 452 and (13) one at 516. Nos. 2 and 3 are very marked ; Nos. 5, 6, and 8 are less
characteristic : Nos. 10 and 11 still less characteristic ; Nos. 12 and 13 still less so ;
Nos. 1, 4, 7, 9, 12, and 13 can be recognized only under special conditions. Liquid
ozone does not show any absorption bands in the visible spectrum, but there is a
masking in the region at about 500/x/x ; there are no bands in the ultraviolet, and the
absorption extends into the visible region as the concentration increases, but not so
far as in the case of the gas. The absorption spectrum of the gas in the ultraviolet
is so marked that, as previously indicated, W. N. Hartley supposed that in passing
through the atmosphere the 293)Lt/x radiations from the sun are suppressed. The
hypothesis was confirmed by the work of E. Meyer and A. Levy. From Lambert's
law, /=/oXlO~*^ when/o denotes the intensity of the incident energy, and / that
after traversing a layer of ozone of thickness I, the absorption constant a of ozone
896
INORGANIC AND THEORETICAL CHEMISTRY
at 0° and 760 mm. in H. Kreusler's photometer is, for light of wave-length A, and
with the centimetre as unit of length, E. Meyer found :
A
. 193
200
220
240
260
280
300/x/z
a
. 11-7
7-8
19-2
105
126
73-4
30-3
There is thus a maximum near A=260jLt/x, and a minimum near X—lSfjufi. A. Levy
estimated that the atmosphere contains 07696 XlO""^ per cent, of ozone by volume.
The corresponding absorption can be calculated. E. Meyer found for Iq the
intensity of the incident radiant energy before it enters the atmosphere, and /,
the intensity at the earth's surface, on the assumption that for A=300)Lt/x, the
intensity of the radiant energy =100,
A
. 193
200
220
240
260
280
ZOOfifx
^0 .
. 16-76
20-75
34-36
50-43
67-54
76-07
100
I .
. 14-21
18-59
26-19
11-46
11-28
]4-73
65-10
100
180 200 220 240 260 280 300
These curves are plotted in Fig. 6. The intensity of the radiant energy received at
the earth's surface owing to absorption by ozone rapidly falls from SOOft/x down to
about 260ju,jLt, and then rises again. The beam of radiant energy from the sun is
assumed to be directed normally to the earth's surface ; in reality, the beam is more
or less slanting, this could make the absorption greater than the calculated value —
even if the data were otherwise correct. The
absorption band commencing near A=200/x/>t is
probably responsible for the de-ozonizing effect
of ultraviolet rays ; and the rays below 170jLt/x,
which V. Schumann found to be entirely absorbed
by a layer of air 1 mm. thick and 760 mm.
pressure, are probably responsible for the forma-
tion of ozone, so that the stability of ozone is a
function of the active mass of oxygen, and of
the intensity of the ultraviolet rays beyond, say,
180/xu. K. Stuchtey examined the spectrum of the
Fig. 6.-The Effect of Atmospheric i^^linous glow on an ozone tube.
Ozone on the Radiant Energy ^5 _
received by the Earth. E. Ladenburg and E. Lehmann have observed
a second absorption spectrum in the ozone remain-
ing after much of a mass of liquid ozone has evaporated; this spectrum has
some bands in the red portion. It has been suggested that this is due to the
presence of C. D. Harries' oxozone — but no satisfactory proof has been yet
adduced.
H. Becquerel 8 found the Specific magnetism of ozone to be very high.
M. Berthelot found that ozone is stable under the influence of sound waves.
Solubility .^ — ^H. Erdmann ^ found that liquid nitrogen dissolves ozone,2 forming
a clear sky-blue liquid. M. A. Hunter attempted to measure the molecular weight
of ozone dissolved in liquid oxygen and in liquid nitrogen. The last-named solution
was too explosive. He found that solutions of ozone in oxygen have a minimum
boiling point, — 183'3°, when 4 to 5 per cent, of ozone is present — thus :
6-5 7-5 per cent.
-182-75° -182-63^
-Many measurements have been made of the solubility of ozone in water. C. F.
Schonbein ^^ said that ozone is not perceptibly soluble in water — this statement is
incorrect, since ozone is appreciably soluble in this menstruum ; E. Schone (1873)
found that when ozonized oxygen is passed into distilled water, the concentration
of the ozone remains constant, although about one-fourth of the ozone is decomposed ;
hence, H. J. K. Inglis argues that the concentration of the ozone must depend on
the rate the gas is passed through the solvent, and consequently the solubility
Ozone
0
2
4
5
Boiling point .
-182-8°
-182-85°
-183-3°
-183-25
OZONE AND HYDROGEN PEROXIDE
897
coefficient cannot be accurately determined. However, many determinations of
the solubility have been attempted. According to A. Ladenburg, a litre of water
at 0° dissolves 20 mgrm. of ozone ; at 2°, 10 mgrm. ; at 28°, lb mgrm. A. Mailfert's
results for the equilibrium conditions in the partition of ozone between gas and
liquid are indicated in Table II.
Table II.- — -Solubility
OF Ozone
tN Water.
Temperature.
Mjrrm. ozone
per litre
solvent.
Mgrm. ozone
per litre
gas.
Ratio.
0-641
Temperature.
27°
Mgrm. ozone
per litre
solvent.
Mgrm. ozone
per litre
gas.
Ratio.
0° '
39-4
61-5
13-9
51-4
0-270
6°
34-3
61-0
0-562
33°
7-7
39-5
0195
11-S°
29-9
59-6
0-500
40°
4-2
37-6
0112
130^
28-0
58-1
0-482
47°
2-4
31-2
0077
150°
25-9
56-8
0-456
55°
0-6
19-3
0031
190°
210
55-2
0-381
60°
00
12-3
0-000
The solubility decreases with a rise of temperature ; and according to 0. Froh-
lich, increases with an increase of pressure. L. Carius reported that the absorption
coefficient of ozone prepared by electrolysis is 0*834 at 1° ; and for ozone prepared
by the electric discharge 0'635 at 1°. A. Mailfert, E. Moufang, F. Biirger, and
V. Rothmund and A. Burgstaller found that the solubility of ozone in water is aug-
mented by the addition of acids. ^^ The presence of oxidizable substances must be
avoided. R. Luther and J. K. H. Inglis find that the absorption coefficient for water
at 0° is 0-494 ; and for —iV-HgSO^, at 0°, 0-487, so that the solubility appears to
decrease slightly with increasing concentration of the acid ; but the rate of decom-
position of ozone in water increases with decreasing acid concentration. The
solution of ozone in water soon decomposes, so that aqueous solutions cannot be used
for the determination of the molecular weight of this gas. Solutions of ozone in
Y\^iV-H2S04 follow Henry's law. The solubility of ozone in neutral salt solutions —
e.f). sodium or magnesium chloride. — is in many cases greater than in water, and
the solutions are more stable. B. Graf's patent for stabilizing aqueous solutions
of ozone is based on this fact. The solubility diminishes in feebly alkaline solutions
— say 0*005 and 0-00002 iV-solutions of sodium carbonate ; in concentrated solutions
of the alkalies, the ozone is rapidly decomposed. No signs of the formation of
hydrogen peroxide has been observed in the decomposition of acid or alkaline
solutions of ozone. E. Langheld i^ recommends quinine salts for increasing the solu-
bility and stability of aqueous solutions of ozone. Contrary to F. Jeremin's views,
R. Bottger holds that oxalic acid does not make ozone solutions more stable, since
it is readily oxidized by ozone. Acetaldehyde and, better still, paraldehyde have
been recommended by S. Eraser for stabilizing solutions of ozone.
Ozone is apparently dissolved by essential oiYs— turpentine, thyme, cinnamon
oil, etc. According to J. L. Soret,^^ these oils absorb ozone from ozonized oxygen
and leave the oxygen unaffected. B. Stelzer found ozone is copiously absorbed by
fats, C. T. Kingzett by ethereal oils. C. D. Harries assumes that the action is
here not a true solution, but rather a case of chemical combination ; addition
compounds are formed called ozonides. According to E. Molinari (19C6), un-
saturated carbon compounds with a double or ethylene bond between the carbon
atoms appear to absorb ozone quantitatively forming ozonides, whereas those with
a triple or acetylene bond do not combine with ozone. C. D. Harries, however,
does not accept E. Molinari's conclusion, since some compounds with a triple bond
combine with ozone more rapidly than some with a double bond. CD. Harries
used ozonized oxygen, E. Molinari ozonized air. C. D. Harries and R. Koetschau,
H. Erdmann, and F. Fischer and H. Tropsch ^^ find that a blue solution is obtained
when ozonized oxygen is passed into acetic acid, acetic anhydride, ethyl acetate,
VOL. I. 3 m
898 INORGANIC AND THEORETICAL CHEMISTRY
chloroform, or carbon tetrachloride. The colour persists for 15 to 20 hrs. with acetic
acid and carbon tetrachloride, but disappears more rapidly with the other solvents.
Water and formic acid do not form coloured solutions ; the latter is oxidized to
carbon dioxide. Carbon tetrachloride dissolves seven times as much ozone as an
equal bulk of water.
The so-called ozone ivater of commerce usually contains no ozone at all.
R. Bottger 15 found nitrous acid ; E. A. Behrens and G. Sonntag, hypochlorous acid;
C. F. Rammelsberg, L. Keutmann, H. Thoms, and C. G. Egeling, chlorine ; and
G. Vulpius, chloride of lime in commercial ozone water.
Kefebenges.
1 L. E. Hill and M. Flack, Proc. Roy. Soc, 84. B, 404, 1911 ; D. Labbe and S. M. Oudin,
Conipt. Bold., 113. 141, 1891.
2 P. Hautefeuille and J. Chappius, Comjd. Ee)id., 91. 522, 1880.
3 G. W. A. Kahlbaum and F. V. Darbishire, TJie Letters of Faraday and Schoenbein, London,
205-209, 1899.
4 J. Olschewsky, Moruitsh., 8. 230, 1887; A. Ladenburg, Ber., 31. 2508, 1898; L. Troost,
Compt. Bend., 126. 1751, 1898 ; H. Erdmann, Ber., 37. 4739, 1904 ; G. D. laveing and J. Dewar,
Phil. Mag., (5), 34. 205, 1892.
» F. Richarz and A. Jacobs, Ann. Physik, (4), 19. 639, 1909.
« T. Wood, Phil. Mag., (4), 28. 106, 1864 ; A. F. Hollemann, Arch. Neerl. Sciences, 3. 26C,
1868 ; M. Berthelot, Compt. Bend., 82, 1281, 1876 ; Ann. Chim. Phys., (5), 10. 162, 1877 ;
S. .Tahn, Zeit. anorg. Chem., 60. 337, 1908 ; S. Jahn and A. Kailan, ib., 68. 243, 1910 ; G. N. Lewis
and M. Randall, Journ. Amer. Chem. JSoc, 36. 1969, 1914; 34. 1128, 1912; F. Fischer and
F. Brahmer, Ber., 39. 940, 1906 ; R. Luther, Zeit. Elektrochem., 11. 832, 1905 ; R. Luther and
J. K. H. IngUs, ib., 8. 645, 1902; Zeit. phvs. Chem., 43. 203, 1903 ; L. Grafenbcrg, Zeit. anorg.
Chem., 36. 355, 1903 ; F. Fischer and K. Massenez, ib., 52. 202, 229, 1907.
' W. N. Hartley, Journ. Chem. Soc, 39. 57, 111, 1881 ; Nature, 39. 474, 1889 ; 30. 394, 1884 ;
P. Hautefeuille and J. Chappius, Compt. Bend., 91. 985, 1880 ; 94. 853, 1882 ; Bull. Soc. Chim.,
(2), 35. 2, 1881 ; K. Angstrom, Arkiv. Math. Astron. Fysik, 1. 347, 395, 1904; J. Chappius,
Ann. Jtcole Norm. Sup., (2), 11. 137, 1882 ; E. Schone, Journ. Bussian Phys. Chem. Soc,
16. 250, 1884; Zeit. anorg. Chem., 6. 333, 1894; E. Meyer, Ann. Physik, (4)," 12. 849, 1903;
E. Ladenburg and E. Lehmann, ib., (4), 21. 305, 1906; H. Kreusler, ib., (4), 6. 419, 1901;
C. Fabry and H. Buisson, Compt. Bend., 156. 782, 1913 ; J. Tyndall, Heat a Mode oj Motion,
London, 333, 1863; PM. Trans., 152. 59, 1862; A. L6vy,Cieletterre, 19. 297, 1898 ; V. Schumann,
Smithsonian Contrib. Knowledge, 1413. 29, 1903; K. Stuchtey, Zeit. iviss. Phot., 19. 161,
1920.
8 H. Becquerel, Comi^t. Bend., 92. 348, 1881 ; M. Berthelot, ih., 90. 487, 1880.
» H. Erdmann, Ber., 39. 1208, 1906 ; M. A. Hunter, Journ. Phys. Chem.., 10. 330, 1906.
i» G. F. Schonbein, Fogg. Ann., 66. 293, 1845; Neue>i Bepert. Pharm., 14. 289, 1866;
M. Berthelot, Compt. Bend., 90. 656, 1880 ; H. J. K. Inglis, Jourii. Chem. Soc, 83. 1010,
1903; E. Schone, Ber., 6. 1224, 1873; L. Carins, Ber., 25. 520, 1872; 6. 806, 1873; 7.
1481, 1874 ; Liebig's Ann., 174. 1, 1874 ; A. Ladenburg, ib., 31. 3508, 1898 ; A. Mailfert,
Compt. Bend., 119. 951, 1894; E. Moufanir, Wochschr. Brauerei, 29. 434, 1911; 0. Frohlicb,
Promethem, 2. 625, 1891 ; A. Houzeau, Ann^ Chim. Phys., (4), 27. 15, 1872 ; J. C. G. dc Marignac,
ib., (3), 14. 254, 1845 ; Compt. Rend.., 20. 808, 1845 ; M. Berthelot, ib., 90. 656, 1880 ; J. L. Soret, ib.,
56. 390, 1863 ; Phil. Mag., (4), 25. 209, 1863 ; T. Andrews, Phil. Trans., 146. 1, 1856 ; C. Hoff-
mann, Fogg. Ann., 132. 617, 1867; C. Gianitti and A. Volta, Gazz. Chim. Ital, 4. 471, 1874;
R. H. Ridout, Chem. News, 41. 73, 1880 ; G. Meissner, Untersuchungen iiber den Sauerstoff, Han-
over, 1863 ; C. Engler and A. Nasse, Liebig's Ann., 154. 215, 1870 ; L. Carius, Ber., 5. 520, 1872 ;
6. 806, 1873 ; 7. 1481, 1874 ; C. F. Rammelsberg, ib., 6. 603, 1873 ; A. E. Leeds, ib., 12. 1831,
1879; L. I. de N. Ilsova, ib., 27. 3500, 1894; V. Rothmund, Nernst's Festschrift, 391, 1912 j
A. W. Williamson, Journ. Chem. Soc, 22. 360, 1869.
11 A. Mailfert, Compt. Bend., 119. 951. 1894 ; E. Moufang, Wochschr. Brauerei, 29. 434, 1911 ;
F. Burger, ib., 30. 285, 1913; V. Rothmund and A. Burgstaller, Monatsh., 34. 665, 1913;
R. Luther and J. K. H. Inglis, Zeit. phys. Chem., 43. 203, 1903 ; R. Luther, Zeit. FAcktrocMm.,
11. 832, 1905.
12 B. Graf, German Pat. D.B.P., 52452, 1890; Zeit. angew. Chem., 3. 494,1890; S. Fraser,
ib., 23. 84, 1910; German Pat. D.B.P., 216092, 1908; E. Langheld, Chem. Ztg., 22. 212,
1898 ; F. Jeremin, Ber., 11. 988, 1878 ; R. Bottger, Jahresber. Phys. Ver. Frankfurt, 24, 1878.
13 J. L. Soret, Ann. Chim. Phys., (4), 7. 113, 1866; B. Stelzer, Pharm. Centralh., 38. 453,
1897 ; C. T. Kingzett, Journ. Soc Chem. Ind., 12. 511, 1893 ; C. D. Harries, Liebig's Ann., 343.
311, 1905 ; E. Molinari, Ber., 40., 4154, 1907 ; 41. 585, 2782, 1908.
1* F. Fischer and H. Tropsch, Ber., 50. 765, 1917; C. D. Harries and R. Koetechaii, /''
42. 3305, 1909 ; H. Erdmann, Liebig's Ann., 362. 133, 1908.
OZONE AND HYDROGEN PEROXIDE 899
15 W. Waldmann and R. Bottger, Pharm. Centralhalle, 13. 114, 1872 ; L. Keutmann, ib., 30. 750,
1889 ; H. Thorns, ib., 31. 68, 1890 ; E. A. Behrens, Dingier' s Jourv., 208. 78, 1873 ; G. Sonntag,
Zcit. Hygiene, 8. 95, 1890 ; 0. F. Rammelsberg, Ber., 5. 603, 1873 ; C. G. Egeling, Apoth. Ztg., 4.
295, 1889; G. Vulpius, Archiv. Pharm., (3), 22. 268, 1884.
§ 5. Oxozone, Ozonides, and Oxozonides
It was thought for some time that ordinary oxygen is a compound of negative
and positive oxygen, the former was called by C. F. Schonbein ozone and symbolized
0, and the latter antozone, symbolized © ; so that ordinary oxygen=ozone 0
-f antozone 0. The existence of the two different forms has not been satisfactorily
proved, and the term ozone is reserved for polymerized oxygen O3, while the term
antozone is not used. C. F. Schonbein used the term ozonides for certain
peroxides which gave off ozone when decomposed ; this term was then extended
to certain compounds formed by the action of ozone on various derivatives of
unsaturated organic compounds ; and it is now applied to compounds formed by
direct union with ozone, and which contain the, presumably dyad, radicle O3.
Thus, ethylene CH2 : CH2 forms ozoethylene or ethylene ozonide, C2H4O3, where the
group O3 is thought to act as a dyad, —0 . 0 . 0— , or —0 : 0 : 0— ; and benzene,
CeHgjforms ozohenzene, CgHeOg, where three dyad O3 groups are united to the ben-
zene. The graphic symbols of these two ozonides are considered to be respectively
H
H
C
C__03
H>C=C<H
H>Q-Q<H
H.C^'^C.H
HC/H.H
^^J
C
H.Cls^/C.H
0
H
ri O3
Ethylene.
Ozoethylene or
Benzene.
Ozobenzene or
ethylene ozonide.
benzene ozonide.
The ozonides are usually prepared by slowly passing oxygen containing 3 to 18
per cent, of ozone over the dry substance or into a solution of the substance in an
inert solvent — methyl or ethyl chlorides — free from water. Solvents, like acetone,
chloroform, hexane, and carbon tetrachloride, are more or less attacked by the ozone ;
benzene forms ozobenzene. The ozonides are usually decomposed by water, and
they are often explosive.
Some evidence has been cited to show that a still more condensed form of oxygen,
O4, and called OXOZOne, is present in the residues obtained when liquid ozonized
oxygen — prepared by a brush discharge of high voltage, say 8000 volts — is
fractionally distilled. The alleged oxozone has not been isolated, although com-
pounds called oxozonides containing the group O4 are known ; as well as ozonides
with the group O3. Thus, the hydrocarbon butylene, C4H8, forms both ozo-
butylene, C4H8O3, and oxozobutylene, C4H8O4. Indeed, said C. D. Harries (1911),
" experimental results in ozonization lead to the conclusion that all organic com-
pounds containing an ethylene linkage (double bond) add one molecule of ozone,
and give rise to ozonides.'' Thus, ozone may oxidize in one of two ways : (i) One
atom of oxygen per molecule of ozone is given up to the reducing agent, and the
other two atoms unite to form a molecule of oxygen ; and (ii) the whole molecule
may unite with the reducing agent to form an ozonide.
C. D. Harries (1911) ^ claims that oxozone has an identity of its own on the
following grounds : (1) The specific gravity of the gas from the last fraction of
liquid ozonized oxygen to evaporate, is less than corresponds with the amount of
iodine it liberates from potassium iodide. C. D. Harries takes this to mean that
some O4, as well as O3, is present, and that the former on decomposition furnishes
900 INORGANIC AND THEORETICAL CHEMISTRY
two atoms of oxygen : 04->02+20 ; and the latter, one such atom : 03->02+0.
This inference is not an adequate explanation of the alleged discrepancy because a
mixture of the two should give the same result if determined gravimetrically as if
determined iodometrically. (2) Moderately ozonized oxygen can be passed through
potassium hydroxide solution or through sulphuric acid without appreciable loss of
ozone, but with very concentrated ozonized oxygen or with oxygen ozonized by means
of a high voltage, there is a 3 to 4 per cent, loss of ozone in the potassium hydroxide
solution, and a 2 to 3 per cent, loss of ozone in the sulphuric acid. This may or
may not show that there is an equilibrium condition between the ozone and the
solutions in question. (3) C. D. Harries and his co-workers (1912) have shown
that washed ozonized oxygen in contact with butylene, C4Hg, produces the
ozonides C4H8O3 and the polymer (C4H803)2, while the unwashed gas produces
the oxozonide C4H8O4 and the polymer (C4H804)2. They also found that when
washed ozonized oxygen is passed into a solution of tetrahydrobenzene, CeHio, in
hexane, it gives the solid ozonide CgHigOs, while the unwashed gas gives a mixture
of the ozonide and oxozonide ; and similarly, with caoutchouc, the compounds
CioHigOg and CioHieOg are formed. Hence, argues C. D. Harries, the so-called
ozone in ozonized oxygen with which this work was done contained about one-third
of oxozone. (4) E. Ladenburg and E. Lehmann found some absorption bands in
the red portion of the spectrum of liquid ozone which are only visible in the fractions
remaining when three-quarters has evaporated ; these bands are the first to dis-
appear, and the pressure increases when this occurs. (5) E. Ladenburg and E. Leh-
mann consider that the changes of the pressure, and also of the density of liquid ozone
— 1'78, 1*75, 1"83 — indicate the presence of a higher molecular modification of ozone.
According to E. H. Riesenfeld and F. Bencker, although the reaction between potas-
sium iodide and ozone starts instantaneously, the final equilibrium resulting in the
formation of potassium hydroxide, iodide, hypoiodite, iodate, and periodate, is
attained only after some days. The oxidation nmnber of ozone, that is, the number
of oxygen atoms consumed per gram-molecule of ozone, in a neutral solution of
potassium iodide is unity ; in acid solutions the oxidation number ranges from
1*0 to 2*7 — it is not affected by the concentration of ozone, and increases with
decreasing temperatures ; and in alkaline solutions also the oxidation number
increases. The greater the concentration of the ozone in the oxygen, the
greater the influence of the hydroxyl ions, resulting in the formation of iodate
or hydrogen peroxide. The differences observed by C. D. Harries are to be
attributed to the action of the hydroxyl ions of the alkali and not to the existence
of a modification of oxygen containing more than three atoms. CD. Harries used
a concentrated solution of potassium iodide and concentrated ozone, so that the
oxidation number was increased by the potassium hydroxide which was formed.
In conclusion, E. H. Riesenfeld and F. Bencher say that there is no evidence
of the existence of a modification of ozone containing more than three atoms per
molecule. To this, C. D. Harries replied that the explanation suggested by
E. H. Riesenfeld and F. Bencher is wrong ; a 2*5 per cent, solution of potassium
iodide was used ; and moreover the action of crude and washed ozone on butylene
is evidence of the presence of more than one compound in ordinary ozone.
Beferences.
1 C. 1). Harries and F. Evcrs, Licblg^? Ann., 390. 235, 19J2; C. D. Harries, ib., 343. 311,
lOfHi ; 374. 288, i910 ; Her., 36. 1933, 2997, 3001, 3058, 1903 ; 37. 612, 839, 1904 ; C. D. Harries
and A. de Osa, ih., 37. 842, 1904; C. D. Harries and R. Weil, ib., 37. 845, 1904; Zeil.
EleklrorMm.,YJ. 029, 1911; 18. 129,1912; A. Kailan, ib., 17. 900, 1911; C D. Harries and
W. Frank, Her., 42. 440, 1909 ; C. J). Harries and C. Thiemer, ib., 39. 2844, 1906 ; C. D. Harries
and H. Nereseheimer, ib., 39. 2840, 1900 ; E. E. Molinari and E. Soncinc, i6., 39. 2735, 1900 ;
C. D. Harries, F. Hagedor, and R. Seitz, ib., 45, 930, 1912 ; C. I). Harries, Untersuchungen iiber
das Ozon und wine Einwirkuiuj miforganificke Verbindungcn, Berlin, 1910 ; E. H. Riesenfeld and
F. Bencker, Zeit. anorg. Chcm., 98. 107, 1910 ; C. D. Harries, ib., 99. 195, 1917 ; E. Ladenburg
and E. Lehmann, Her. dent. phys. 6'e.y., 4. 125, 1900.
OZONE AND HYDROGEN PEROXIDE 901
§ 6. The Chemical Properties of Ozone
Ozone slowly and spontaneously passes into ordinary oxygen at ordinary tem-
peratures. L. von Babo i kept a sample for a week over concentrated sulphuric acid
and found some undecomposed ozone. The gas seems to be more stable in contact
with acid than with water. CD. Harries found that ozonized oxygen lost one
per cent, of ozone by bubbling through sulphuric acid. The decomposition of ozone
gives ordinary oxygen, and is attended by an expansion corresponding with 2O3
(2 vols.) =302 (3 vols.). 2 The speed of the spontaneous decomposition of ozone in
ozonized oxygen is greater the more concentrated the ozone. ^
The rate of the decomposition is accelerated by reducing the pressure. Accord-
ing to D. L. Chapman and H. E. Jones, the presence of oxygen, carbon dioxide,
nitrogen, and moisture do not appreciably affect the rate of decomposition, while
the presence of traces of nitrogen peroxide, chlorine, and phosphorus pentoxide
accelerate the rate of decomposition. The effect of moisture has been previously
discussed. Ozone is said to be decomposed by agitation with powdered glass,
by passage through a long glass tube * — though this is doubtful — and by mere
contact with certain agents : 5 finely divided platinum and other metals ; silver
foil; the rare earths; dioxides of manganese, lead, nickel, and cobalt; oxide
of iron, silver, or copper ; mercury ; soda lime ; etc. The ozone is converted
into ordinary oxygen without decomposing the oxides. Hence, the reactions
are grouped among catalytic reactions. The effect can be shown by passing
ozonized air through a tube containing copper oxide and testing the issuing gas by
ozone test paper. No indication of ozone is obtained. The action is probably due
to the cyclic formation and decomposition of the higher oxides ; with silver foil,
the alternate formation and decomposition of silver oxide can be observed. Accord-
ing to D. L. Chapman and H. E. Clarke the effect of the surface of the glass on the
decomposition of the contained ozone is so slow that even in moderately small
globes, the amount of ozone destroyed on the internal surface of the vessel may be'
neglected in comparison with that decomposed in the interior of the gas ; and that
the conversion of ozone into oxygen may be regarded as a homogeneous reaction.
C. F. Schonbein ^ found that if clean and dry plates of gold or platinum be placed
in ozonized oxygen, the metal becomes negatively polarized, and an electric current
can be obtained by connecting up these plates with plates of ordinary gold or plati-
num— the polarization disappears on heating the plates.
The action of heat is very peculiar, as previously discussed. In every case, the
decomposition and formation of ozone by ultraviolet light, electrical discharge, and
heat, a balanced reaction is involved, and the conditions of equilibrium are in accord
with the law of mass action. According to S. Jahn,^ and E. P. Perman and R. H.
Greaves, the rate of decomposition varies inversely as the pressure of the oxygen ;
but D. L. Chapman and H. E. Jones found the velocity of decomposition not to be
affected by increasing the partial pressure of the oxygen mixed with the ozone at
100°, the reaction is almost irreversible, and the rate of decomposition at 127° is
bimolecular, so that if C denotes the number of gram-molecules of ozone per litre,
and ^ is a constant, t the time, the velocity of the decomposition is (lCjdt=kC^;
if C is unity, k represents the number of gram-molecules of ozone which would
decompose in one minute if the initial concentration of the ozone were one gram
per litre. E. Warburg found at 16°, yt=0-0000492 ; at 100°, A;=0-157 ; and at
1269°, A;=l-77. The water had a pressure of 0-0021 mm. of mercury ; when the
water pressure was 0*154 mm., the velocity of decomposition was 22 per cent,
faster than the results just recorded. In order to explain this result, it has been
assumed that a rapid reversible change 03^02+0 occurs, and this is followed
by a relatively slow reaction 03-[-0=202. This latter reaction is alone accessible
to measurement.
The decomposition of ozone at high temperatures is accompanied by phos-
nhorescence or luminGSCfincft — this is shown bv holflin<T a hot dass rod near the
902 INORGANIC AND THEORETICAL CHEMISTRY
surface of liquid ozone (M. Beger, 1910), or by passing ozonized oxygen through a tube
with a e^ipillarv opening into an evacuated vessel (J. Dewar, 1888) .» When ozone is
aspirated with a water pump, the water becomes luminous, and retains its luminosity
for five or six seconds. Flasks filled with the issuing water are distinctly luminous,
and when water is shaken with ozonized oxygen, the phenomenon of luminescence
can be reproduced five or six times and then ceases ; it can be reproduced with the
same gas if fresh water be introduced. Variations of pressure had no appreciable
effect on the luminosity. If ozonized oxygen be shaken with alcohol, the lumi-
nescence is feebler biit more persistent ; with benzene it is very feeble, and the ozone
is absorbed or decomposed. If the water be carefully freed from organic matter
it shows no luminescence even with highly concentrated ozone. The luminescence
with ordinary tap water is therefore due to the action of organic matter on the
ozone. According to M. Otto, the luminescence always appears on contact of
ozone with organic bodies. The decomposition of ozone gives gaseous ions which
discharge a charged electroscope. According to R. Schenck,^ ozone acts on a photo-
graphic plate, and makes zinc blende, but not zinc oxide or barium platinocyanide
fluorescent. These statements have also been denied ; others say the fluorescence
of zinc blende is an effect which attends its oxidation to sulphate, and the action on
a photographic plate is due to the chemical action of ozone on the organic matter.
Ozone is a very powerful oxidizing agent. B. C. Brodie i^ (1872) has made
an interesting study of its action with different substances. Let the so-called
oxidation number represent the number of atoms of oxygen [0] taken up by the
oxidized compounds, per molecule of ozone, then
(1) If the ozone is catalytically decomposed, say by concentrated alkali lye at not too
low a temperature, 203 = 302, the oxidation number is zero, and the increase in volume is
equal to half the volume of the ozone employed.
(2) If the ozone gives up one atom of oxygen per molecule, as is the case when it acts
on ferrous sulphate, manganese sulphate, potassium ferrocyanide, or potassium iodide,
03 = 02 + [0], and there is no change of volume. The oxidation number is unity.
(3) If the ozone gives up all three oxygen atoms to the oxidized substance as is the case
with turpentine, sodium thiosulphate in neutral solution, the oxidation of sulphur dioxide :
380,4-03 = 3863, sodium sulphide : Na£S4-03=Na2S03, and in the formation of ozonides,
03=[30J ; the oxidation number is 3, and the decrease in volume is equal to the volume
of the ozone.
According to E. H. Riesenfeld and T. F. Egidius (1913), n it is probable that the
ozone does its work in one of two ways : (i) One atom of oxygen per molecule of
ozone is given up to the reducing agent, and the remaining two atoms form a molecule
of gaseous oxygen ; or (ii) the whole molecule may be added or coupled with the
compound oxidized, as in the formation of ozonides, and the oxidation number
is 3. The more complicated reactions are probably the result of a combination of
these two. Thus, the reaction 203=02-f[40] observed by B. C. Brodie in the oxida-
tion of sodium thiosulphate, is a combination of 03=02+ [0] and 03=[30]. In
the oxidation of sodium thiosulphate in neutral solution, Y. Yamauchi found that
the main reaction is symbolized Na2S203-f 03=S02-f Na2S04, with a side reaction
Na2S203-f03=Na2S206 (sodium dithionate). In alkaline solutions, sodium thio-
sulphate seems to unite directly with ozone, forming an ozonide, Na2S203.03 which
subsequently decomposes, giving off ordinary oxygen, forming as principal products :
sodium sulphate, sulphite, and dithionate— with sulphate as the end-product.
Similar remarks apply to the oxidation of sulphides and sulphites. The result
may be complicated by the catalytic decomposition of the ozone or of the thiosul-
phate (Na2S203=S+Na2S03) itself.
Ozone oxidizes hydrogen very slowly. Although C. F. Schonbein and M. Ber-
thelot failed to obtain any evidence of combination, others i- have found that by
exposing electrolytic gas to the brush discharge^ — either under reduced pressure, or
by diluting the gas with argon or carbon dioxide so as to avoid an explosion, or by
heating the gaseous mixture— water is formed. Thus, F. Fischer and M. Wolf, by
OZONE AND HYDKOGEN PEKOXIDE 903
heating electrolytic gas containing ozone, for an hour, at 100°, obtained 3*7 mgrm.
of water and 4'4:5 mgrm. at 174°. According to F. Weigert, when electrolytic gas
is exposed to ultraviolet light, larger quantities of water are formed than corresponds
with the ozone produced. H. Thiele says that the action is not to be ascribed to the
primarily formed ozone, as in the case of other reactions in ultraviolet light.
C. F. Schonbein, G. Meissner, and L. von Babo i^ supposed that when a solution
of ozone in water decomposes, hydrogen peroxide, H2O2, is formed, but many investi-
gators— ^C. Engler and A. Nasse, L. Carius, E. Schone, C. Gianetti and A. Volta,
and M. Berthelot — have denied this reaction, and claim that the solution decomposes
into oxygen and water. There is no reliable evidence of the alleged formation of
hydrogen peroxide. The difficulty arose because of the confusion of ozone reactions
with those of hydrogen peroxide, and C. Arnold and C. Mentzel, V. Kothmund and
A. Burgstaller, and P. Jannasch and W. Gottschalk, have employed tests which
leave no ambiguity. The latter have shown that neither hydrogen peroxide nor
persulphuric acid is formed in solutions of ozone in sulphuric acid. Ozone precipi-
tates manganese dioxide quantitatively from a solution of manganous sulphate,
but if a trace of hydrogen peroxide were formed the reaction would not be quantita-
tive since the precipitate is dissolved in the presence of that reagent. The stability
of a solution of ozone decreases as the concentration of the acid decreases, but no
relation between the concentration of the acid and the velocity of decomposition
has been discovered. Alkaline solutions of ozone are very unstable. Colloidal
platinum, copper sulphate, stannous sulphate, etc., do not accelerate the speed
of decomposition of aqueous solutions of ozone. V. Rothmund also found that in
0"01iV-acid solutions the speed of decomposition of ozone is bimolecular, and in
weaker acid and in alkaline solutions the speed is between a bi- and uni-molecular
process ; they therefore assume that a uni-molecular reaction is superposed on a
bimolecular one, and this furnishes the velocity equation dxl(lt=ki(a—x)'^ -\-k<2{a—x) ,
where k^ and k^ are constants — in acid solutions the first term is the more important,
and in alkaline and weakly acid solutions the second term is the more important.
C. D. Harries found almost all ozonides react with water forming hydrogen
peroxide. M. Berthelot and C. T. Kingzett also found ozonized ether, and many
other substances after treatment with ozone, give hydrogen peroxide when shaken
with water. The hydrogen peroxide may be formed by the autoxidation of these
substances, and the ozone may act indirectly as a carrier of oxygen. C. F. Schonbein,
L. von Babo, E. Schone, A. Schmidt, etc., have reported that commercial ether
nearly always contains hydrogen peroxide, and this may mean that an ozonide has
been formed. Ozone and hydrogen peroxide react slowly : H202+03=H20+202.
H. J. K. Inglis found the reaction is catalytically accelerated by manganese salts.
So slow is this reaction in aqueous solutions that C. Arnold and C. Mentzel believed
the mixture to be Bestdndigkeit, and C. Engler and W. Wild 1* applied the same
remarks to the gaseous mixture. It is very remarkable that the reaction between
a vigorous oxidizing agent like ozone, and a reducing agent like hydrogen
peroxide, should be so slow. Theoretically it might have been anticipated that
0:0: iO-f-H2i : 0 : 0->H20+202 would readily occur. In concentrated solutions
this is probably what happens, because, in F. Fischer and M. Wolf's experiments on
the synthesis of hydrogen peroxide by the action of the brush discharge on mixtures
of hydrogen and oxygen at low temperatures, it was found that a mixture with 97
per cent, of oxygen and 3 per cent, of hydrogen gave much ozone and the merest
traces of hydrogen peroxide, while if these proportions were reversed, a highly
concentrated form of hydrogen peroxide was obtained. It is assumed that the trace
of hydrogen peroxide, in the former case, represents what has escaped decomposi-
tion in the reaction between ozone and hydrogen peroxide. V. Rothmund and
A. Burgstaller measured the velocity of the reaction between hydrogen peroxide
and ozone in 0"01iV-sulphuric acid at 0°, and found that in the presence of a large
proportion of hydrogen peroxide the decomposition of the ozone follows the course
of a unimolecular reaction: H202-f 03=11204-202; but if a less proportion of
904 INORGANIC AND THEORETICAL CHEMISTRY
the peroxide be present, the ozone disappears more rapidly than the peroxide, and
this the more the lower the ratio of peroxide to ozone. Hence it was inferred
that the interaction of hydrogen peroxide and ozone is accompanied by the
spontaneous decomposition of ozone, and that hydrogen peroxide catalyzes the
latter reaction. According to H. McLeod (1880), if the hydrogen peroxide be in an
alkaline or neutral solution, the ozone is destroyed very quickly ; the reaction is slow
only in acid solutions. In some cases, therefore, ozone acts as a reducing agent.
Thus, barium peroxide is reduced to the monoxide, and hydrogen peroxide to water.
When ozone is brought into contact with sodium peroxide, the two substances
mutually decompose and oxygen is liberated : 03+Na202+H20=2NaOH+202.
There are very uncertain signs of the formation of a very unstable compound
of ozone with fluorine.^^ H. Moissan postulated the possible formation of an inter-
mediate compound of fluorine and ozone when fluorine acts on water ; but 0. Ruff
and J. Zedner could detect no signs of such a compound by the action of fluorine
on oxygen under the influence of an electric arc, or in the induction coil. G. Gallo,
however, believed that he did prepare an unstable endothermal compound by the
action of ozone on fluorine at low temperatures, and that above — 23° this com-
pound is liable to decompose explosively. E. Comanducci claims to have converted
a mixture of chlorine and oxygen into chlorine dioxide by treating the mixed gases
in Siemens' ozonizer.i^ D. L. Chapman and P. S. McMahon found that the presence
of ozone retards the photochemical action between chlorine and hydrogen. No
reaction was observed with bromine and oxygen. C. F. Schonbein found that iodine
is oxidized in a stream of ozonized oxygen to I2O3 or I4O9, and according to H. B.
Baker and R. J. Strutt, the oxidation ^^ is accompanied by an orange-coloured glow.
By exposing a mixture of iodine vapour and oxygen to the brush discharge in an
ozonizer, J. Ogier obtained different oxidization products in different parts of the
tube. F. Fichter and F. Rohner passed 8 per cent, of ozonized oxygen into a chloro-
form solution of iodine, and obtained a yellowish- white oxide, I4O9, and the same pro-
duct was obtained by the action of ozone on dry iodine. By treating a solution of
iodine in acetic anhydride and concentrated sulphuric acid with ozonized oxygen,
M. Beger obtained a white substance which unites with sulphuric acid forming a
citron-yellow hygroscopic powder, I2O3SO3JH2O, analogous to the substance
prepared by P. Chretien by the action of iodine on a solution of iodic acid in sulpluiric
acid. Ozone decomposes all the hydrogen haloids with the exception of hydrogen
fluoride: 2HCl-f03=Cl2+H20-|-O2; and E. Comanducci claims to have made
hypochlorous acid by the action of ozone on hydrogen chloride in an ozonizer.
The action of ozone on potassium iodide. — Unlike oxygen, ozone liberates
iodine from a neutral solution of potassium iodide, KI. This can be shown by dipping
paper in a solution of potassium iodide and holding it at the exit tube of the ozonizer.
The paper turns brown owing to the liberation of iodine. If a little starch be mixed
with the potassium iodide, the paper will appear blue if ozone be present. The
reaction is usually represented 03+2KI+H20=02+l2H-2KOH. The solution
is then alkaline in virtue of the potassium hydroxide. Hence, if red litmus paper
be steeped in water containing a trace of potassium iodide, the moist paper, when
exposed to ozonized air, will be coloured blue owing to the action of the potassium
hydroxide on the red litnkus. The simple equation just indicated gives no idea of
the great complexity of this reaction. According to 0. Brunck i^ and others, there
is a slow reaction between the iodine and the potassium hydroxide which results in
the formation of potassium hypoiodite, KIO, which, in consequence of another
consecutive reaction, slowly forms potassium iodate, KIO3, and iodide, 3KIO=2KI
-I-KIO3. E. Pechard believes that potassium periodate is formed at an intermediate
stage in the oxidation of potassium iodide by ozone, and that the periodate then
reacts, 3KIO4+2KI+3H2O-KIO3+2K2H3IO6+I0, and the two constituents last
symbolized react 2K2H3l06+l2=3KI03+KI+3H20, which form a neutral solu-
tion. Some periodate and peroxide have also been reported to be produced by a
reaction between the iodide and ozone. According to J. N. Pring, the ratio of iodate
OZONE AND HYDROGEN PEROXIDE 905
and periodate to free iodide and hypoiodite increases with the amount of ozone which
acts on the solution. AVith a dilution less than 160 parts of ozone per million of air,
no iodate is formed, but only iodine and hypoiodite. If the temperature is less than
— 24°, and the ozone acts on the solid salt, the smallest quantity ot ozone gives
more iodate than free iodine and hypoiodite. If ozone acts upon an acid solution
of potassium iodide, the result is different i^ from that which occurs with a neutral
solution, for there is a side reaction resulting in the formation of hydriodic acid :
KI+HC1=HI+KC1, in addition to the normal reaction for neutral solutions.
The ozone reacts with the hydriodic acid, HI, forming hydrogen peroxide, possibly
by the side reaction : 4HI-f 63=H20 +H2O2+2I2. The resultant equation is repre-
sented : 4O3-fl0HI=H2O2+5l2+4H2O+3O2. The hydrogen peroxide can be
detected with titanium sulphate which gives a yellow coloration. It therefore follows
that a gram-molecule of ozone gives more iodine in an acid than in a neutral solution.
According to G. Lechner (1911), an alkaline solution of potassium iodide reacts
with ozone forming potassium iodate, KIO3, thus KI-f03=KI03. If the mixed
solution of iodine and iodate be acidified, iodine is liberated : KlOs-f 5KI+3H2SO4
=3K2S04+3H20+3l2. The free iodine can then be determined by titration with
a standard solution of sodium thiosulphate. In this way, E. Czako (1912) has
shown that it is possible to determine 0*00002 grm. of ozone in 100 c.c. of ozonized
air. According to C. Engler and A. Nasse, dry ozone and dry potassium iodidev do
not react. H. Riesenfeld and F. Bencker found that ozone has no action on neutral
or acid solutions of potassium iodate, but it oxidizes alkaline solutions to periodate ;
and it is without action on potassium periodate — vide oxozone.
The white mist — ozone fog — produced when ozone acts on iodides, sulphur dioxide, etc.,
was thought by C. Meissner to be the antozone of C. F. Schonbein. V. Rothmund ^o noted
that analogous fogs are produced in many other reactions- — e.g. ammonium chloride fogs,
fogs from fuming acids, fogs produced by the action of radium emanation on sulphur,
carbon disulphide, etc. — and that it is caused by water in which a small quantity of the
products of the reaction are dissolved. The fogs with ozone are produced only when the
reducing agent is of a volatile nature and the reaction products are soluble in water. The
size of the mist spheres were foimd to be practically the same in a number of very different
reactions, and from their rate of subsidence they are approximately 10~* cm. in diameter.
The reports which have been published to the effect that SUlphUT is not attacked
by ozone are probably based on experiments in which the concentration of the ozone
was very small. E. Pollacci^i has shown that at ordinary temperatures sulphur is
oxidized by ozone, but not by oxygen ; and H. B. Baker and R. J. Strutt observed
that a blue luminescence is obtained when ozonized oxygen is passed over sulphur.
A. Malfert supposed that sulphur dioxide is formed if moisture be excluded ; sul-
phuric acid if moisture be present ; and alkali sulphates if alkalies be present.
T. Weyl obtained sulphuric acid by leading ozonized oxygen into hot water
with finely divided sulphur in suspension, but there were no signs of oxidation
if oxygen alone be used. A. Stock and K. Friederici also found that sulphur dis-
solved in carbon tetrachloride was oxidized by ozone to sulphur dioxide. C. D.
Harries looks upon sulphur trioxirle as a kind of ozonide, SO3. Hydrogen sulphide
is oxidized by ozone to sulphur, and D. Helbig found that hydrogen sulphide can be
oxidized to sulphur when passed into a solution of potassium permanganate and
sulphuric acid, which furnishes ozonized ox^^gen. Sulphur dioxide and sulphurous
acid are oxidized to sulphur trioxide and sulphuric acid respectively : 3SO2+O3
=3S03. P. V. Langlois and L. S. Thomassin, and E. H. Riesenfeld have suggested
technical processes for the oxidation of sulphur dioxide based on this reaction.
With the same object in view% A. Reynoso and B. Hunt proposed to pass electric
sparks through a strongly compressed mixture of sulphur dioxide and air at a low
temperature. A. Borchers assumes that an ozonide, (802)303, is first formed as an
intermediate product which decomposes into 3SO3. The action of ozone on sodium
sulphide, hydrosulphide, or polysulphide ; sodium thiosulphate, sodium sulphate,
and sodium bisulphite, gives sodium sulphate as an end-product when an excess of
906 INORGANIC AND THEORETICAL CHEMISTRY
the oxidizing agent is used ; when the ozone is not in excess, intermediate products .
are formed. With sodium sulphide, E. H. Riesenfeld and T. F. Egidius detected
the intermediate products :
-y Hyposulphite v
Sulphide — > Thiosulphate ^ Sulphite \ > Dithionate — > Sulphate
^ Polythionate /
Ozone is dissolved by sulphuric acid ; and some unknown change occurs, for
J. Brand and L. Grafenberg found that the sulphuric acid possesses oxidizing
properties — e.g. it blues starch and potassium iodide test paper. These qualities
are not lost by boiling, or by leading air through the liquid. The titanic acid
reaction for hydrogen peroxide does not occur. The presence of platinum sponge
favours the oxidation.
The solution gives a similar precipitate with acetone as does Caro's acid, but,
according to J. Brand and L. Grafenberg, neither persulphuric nor Caro's acid
appears to be present.22 The electrical conductivity of the dilute acid seems to be
lowered by the presence of the unknown oxidizing substance which is present, and
which some have supposed to be an unknown ozonic acid. Persulphuric acid has
not yet been formed by oxidizing sulphuric acid with ozone. M. Berth elot obtained
a persulphuric anhydride by subjecting a mixture of sulphur dioxide and oxygen
to the brush discharge, and J. Schmidlin and P. Massini obtained persulphuric
acid by treating the product of the action of ozone on sulphur trioxide or fuming
sulphuric acid and with water. The presence of sulphur trioxide seems to be
necessary, since no persulphuric acid is obtained with sulphuric acid alone. It is
thought that the reaction involves the formation of sulphur heptoxide as an inter-
mediate product. The salts of persulphuric acid with concentrated sulphuric acid
give ozonized oxygen. Selenium and tellurium, like sulphur, are oxidized by ozone.
Attempts by J. Jannek and J. Meyer to make pure selenium dioxide and selenium
trioxide by this method have not yet been successful.
It has not been definitely proved that ozone can react directly with nitrogen.
L. Carius andM. Berthelot 23saidnot. C. F. Schonbein stated that a mixture of ozonized
oxygen and nitrogen furnishes calcium nitrate with lime water, but M. Berthelot
says that the nitrate was due to an impurity in the lime water, and that nitrates or
nitrogen oxide may be produced as a by-product in the formation of ozone by the
phosphorus process. However, there is no doubt that, as shown by J. Chappius
and P. Hautefeuille, nitrogen is oxidized when air is exposed to the brush discharge
in the ordinary method of preparing ozone ; 24 but, as E. Fonrobert remarks, this
does not prove that ozone can react with nitrogen. C. Montanari obtained no signs
of the oxidation of nitrogen by leading a mixture of nitrogen and ozonized oxygen
over platinized lime. The chemically active variety of nitrogen obtained by R. J.
Strutt 25 by the action of a spark discharge on nitrogen gives no trace of nitrogen
oxide after it has been mixed with ozone, and cooled with liquid air, but T. M.
Lowry has shown that air which has been subjected to the brush discharge and
afterwards sparked gives a greater yield of nitrogen oxides than when the air supply
has not been previously ozonized. Hence, it is inferred that in the sparking the
nitrogen in the air is converted into a form which enables it to unite rapidly with
the ozone. Nitric oxide once formed can be oxidized to a higher oxide, which
is decomposed by ozone to oxygen and nitric oxide. 26 While two volumes of nitric
oxide and one volume of oxygen unite to form nitrogen peroxide, the reaction is
not complete if the oxygen employed be previously ozonized — this may be due to
the direct retarding action of ozone, or to the destruction by ozone of some catalytic
agent necessary for the reaction between nitric oxide and oxygen. Ozone oxidizes
nitrogen trioxide to the tetroxide, and nitrogen tetroxide to the pentoxide ; in
aqueous solution, nitrous acid is first formed and then nitric acid. The reaction is
very much faster with ozone than with oxygen.27 D. Helbig obtained a volatile sub-
stance by the action of ozone on nitrogen trioxide at the temperature of liquid air ;
OZONE AND HYDROGEN PEEOXIDE 907
it is possible an ozonide, N2O6, identical with E. Miiller's nitrogen hexoxide. The
oxidation of the tetroxide to the pentoxide by ozone has been recommended as a
method of making nitric anhydride. 28
L. Carius, and L. I. de N. Ilosva failed to obtain evidence of any combination be-
tween dry ammonia and ozone ; but G. Baumert and C. F. Schonbein 29 noted the for-
mation of a cloud immediately moisture was added, and the ammonia was oxidized to
ammonium nitrite and nitrate. L. Carius reported that some hydrogen peroxide
was formed at the same time : 2NH3+403=NH4N02+H202+402 ; and NH4NO2
+H2O2— NH4NO34-H2O. If the gases are very dilute, the reaction is so slow
that it is not thought the occurrence of ammonium nitrite and nitrate in the atmo-
sphere can be explained by the oxidation of ammoniacal products by atmospheric
ozone. D. Helbig reported the oxidation of ammonia with incandescence when the
gas is passed over a mixture of potassium permanganate and sulphuric acid ; while
P. Jannasch and W. Gottschalk found a cloud is formed by leading ozone into
aqueous ammonia, and the cloud condenses to a white film on the walls of the vessel.
E. Warburg also obtained a very marked contraction by exposing a mixture of
ammonia and oxygen to ultraviolet rays. A. W. Browne and F. F. S. Netterly ^o
reported that if ozone be passed into a boiling alkaline solution of hydrazine sulphate,
a small quantity of azoimide, HN3, but no ammonia, NH3, was formed.
Dry yellow and red phosphorus are oxidized to phosphorus pentoxide by ozone ;
if water is present phosphorous acid is first formed and afterwards phosphoric acid.
Ozone oxidizes arsenic in presence of water to arsenic acid, but antimony is not so
easily oxidized, and C. F. Schonbein (1847) proposed to distinguish the two latter
elements by this reaction. J. Schmidlin and P. Massini^i tried to convert phosphorus
pentoxide into a higher state of oxidation by means of ozone, but without result.
Arsenious oxide is oxidized quantitatively to arsenic acid by ozone, and C. F. Schon-
bein^- proposed to determine ozone by the reaction: As203+203=Aso05+202,
but Y. Yamauchi found that one gram of ozone oxidized 2 '14 grms. of arsenic trioxide
while theoretically 2*06 grms. passes into the pentoxide. Phosphine and arsine are
immediately decomposed by ozone ; and A. Stock and W. Siebert found that
stibine at —90° explodes in contact with ozone. A solution of ozone in liquid
oxygen does not attack solid stibine, but as the temperature is allowed to rise
slowly an explosion occurs. According to A. Besson,^^ ozone in sunlight acts on phos-
phorus trichloride forming the oxy chloride ; with phosphorus tribromide> the
pentabromide and trioxide are formed, but no oxybromide ; and with arsenic
trichloride, at 50°, chlorine is liberated and arsenic oxide formed. Phosphorous
iodide is decomposed with the liberation of iodine and the formation of complex
oxyiodide. Ozonized oxygen has no action in the cold on solutions of phosphorus
pentachloride or pentabromide in carbon tetrachloride. A. Stock and K. Friederici 34
find that when solutions of phosphorus trisulphide, P4S3, are treated with ozone a
yellowish- white oxysulphide, P4S3O4, which with further treatment forms P4S3O7, is
precipitated.
Ozone is destroyed by carbon, the smell disappears, but no oxidation has been
detected. A. R. Leeds and F. Baumann stated that ozone oxidizes carbon monoxide,
while according to I, Remsen and M. S. Southworth, and E. H. Keiser,35 ozone does
not oxidize this gas, but the negative results are possibly due to the use of ozone of
too great dilution. M. Berthelot has shown that there is a kind of equilibrium
between ozone and carbon monoxide since a mixture of the two gases in the brush
discharge forms about 90 per cent, carbon dioxide, while carbon dioxide is partially
decomposed under similar conditions into carbon monoxide and ozone. E. Gold-
stein found that when sparks were passed through a Geissler's tube filled with
oxygen and carbon monoxide, at the temperature of liquid air, the spectrum of
carbon monoxide gradually disappeared. R. Clausmann exposed a mixture of
ozonized oxygen and carbon monoxide in sunlight for eight days and found 2*83
parts of carbon dioxide were formed, in darkness only 0"88 part was formed.
J. Thiele also exposed a similar mixture to ultraviolet radiations and obtained small
908 INORGANIC AND THEORETICAL CHEMISTRY
quantities of carbon dioxide. According to W. A. Jones and C. E. Waters,36
the oxidizing power of ozonized oxygen on carbon monoxide is greater the higher
the temperature, and the greater the concentration of the ozone. The action
is appreciable in the cold if the concentration of the ozone be high. A. Besson
and L. Fournier^^ converted silicochloroform into a volatile oxychloride, Si2Cl60,
by the action of ozone.
With the exception of gold and the metals of the platinum family, moist ozone
oxidizes all the metals which have been tried — copper, iron, nickel, etc. In some
cases the metals require heating at ordinary temperatures. Copper, nickel, and tin
withstand the gas fairly well ; aluminium, zinc, brass, and lead are quickly corroded.
Iron is not so readily oxidized if the carbon-free metal be alloyed with chromium,
and the use of ferrochromium containing 25 per cent, of chromium hasbeen patented 38
for chemical apparatus required to resist the fumes of ozone or nitric oxide. A
coating of shellac varnish as well as many of the acid-proof paints of the graphite or
asphalt type protects the metals quite well. Silver and lead form the higher oxides.
If the gas and metal be thoroughly dried, C. F. Schonbein ^^ found that in many
cases no reaction occurs ; the presence of moisture is necessary to start the oxida-
tion. C. F. Schonbein found that similar remarks apply to the action of ozone on
several other substances — metal sulphides, iodides, manganese and lead salts,
potassium cyanide, organic colouring agents, etc. W. Manchot and W. Kamp-
schulte hold that dry ozone can react with the metals forming ozonides under certain
conditions. These ozonides are usually very unstable and only a very small quantity
is formed.' If, however, water be present, or the system be heated, the ozonide is
formed in greater proportions and decomposed into the metal oxide and oxygen.
The many analogies between ozone and sulphur suggest the hypothesis that
ozone and sulphur dioxide, SO2, are related as indicated in the respective graphic
formuljB 0=0=0, and 0=S=0 ; and that just as sulphur dioxide is the anhydride
of sulphurous acid, so is ozone the anhydride of an ozonic or ozonous acid. Although
ozonic acid is unknown, it is supposed that the well-known potassium tetroxide,
K2O4 — that is, K2O.O3 — is the corresponding salt analogous with potassium sulphite,
say K2O.SO2. The argument, though very feeble, has been pushed still further.
C. F. Schonbein (1844) found that when ozone is passed through an aqueous solution
of, say, 40 per cent, potassium hydroxide, something is formed which gives a blue
coloration with a mixture of potassium iodide and starch ; and A. von Baeyer and
V. Villiger (1902) *^ showed that if the potash lye be strongly cooled, the ozone forms
an orange-brown solution, and the colour disappears when the lye is removed
from the freezing mixture. In these experiments, it was assumed that ozonsdures
Kalium — potassium OZOnate — is formed, and that the salt is more stable at low
temperatures than it is at ordinary temperatures. A. von Baeyer and V. Villiger
also obtained a coloured product with solid potassium hydroxide. Similar results
were obtained with the other four alkali hydroxides, and the stability of the
product decreased with decreasing atomic weight in passing from ccosium to lithium.
Indications of the formation of an analogous unstable product were obtained with
dry liquid ammonia. 41 Again, according to L. Griifenberg, the hydroxides of the
alkaline earths form coloured peroxidized compounds ; wihen ozone is passed into
lime water under the same conditions, a granular precipitate is formed which does
not colour a mixture of potassium iodide and starch blue ; but does so if it be acidified.
The acidified liquid does not smell of ozone. The corresponding product with
magnesium hydroxide is not coloured.
It is further assumed, without proof, that potassium tetroxide and potassium
ozonate are the same, and hence, ozonic acid, H2O4, is regarded as a hydrate o£
ozone, O3H2O, or (H0)202, i.e. (H0)2=0=0 by analogy with the corresponding
sulphur compound, (H0)2=S=0. Against this assumption it has been urged : (1)
Ozonized oxygen led through water does not make the liquid appreciably more
conducting, electrically, which it probably would do if traces of an acid were formed ;
and (2) no appreciable difference has been ^letected in the solubility of ozone in water
OZONP: and hydrogen peroxide 909
and in normal acids— if an acid were formed in water, the solubility in that
menstruum would probably be the greater. The latter argument has not much
weight.
W. A. Shenstone and J. T. Cundall found that although dried ozone is destroyed
by mercury, the metal is not attacked as it is by imperfectly dried ozone. The
action of ozone on mercury is superficial.*^ Put a globule of mercury in a small
flask, pass ozonized air into the flask and shake the globule of mercury about. The
mercury loses its lustre, and spreads a film over the walls of the flask. The globule
of mercury is restored when the film is shaken up with water. According to
W. Manchot and W. Kampschulte, the mercury is but slightly attacked at ordinary
temperatures by oxygen with 1*5 per cent, of ozone by volume ; at 55°, a brown film
is formed which becomes deeper and deeper in colour as the temperature is raised,
and attains a maximum at 180° to 190°. The surface of the mercury then acquires
a deep steel-blue colour and begins to develop a brown vapour. At higher tempera-
tures the action becomes less marked ; at 238°, only a yellow film is perceptible,
and at 250° there is no apparent action. There appears to be a higher oxide, or an
ozonide of mercury formed in this reaction, but the product has not been isolated.
Polished silver foil is attacked by moist ozone, and, according to A. Volta,*^ negatively
polarized ; dry ozone also polarizes the metal, and the ozone is at the same time
decomposed. If a piece of silver foil cleaned with silver sand be heated in a Bunsen's
burner for a moment, and while still warm, held in a stream of ozonized air, the
silver is browned or blackened, owing, it is said, to the formation of a higher oxide
of silver. E. Fremy recommended the reaction as a test for ozone, but A. Houzeau
considered it not sufficiently delicate. Ozone can be recognized by its smell long
before the silver is coloured. If the silver is dirtied by contact with the fingers
the test is not so good. If the surface of the silver be not specially purified it will
be darkened by ozone at the ordinary temperature. Thus, if a plate which has once
been blackened be strongly heated, the colour disappears, but the plate is blackened
by ozone at ordinary temperatures, owing, it is supposed, to the presence of a trace
of undecomposed oxide ; similarly, a plate which has been dipped in dilute nitric
acid, washed, and dried, is blackened at ordinary temperatures. Scouring the plate
with sand destroys its activity at ordinary temperatures. Minute traces of the
oxides of the heavy metals — e.g. nickel, cobalt, lead, chromium or iron oxide — or
the platinum metals— e.^. ruthenium, palladium, or platinum — act as catalytic
oxygen carriers and enable the silver to be blackened by — say, 0'2 per cent, of
ozone— at ordinary temperatures. To clean the metal surface, it is washed a few
times with benzene, polished with moist sand, and rubbed with a clean dry cloth.
The sensitiveness gradually diminishes when the plate is kept. Ozone produces a
white iridescent film on a clean silver plate at 100° ; a steel blue film at 154° ; a
pronounced blackening between 220° and 240° ; this is an optimum temperature
since at higher temperatures the effect becomes more and more feeble, until at 450°
no change can be observed — vide silver.
Copper is attacked by ozone, but reactions analogous with those presented by
mercury have not been observed because air itself oxidizes the heated metal. As in
the case of mercury, W. Manchot found that molten tin at 500° is attacked by ozonized
air (1 per cent, of ozone). Ozonized air slowly attacks metallic lead forming a layer
of brown dioxide ; in the presence of water, A. R. Leeds found that lead hydroxide
is formed, W. Manchot observed only a slight reaction on lead at 100°, with rising
temperatures the attack is more marked ; at 227°, the metal is steel-blue like silver ;
at 385° it is brownish-blue and the action cannot then be distinguished from that
of ordinary oxygen. Nickel is not perceptibly attacked by ozone at 240° ; at 300°
a yellow film is formed ; at 415° the film is golden yellow. This change occurs more
quickly with ozonized oxygen than it does with ordinary oxygen ; at higher tempera-
tures, the difference is inappreciable. The action of ozone on silver sulphide is
very slow ; cobalt and nickel sulphides first form sulphites and then the dioxides ;
lead, manganese, and palladium sulphides give the dioxides and sulphuric acid
910 INORGANIC AND THEORETICAL CHEMISTRY
without the intermediate formation of the sulphite being perceptible ; and gold
sulphide gives metallic gold.
Ozone oxidizes mercurous salts to the mercuric state : ** 2HgN03+03=:HgO
4-Hg(N03)24-02 ; mercurous chloride or bromide also forms some oxyhalide.
Similarly thallous salts are oxidized to thallic salts ; and at the same time the colour
turns brown ; this was proposed as a test for ozone by R, Bottger (1865) and
E. Schone (1879). According to Y. Yamauchi, the reaction, 2T10H+203=Tl203
-f-H20-f202, is quantitative. Ozone precipitates dark-brown lead dioxide from
solutions of lead salts, slowly in neutral solutions, rapidly in basic solutions ; hence,
basic lead acetate test papers have been recommended as a test for ozone. Ozone
converts dark-brown lead sulphide into white lead sulphate : PbS+403=PbS04
-I-4O2. This can be shown by holding a strip of paper which has been steeped in
a solution of lead acetate and browned by hydrogen sulphide, in a stream of ozonized
oxygen or ozonized air. Many other sulphides — copper, antimony, zinc, cadmium —
behave in a similar manner. Attempts by L. Moser to oxidize cupric salt solutions
by the action of ozone on neutral or alkaline, hot or cold, solutions have given no
result. Ozone acts on Fehling's solution like chlorine or hypochlorites. Feebly
acid solutions of bismuth nitrate gave P. Janhasch and W. Gottschalk no precipitate,
but alkaline solutions are coloured yellow or brown. C. F. Schonbein and A. W.
Willianason represented the oxidation of stannous chloride to stannic chloride by
the equation: 3SnCl2+6HCl-|-303=3SnCl4+3H20+302 ; but Y. Yamauchi has
shown that the relation between tJie ozone and the stannous chloride oxidized
corresponds with 3SnCl2+6HCl+03=-3SnCl4+3H20. This latter equation is
unusual in that the whole of the oxygen of ozone is consumed by the oxidation.
The action of ozone on manganous salts is peculiar, and is largely determined by
the concentration and acidity of the solution. With neutral solutions, a brown
precipitate of manganese dioxide is formed ; paper soaked in a solution of man-
ganous sulphate and dried was recommended by D. Huizinga as a test for ozone ; and
paper similarly made with manganous chloride was recommended by C. Engler and
W. Wild. The test papers are not very sensitive. A solution of manganous
sulphate can be used in place of ink, and the writing is turned brown by exposure
to ozone owing to the formation of a brown manganese dioxide. C. F. Schonbein
wrote M. Faraday a letter to this effect, soon after he had discovered ozone. Accord-
ing to L. Maquenne, in feebly acid solutions of manganous sulphate, the solution which
remains after the separation of the manganese dioxide has a rose-pink colour, and
this is the more intense, the greater the acidity of the solution, until, with 10 per
cent, of sulphuric acid, no dioxide is precipitated, and the manganese is converted
into permanganate. Ozone does not carry the oxidation further. If the concen-
tration of the acid exceeds 30 per cent., manganic sulphate, not a permanganate, is
formed. The greater the concentration of the manganous sulphate, the more the
sulphuric acid required. Analogous results are obtained with manganous nitrate
and nitric acid, and with hydrochloric acid and manganous chloride. The fact
that the higher manganese oxides give ozone, led 0. Brunck, as we have seen, to
postulate an ozonide structure for these compounds, and the conversion of the lower
mangansese salts into the higher oxides, by ozone, is taken to confirm this.
P. Jannasch and W. Gottschalk 45 have studied the use of ozone as a reagent for the
qualitative detection and the gravimetric precipitation of manganese salts in analysis.
Colourless glass, in which the colour of the iron salts has been bleached by manganese
oxide, is slowly coloured violet by exposure to ozone or to ultraviolet rays. Accord-
ing to F. Fischer, this transformation does not occur so readily in the near sea-level,
as on high mountains where the ozone and the ultraviolet rays of the sun are more
prevalent.
Ozone oxidizes ferrous salts to ferric salts 46 in neutral and acid solutions. The
reaction, according to R. Luther and H. J. K. Inglis is in accord with 2Fc(NH4)2(S04)2
-|-03=Fe20(S04)2-f2(NH4)2S04+02, but Y. Yamauchi found rather more ferrous
iron was oxidized than corresponds with this equation, and hence he inferred that
OZONE AND HYDROGEN PEROXIDE 911
some of the oxygen also does work in oxidation. Ferrocyanides are oxidized to
ferricyanides. Potassium carbonyl ferrocyanide, KsFeCOCys, is not oxidized to
the fcrri-salt by ozone. In alkaline solutions ferri-salts are oxidized to ferrates by
ozone: Fe203-j-03+2H20=2H2Fe04, which suggests an ozonide structure for the
ferrates. P. Jannasch and W. Gottschalk could not oxidize nickel nitrate solutions
by ozone, but cobalt sulphate gave a dark-brown precipitate. Cobalt and nickel
sulphides or hydroxides are immediately oxidized to the peroxide — the sulphur of
the sulphide forms sulphuric acid. Ozone oxidizes chiomic salts to chromates or
dichromates, but it does not carry the oxidation further.
Ozone precipitates gold from acid-free solutions of gold chloride or colours the
solutions dark-violet. This was therefore recommended as a test for ozone by
R. Bottger.47 Ozone precipitates palladium dioxide from palladium salt solutions,
and it forms complex salts, palladiates, in alkaline solutions. E. Schneider recom-
mended palladous chloride for making test papers for ozone. The reaction is
usually represented by the equation: HaPdC^-f 4H20+03=Pd(OH)4+4HCl
-fH20+02.
Ozone rapidly corrodes rubber and rubber compounds ; concentrated ozone
will eat through rubber tubing in a few moments ; cork withstands dilute ozone
for a short time, but its use should be avoided. Ozone attacks many organic com-
pounds in the cold. Methane gives formaldehyde and formic acid ; ethylene forms
acetaldehyde and acetic acid ; ethylene reacts explosively, giving carbon and water ;
alcohol forms aldehyde and acetic acid ; ether forms aldehyde and acetic acid and a
syrupy liquid ethyl peroxide, (02115)202, which is explosive, and which forms alcohol
and hydrogen peroxide with water. Nitroglycerol, dynamite, nitrogen chloride,
and nitrogen iodide explode in an atmosphere rich in ozone. Benzene forms formic,
acetic, oxalic, and other acids as well as a white gelatinous explosive compound
called OZObenzene. The phenols are slowly attacked ; aniline forms ozobenzene,
etc. The vegetable colours are quickly bleached by ozone — indigo blue forms
colourless isatine. Tincture of guaiacum is coloured blue by ozone. The two last-
named reactions have been used as tests for ozone. The colouring matter of blood
as well as the albumins are destroyed by ozone.
The uses of ozone. — Ozone is used for the purification of water. Its function
is to oxidize the organic matter, and sterilize the water. So satisfactory is the
ozone treatment in the sterilization of water that it is declared by recognized
authorities : " The purification of drinking water has ceased to be a problem."
Water of average quality can be satisfactorily treated with ozonized air containing
one part of ozone per million, and the cost of treatment lies between 45. Od. and 85. Od.
for 1,000,000 gallons per hour. There are not far from 100 ozone water-plants
giving satisfactory results. Ozonized air is also used in ventilation. The ozone
treatment of air is not intended to supplant ventilation, but is to supplement it,
by providing an agent which destroys bad odours, and the feeling of closeness
experienced when a number of people are in a confined space. The concentration
of the ozone for this purpose should be smaller than can be recognized by smell —
say less than a milligram per cubic metre. Ozone or galvanized air, said Leuch,
in 1849, is a most powerful bleaching agent. It is now used for bleaching purposes
particularly for paper pulp ; starch ; oils ; and for oxidizing oil in the manufacture
of linoleum, etc. H. de la Coux ^8 has given a long list of possible applications,
but many have been reported failures on trial — ^sometimes because the ozone has
been wrongly applied, e.g. in too concentrated a form for bleaching certain textiles ;
and sometimes because some unexpected complication has arisen, e.g. in bleaching
flour where the taste is affected, and in bleaching dextrine and glue where the
adhesive properties are impaired by ozone bleaching. The industrial applications
have stimulated inventors, and accordingly, a number of fairly efficient ozonizers
have been placed on the market. In most of these, a high tension alternating
electric discharge is sent across a space through which the air to be ozonized
passes.
912 INORGANIC AND THEORETICAL CHEMISTRY
References.
» B. C. Brodic, Phil. Trans., 162. 435, 1S72 ; V. Rothmund and A. Burgstaller, Monatsh.,
34. CiGu, 1913 ; C. D. Harries, ZciL Elcktrochem., 18. 129, 1912 ; L. von Babo, Liehig's Ann. Svppl.,
2. 265, 1863.
» T, Andrews and P. G. Tait, Phil. Trans., 150. 113, 1800 ; L. von Babo and C. E. Claus,
Liebig'.^ Ann. Suppl, 2. 297, 1863.
» M. Berthelot, Compt. Bend., 86. 76, 1878.
« L. Palmieri, Compt. Rend., 74. 1266, 1872 ; A. Houzoau, ib., 74. 1267, 1872.
6 C. F. Schonbein, Verh. Nat. Gcs. Basel, 4. 67, 1842; W. A. Shonstone and T. A. Cundall,
Journ. Chem. Soc., 51. 610, 1887 ; D. L. Chapman and H. E. Jones, ib., 97. 2463. 1910 ; 99.
1811, 1911 ; D. L. Chapman and H. E. Clarke, ?7>., 93, 1638, 1908 ; K. Luther and J. K. H.
Inglis, Zeit. phys. Chem., 43. 203, 1903; S. .Tahn, Zeit. anorg. Chem., 60. 292, 337, 1908;
E. Mulder and H. G. L. van der Meuien, Bee. Trav. Chim. Pays-Ba-s, 1. 167, 1882.
• C. F. Schonbein, Verh. Nat. Ges. Basel, 4. 67, 1840.
' E. Warburg, Ann. Physik, (4), 9. 1286, 1902 ; (4), 13. 1080, 1904 ; S. Jahn, Zeit. anorg.
Chem., 48. 260, 1906; E. P. Perman and R. H. Greaves, Proc. Boy. Soc., 80. A, 353, 1908;
E. Warburg, Sitzber. Akad. Berlin, 1120, 1901 ; D. L. Chapman and E. H. Jones, Journ. Chem. JSoc,
97. 2463, 1910.
« M. Beger, Zeit. Elektrochem., 16. 76, 1910; J. Dewar, Proc. Boy. Soc, 43. 1078, 1888;
J. Chappius, Bidl. Soc. Chim., {2), 35. 419, 1881 ; A. Schuller, Wied. Ann., 15. 289, 1882 ; J. Meyer,
Journ. prakt. Chem., (2), 72. 293, 1905 ; E. Fahrig, Chem. News, 62. 39, 1890 ; E. Ritsert, PJmrm.
Ztg., 35. 372, 1890 ; M. Otto, Compt. Bend., 123. 1005, 1896 ; Ann. Chim. Phys., (7), 13. 77,
1898 ; P. Richarz and R. Schenek, Sitzber. Akad. Berlin, 490, 1904 ; 1 102, 1903.
» R. Schenek, Sitzber. Akad. Berlin, 37, 1904 ; R. Schencic and F. Mihr, Ber., 37. 3464, 1904 ;
L. (iraetz, Phys. Zeit., 5. 688, 1904 ; F. Richarz, ib., 6. 1, 1905 ; K. Schaum, ib., 6. 73, 1905.
i» B. C. Brodie, Phil. Trans., 162. 435, 1872.
" E. H. Riesenfeld and T. F. Egidius, Zeit. anorg. Chem., 85. 217, 1914 ; Y. Yamauchi, Amer.
Chem. Journ., 49. 55, 1913.
" M. Berthelot, Cornpt. Bend., 88. 50, 1879 ; Ann. Chirn. Phys., (5), 17. 142, 1872 ; E. Coman-
ducci. Bend. Accad. Napoli, (3), 15. 15, 1909; C. Pickel, Zeit. anorg. Chem., 38. 307, 1904;
F. Fischer and M. Wolf, Ber., 44. 2956, 1911 ; F. Woigert, ib., 46. 815, 1913 ; Zeit. phys. Chem.,
90. 189, 223, 1915 ; H. Thiele, Zeit. angew. Chem., 22. 2472, 1909.
** C. F. Schonbein, Journ. prakt. Chem., (1), 83. 86, 1861; G. Meissner, Untersuchungen ilber
den Sauerstoff, Hanover, 1863; L. von Babo, Liebig's Ann. Suppl., 2. 265, 1863; C. Engler
and A. Nasse, Liebig's Ann., 154. 215, 1870 ; L. Carius, ib., 174. 12, 1874 ; Ber., 5. 520, 1872 ;
E. Schone, ib., 6. 1224, 1873: Liebig's Ann., 171. 87. 1874; C. Gianetti and A. Volta, Oazz.
Chim. ftal, 4. 471, 1874 : M. Berthelot, Compt. Bend., 88. 50, 1879 ; C. Arnold and C. Mentzel,
Ber., 35. 2902, 1902 ; P. Jannasch and W. Gottschalk, Journ. prakt. Chem., (2), 73. 496, 1906 ;
V. Rothmund and A. Burgstaller, Moruitsh., 34. 665, 1913 ; A. Schmidt, Hcematolzylin
Studien, Dorpat, 1865.
14 C. F. Schonbein, Journ. prakt. Chem., (1), 75. 80, 1858; (1), 78. 71, 1859; (1), 77. 130,
1858 ; E. Schone, Liebig's Ann., 196. 239, 1879 ; H. J. K. Inglis, Journ. Chem. Soc, 83. 1010,
1903 ; H. Mcleod, ib., 37. 118, 1880 ; C. Arnold and C. Mentzel, Ber., 35. 2902, 1902 ; C. Engler
and W. Wild, ib., 29. 1940. 1896 ; F. Fischer and M. Wolf, ib., 44. 2956, 1911 ; V. Rothmund and
A. Burgstaller, Moruitsh., 38. 295, 1917.
16 H. Moissan, Ann. Chim. Phys., ^6), 24. 224, 1891 ; O. Ruff and J. Zedncr, Ber., 42. 1037,
1909 ; G* GaUo, Atti Accad. Lincei, (5), 19. i, 295, 753, 1910.
1* E. Comanducci, Bend. Accad. Napoli, (3), 15. 15, 1909; P. Hautcfeuille and J. Chappius,
Ann. Chim. Phys., (6), 2. 282, 1884 ; D. L. Chapman and P. S. McMahon, Journ. Chem. Soc, 97.
845, 1910.
1' C. F. Schonbein, Verh. Not. Ges. Basel, 7. 23, 1847; J. Ogier, Compt. Bend., 85. 957, 1877 ;
86. 722, 1878 ; P. Chretien, ib., 123. 814, 1911 ; F. Fichter and F. Rohner, Ber., 42. 4093, 1909 ;
M. Beger, Chem. Ztg., 33. 1232, 1909 : N. A. E. Millon, Journ. prakt. Chem., (1), 34. 321, 1845 ;
H. B. Baker and R. J. Strutt, Proc Phys. Soc, 23. 150, 191 1.
18 0. Brunck, Ber., 33. 1832, 1900 ; E. Pochard, Compt. Bend.. 130. 1705, 1900 ; K. Garzarolli
Thurnlackh, Sitzber. Akad. Wien, 110. 787, 1901 ; Monatsh., 22. 955, 1901 ; W. Brav, Zeit. phys.
Chem., 54. 463, 1906 ; J. N. Pring, Chem. News, 109. 73, 1914 ; Proc Boy. Soc, 90. A, 204, 1904.
1' B. C. Brodie, Phil. Trans., 162. 435, 1872; (}. Meissner, Neue Untersuchungen fiber den
elektrisierten Saue/stoff, Gottingen, 1869 ; L. von Babo and C. E. Claus, Liebig's Ann. Svppl, 2.
297, 1863 ; J. L. Soret, Liebig's Ann., 130. 95, 1864 ; 138. 45, 1866 ; O. Brunck, Ber., 33. 1832,
1900; Zeit. anorg. Chem., 10. 235, 1895: Zeit. angew. Chem., 16. 894, 1903; E. Ackermann,
Chem. Ztg., 24. 235, 1900 ; A. Ladenburg and R. Quasig, Ber., 34. 1184, 1901 ; F. P. Treadwell
and E. Anneler, Zeit. anorg. Chem.. 48. 86, 1906 ; R. Luther and H. J. K. Tnclis, Zeit. jthys. Chem.,
43. 203, 1906 ; E. H. Riesenfeld and F. Bencker, Zeit. anorg. Chem., 98. 167, 1916 ; G. Lechner,
Zeit. Elektrochem., 17. 412, 1911 ; E. Czako, Journ. Gasbeleuchtung, 55. 768. 1912; C. Engler
and A. Nasse, Liebiq's Ann., 154. 215, 1870 ; G. Lechner, Zeit. Elektrochem., 17. 414, 1911.
20 V. Rothmund, Phys. Zeit.. 18. 113, 1917; Monatsh., 39. 571, 1918; G. Meissner, Unter-
nnchun/]fn Uber den Sauerstoff, Hannover, 1 863.
21 E. P'onrobert, Das Ozon, Stuttgart, 89, 1916; E. Pollacci, Bend. 1st. Lombardo, (3), 17. 198,
OZONE AND HYDROGEN PEROXIDE 913
1884 ; H. B. Baker and R. J. Strntt, Proc. Phys. Soc, 23. 149, 1911 ; A. Mailfert, Compt. Bend.,
94.860, 1186, 1882; M. Berthelot, t6., 86. 20,277, 1878; T. Weyl, Chem. Ztg., 25. 291,1901;
A. Stock and K. Friederici, Ber., 46. 1380, 1913 ; J. Schmidlin and P. Massini, ih., 43. 1162, 1910 ;
I). Helbig, Eep. anal. Chem., 1. 148, 1881 ; P. V. Langlois and L. S. Thomassin, Bent. Ind. Ztg.,
372, 1870 ; E. H. Riesenfeld, German Pat., D.R.P. 229, 274, 1909 ; A. Reynoso, Dingler's Jonrn.,
219. 472, 1876 ; B. Hunt, Brit. Pat. No., 2181, 1874 ; E. H. Riesenfeld and T. F. Egidius, Zeit.
anorg. Chem., 85. 217, 1914; A. ^ovcYievs, Studien ilber Gasreaktionen, Freiburg i. B., 1911.
22 J, Brand, Ann. Physik, (4), 9. 468, 1902 ; L. Grafenberg, Zeit. anorg. Chem., 36. 355, 1903 ;
J. Jannek and J. Meyer, ih., 83. 51, 1913; M. Berthelot, Compt. Rend., 86. 20, 277, 1878;
•J. Sohraidlin and P. Massini, Ber., 43. 1162, 1910 ; V. Rothmund and A. Burgstaller, Monatsh.,
34. 665, 1913 ; A. Borchers, Studien iiher Gasreaktionen, Freiburg i. B., 1911.
23 M. Berthelot, Compt. Bend., 83. 933, 1876 ; 84. 61, 1877 ; Ann. Chim. Phys., (5), 12. 446,
1877 ; L. Carius, Ber., 6. 806, 1873 ; Liehig's Ann., 174. 1, 1874 ; C. F. Schonbein, Journ. prakt.
Chem., (1), 84. 193, 1861 ; E. Fonrobert, Das Ozon, Stuttgart, 1916 ; C. Montanari, Chem. Ztg.,
32. 722, 1908.
2* C. F. Schonbein, Denkschrift uher des Ozon, Basel, 1849 ; T. Andrews, Phil. Trans., 150.
113, 1861 ; B. G. Brodie, ih., 162. 450, 1872 ; L. von Babo, Liehig's Ann. Suppl, 2. 294, 1863 ;
J. Ghappius and P. Hautefeuille, Compt. Rend., 92. 80, 1881 ; 94. 946, 1111, 1882 ; A. Houzeau,
ih., 70. 39, 1286, 1870 ; Ann. Chim. Phys., (4), 22. 150, 1871 ; R. Bottger, Jahr. Phys. Ver.
Frankfurt, 12, 1872 ; E. Schone, Ber., 6. 1208, 1873 ; F. Fischer and E. Hene, ib., 46. 603, 1913 ;
G. Meissner, Untersuchungen ilber den Sauerstoff, Hannover, 1863 ; Neue Untersuchungen uher
des electrische Sauerstoff, Gottingen, 1869.
25 R. J. Strutt, Proc. Roy. Soc, 85. A, 219, 1911 ; 86. A, 56, 105, 1912 ; Journ. Chem. Soc,
113. 200, 1918 ; T. M. Lowry, ib., 101. 1152, 1912 ; W. Cramp and B. Hoyle, Proc Inst. Elect.
Eng., 42. 328, 1909.
28 P. Hautefeuille and J. Ghappius, Compt. Rend., 92. 80, 1881; 94. 940, 1111, 1882;
V. Ehrlich and F. Russ, Monatsh., Z2. 917, 1911 ; 36. 317, 1915 ; E. Warburg and G. Leithauser,
A7in. Physik, (4), 23. 209, 1907 ; G. Leithauser and R. Pohl, Ber. dcM. phys. Ges., 6. 249, 1908 ;
W. Manchot, Ber., 41. 471, 1908 ; F. vonLepel, Ber., 30. 1027, 1897 ; F. Fischer and F. Brahmar,
ih., 39. 940, 1906; F. Fischer and H. Marx, ib., 39. 2557, 3631, 1906; 40. 443, 1111, 1907;
A. Mand and F. Russ, 7jeit. angew. Chem., 21. 486, 1908.
27 M. Berthelot, Ann. Chim. Phys., (5), 14. 367, 1878; L. Carius, Liehig's Ann. 174. 1, 1874 ;
D. Helbig, Atti Accad. Lincei, (5), 11. 311, 1902 ; (5), 12. 166, 1903 ; E. Miiller, Zeit. anorg. Chem.,
76. 324, 1912.
28 F. Foster and M. Koch, Zeit. angew. Chem., 21. 2161,2209, 1908 ; F. Russ and V. Ehrlich,
Monatsh., 32. 917, 1911 ; 36. 317. 1915.
29 L. Carius, Liebig's Ann., 174. 49, 1874 ; Ber., 7. 1481, 1874 ; L. I. de N Ilosva, ib., 27. 3500,
1894; G. Baumert, Pogg. Ann., 100. 30, 1857 ; C. F. Schonbein, Journ. prakt. Chem., (1), 89. 7,
323, 1863 ; C. Montanari, Chem. Ztg., Z2. 722, 1908 ; D. Helbig, Rep. anal. Chem., 1. 148, 1881 ;
P. Jannasch and W. Gottschalk, Journ. prakt. Chem., (2), 73. 497, 1906 ; E. Warburg, Sitzber.
Akad. Berlin, 216, 1912.
30 A. W. Browne and F. F. S. Netterly, Journ. Amer. Chem. Soc, 31. 221, 1909.
31 J. Schmidlin and P. Massini, J?er., 43. 1162, 1910.
32 C. F. Schonbein, Liebig's Ann., 89. 257, 1854; Y. Yamauchi, Amer. Chem. Journ., 49. 55,
1913 ; A. Stock and W. Siebert, Ber., 38. 3837. 1905.
33 A. Besson, Compt Rend., 121. 125, 1895.
3* A. Stock and K. Friederici, Ber., 46. 1380, 1913.
35 T. Rerasen and M. S. Southworth, Ber., 8. 1414, 1875 ; Amer. Chem. Journ., 4. 50, 1882;
E. H. Keiser, Hh., 5. 424, 1883 ; M. Berthelot, Com.pt. Rend., 88. 50, 1879 ; Ann. Chim. Phys.,
(5), 17. 142, 1879 ; A. R. Leeds, Ainer. Chem. Journ., 4. 454, 1883 ; F. Baumann, Zeit. physiol.
Chem., 5. 244, 1881 ; E, Goldstein, ib., 36. 3042, 1903; R. Koetschau, Liebia's Ann., 374. 321,
1910
38 W. A. Jones, Amer. Chem. Journ., 30. 40, 1903 ; C. E. Waters, ih., 30. 50, 1903 ; W. Manchot
and W. Kampschulte, Ber., 40. 2891, 1907 ; 42. 3942, 1909 : P. Clausmann, Compt. Rend., 150.
1332, 1910; Bull. Soc Chim., (4), 7. 827, 1910; 9. 1207, 1911 ; J. Thiele, Zeit. avgew. Chem.,
22. 2472 1909.
3' A.' Besson and L. Fournier, Compt. Rend., 148. 1192, 1909.
38 Griesheim Elektron chemischen Fabrik, German Pat., D.R.P. 274872, 1913.
39 C. F. Schonbein, Journ. prakt. Chem., (1), 45. 469, 1866 ; W. Manchot and W. Kampschulte,
Ber., 40. 4984, 1907 ; 46. 1089, 1913 ; W. A. Shenstone and J. T. Cundall, Journ. Chem. Soc,
51. 611, 1887.
40 A. von Baever and V. Villiger, Ber., 35. 3038, 1902 ; 40. 4984, 1907 ; W. Traube, ib., 49. 1670,
1916.
41 W. Manchot and W. Kampschulte, Ber., 40. 4984, 1907; 46. 10S9, 1913 ; A. Bach, ib., 35.
3424, 1902; W. Traube, ib., 45. 2201, 3319, 1912; L. Grafenberg, Zeit. anorg. Chem., 36. 355,
1903.
► "2 rp Andrews and P. G. Tait, Phil. Mag., (4), 17. 436, 1859 ; W. A. Shenstone and J. T. Cundall,
Journ. Chem. Soc, 51, 611, 1887; A. Volta, Gaz~. Chim. Ital.,9. 521, 1879; W. Manchot and
W. Kampschulte, Ber., 40. 2891, 1907 : 43. 7.W, 1910.
43 A. Volta, Gazz, Chim. Ital, 9. 521, 1879 ; E; Fn'my, Compt. Rend., 61. 939, 1865 ; A. Hou-
VOL. I. 3 N
9U INORGANIC AND THEORETICAL CHEMISTRY
zeau, Ann, Chim. Phyf^., (3), 62. 129, 1861 ; 0. Arnold and C. Mentzel, Ber., 35. 1324, 1902 ;
W. Manchot and W. Kampschulte, ib., 40. 2891, 1907 ; 42. 3942, 1909 ; W. Manchot. ib., 42.
3942, 1909 ; 43. 750, 1910 ; A. R. Leeds, ib., 12. 1831, 1879 ; H. Thiele, Zeit. offentl. Chem., 12.
11, 1900.
** Y. Yaraauchi, Amer. Chem. Journ., 49. 55, 1913 ; R. Bottger, Jonrn. prakt. Chem., (1), 95.
311, 1865 ; E. Schone, Liebig's Ann., 196. 58, 1879 ; L. Moser, Zeit. anorg. Chem., 54. 121, 1907 ;
N. SchilofF, Zeit. phys. Chem., 42. 641, 1903 ; P. Jahnasch and W. Gottschalk, Journ. j)rakt.
Chem., (2), 73. 496, 1906; C. F. Schonbein, Liebig's Ann., 61. 13, 1867; A. W. Williamson,
ib., 89. 293, 1854; B. Huizinga, Journ. prakt. Chem., (1), 102. 201, 1867; C. Engler and
W. WUd, ib., (2), 73. 496, 1906 ; L. Maquenne, Compt. Bend., 94. 795, 1882 ; A. Mailfert, ib., 94.
860, 1186, 1882.
" P. Jannasch and W. Gottschalk, Jo%irn. prakt. Chem., (2), 73. 496, 1906.
*« R. Luther and H. J. K. Inglis, Zeit. phys. Chem., 43. 203, 1903 ; Y. Yamauchi, Amer. Chem,.
Journ., 49. 65, 1913 ; J. A. Miiller, Bull. Soc. Chim., (3), 29. 1158,. 1903 ; P. Jannasch and
W. Gottschalk, Journ. prakt. Chem., (2), 73. 496, 1906 ; A. Mailfert, Compt. Rend., 94. 860, 1882.
*' R. Bottger, Jahr. Phys. Ver. Frankfurt, 17. 1879 ; A. Mailfert, Compt. Rend,., 94. 860, 1882 ;
E. Schneider, Repert. anal. Chem., 1. 54, 1881 ; Y. Yamauchi, Amer. Chem. Journ., 49. 55,
1913.
** H. de la Coux, V ozone et aes applications industrielles, Paris, 1910 ; A. Vosmaer, Ozone :
its Manufacture, Properties, and Uses, London, 1916.
§ 7. The Constitution of Ozone
The nature of ozone was the subject of much discussion soon after C. F. Schonbein
had established the individuality of the gas. At first, C. F. Schonbein seems to
have thought the gas was a new elementary body which belonged to the same class
of electronegative elements as chlorine and bromine ; later on, he suggested that it
may be an elementary substance which, when united with hydrogen, forms the
nitrogen of the atmosphere.^ Henry Cavendish had proved, in 1784, that nitrates
are produced when electric sparks pass through air ; and L. Kivier and L. II . de
Fallenberg (1845) showed that nitrous acid is formed during the oxidation of
phosphorus. Nitrous acid like ozone colours starch paper blue ; it was at first
supposed that the reactions ascribed by C. F. Schonbein to ozone were really pro-
duced by traces of nitrous acid. Ozone can be produced imder conditions where
no nitrogen is present ; consequently, it follows that ozone contains no nitrogen.
Oxygen is an invariable antecedent when ozone is formed, and an invariable con-
sequent when ozone is decomposed.
Is ozone a condensed form o! oxygen or an oxide o! hydrogen ?— About
1845, J. C. G. de Marignac and A. de la Rive showed that moist silver, when
exposed to ozone, forms a silver oxide, and that potassium iodide — KI — can be
oxidized by ozone to potassium iodate— KIO3 ; but there is never any sign of the
formation of any nitrogen compound. This narrowed the question, for it appeared
that ozone is either (1) a form of matter identical with oxygen — J. C. G. de Marignac
and A. de la Rive (1845) ; or (2) oxidized water, that is, a peroxide of hydrogen —
C. F. Schonbein and A. W. Williamson (1845). On the one hand, J. C. G. de
Marignac and A. de la Rive, in their memoir, Sur la production et la nature de
rozone,^ showed that ozone can be obtained by the electrolysis of water free from
nitrogen, and that ozone containing nothing but oxygen must be an allotropic form
of that element. This conclusion was supported by J. J. Berzelius. On the other
hand, C. F. Schonbein (1817) contended that ozone is a higher oxide of hydrogen
than L. J. Thenard's bioxyde dliydrogene. This view was supported by A. W.
Williamson because he obtained water by passing ozonized ox5^gen over heated
copper oxide. A. W. Williamson's gas was obtained by electrolysis. E. Fremy
and E. Becquerel demonstrated in their memoir, Recherches electrochiwque sur les
jyroprietes des corps electrises, that pure oxygen can be converted into ozone by the
prolonged action of electricity; they also found the ozone to be absorbed by
mercury or by potassium iodide, as fast as it was produced. G. Baumert then
suggested that the ozone obtained by the electrolysis of acidulated water is different
OZONE AND HYDROGEN PEROXIDE
915
1
1/
n
u
from that obtained by the electrical discharge in oxygen, or by the action of phos-
phorus on moist oxygen, because the electrolytic ozone always contains hydrogen
and the other form of ozone contains nothing but oxygen. In 1856, in a paper,
On tJie constitution and properties of ozone, T. Andrews showed that the difficulty
with electrolytic ozone was due to the presence of impurities in the gas, and if proper
precautions be taken :
No gaseous compound having the composition of a peroxide of hydrogen is formed
during the electrolysis of water ; and that ozone, from whatever source derived, is one
and the same body, having identical properties and the same constitution, and is not a com-
pound body, but oxygen in an altered or allotropic condition.
About this time, ozone was variously styled nascent oxygen, implying that it was
oxygen in an atomic condition ; active oxygen or erregten Sauerstoff, in reference
to its great chemical activity ; and polarized oxygen or electrisierten Sauerstojf.
The hydrogen oxide theory was not given up until 1860, when T. Andrews and
P. G. Tait 3 proved in their paper, On the Volumetric relations of ozone, that if an
electric discharge — brush or spark — be passed through pure dry oxygen, a contrac-
tion occurs amounting to about one-twelfth of the original volume. The oxygen
was sealed in a tube, shaped as indicated in Fig. 7, Q, and subjected to the brush
discharge, via the platinum wires sealed into the glass. In T. Andrews and P. G.
Tait's experiment the contraction in volume was measured by attaching to the
tubes a small manometer, a and h, charged
with concentrated sulphuric acid. A dupli-
cate tube, Q, Fig. 8, containing air was
treated along with the tube containing the
oxygen, R, Fig. 7, so that any changes due
to variations of pressure or temperature
during the experiment could be corrected.
The tubes, during the experiment, were
placed in a water tank as indicated in Fig.
8, ^to keep the temperature uniform. When
ozonized oxygen is heated to 270°, and
allowed to cool, the original volume of
oxygen is obtained ; and when a thin glass
bulb, c, Fig. 7, R, of potassium iodide is
sealed in the tube along with the oxygen, and after ozonization, broken by
shaking the bulb against a piece of glass tubing d, iodine is liberated without any
perceptible change in volume. If the gas which has been treated with potassium
iodide be heated to 270° as before, no change in volume can be detected. Hence,
T. Andrews and P. G. Tait concluded that ozone is a condensed form of oxygen.
This statement, however, gives no information about the weight of oxygen in a given
volume of ozone, i.e. the number of atoms in the molecule of ozone.
The absence of hydrogen in ozone was further confirmed by A. Houzeau (1868),
L. von Babo (1863) and by J. L. Soret (1863) ^ in an experiment in which ozone
was thoroughly dried, and then decomposed by heat. No trace of any compound
of hydrogen — e.g. water — could be detected in the products of decomposition.
Hence, it is inferred that ozone is not a compound of hydrogen with oxygen ;
ozone contains nothing but oxygen ; it is a kind of oxidized oxygen. A similar
experiment was made by C. F. Schonbein in 1849, but its importance does not appear
to have been appreciated at that time.
Ozone is a form of oxygen in which three volumes of oxygen are con-
densed to two volumes. — Bince the volume of ozonized oxygen undergoes
no change when mixed with a solution of potassium iodide, it is inferred that the
oxidation of potassium iodide can only be effected by so much oxygen in ozone
as has been condensed with ordinary oxygen to form ozone. This excess of oxygen
is absorbed b}^ the solution of potassium iodide, and the ordinary oxygen which
R
[G. 7. — Andrews
and Tait's
Ozone Tubes.
Fig. 8. — Andrews
and Tait's Ex-
periment.
916
INORGANIC AND THEORETICAL CHEMISTRY
remains has the same volume as the ozone present before the action of the potassium
iodide. Hence, no new contraction occurs with potassium iodide. In symbols,
the formula for ozone is 02 + n- T. Andrews and P. G. Tait did not determine
the numerical value of n. The formula for ozone might be O3, O4, O5, . . . The
special difficulty involved in this determination arises from the fact that ozone
cannot be obtained free from oxygen ; and, accordingly, the regular methods of
determining the molecular weights — vapour density, etc. — cannot be applied.
W. Odling (1861) ^ proposed to take the simplest possible formula, O3, thus assuming
that three volumes of oxygen are condensed to form two volumes of ozone ; he said :
Tf we consider ozone to be a compound of oxygen with oxygen, and the contraction to
be consequent upon their combinations, then, if one portion of this combined or contracted
oxygen were absorbed by the reagent, the other portion would be set free, and by its
liberation might expand to the volume of the whole. Thus, if we suppose three volumes
of oxygen to be condensed by their mutual combination into two volumes, then on absorbing
one-third of this combined oxygen by mercury, the remaining two-thirds would be set
free and consequently expand to their normal bulk, or two volumes :
000+Hg2=Hg20 + 00
Interpreting this assumption in the light of Avogadro's rule :
.vO.--:-.^!;; ••.*;•..:.■•.•: •.■.••*••..*•.•■:>
A
<8
Fia. 9. — 3 Volumes of Oxygen give 2 Volumes of Ozone.
This beautiful hypothesis, said C. W. Heaton (1868), although accounting perfectly
for all known facts, is yet but a probability. One link in the chain of evidence is
lacking. True, J. Tyndall ^ inferred that ozone contains more atoms per molecule
than does ordinary oxygen because the former has much greater absorptive power
for heat than the latter ; but the missing link was not supplied until J. L. Soret
described some happily devised experiments in his memoir, Recherches sur la
dcnsile de V ozone (1866).
J. L- Soret's experiments.— J. L. Soret (1866) 7 took advantage of the fact,
known to C. F. Schonbein, that essential oils absorb ozone without absorbing any
(B) After the ozone has
been converted hito
oxygen by heat.
(C) After removal
of ozone with
cinnamon oil.
Pio. 10. — Soret's Experiments with Ozone (Diagrammatic).
marked quantity of oxygen. Hence a sample of ozonized oxygen was introduced
into two flasks. A, Fig. 10, and A^. The vessel A was heated so as to doconipose
the ozone. The gas now occupied a greater volume than before— J5, Fig. 10. The
expaasion was measured when the gas had cooled to its former temperature. The
vessel Ai, containing the same mixture of ozonized oxygen, was treated with
cinnamon oil ; the contraction due to the removal of ozone was measured— C.
J. L. Soret (1866) found the following data with measured volumes of a given
sample of ozonized oxygen :
Expansion after heating ....
Contraction after treatment with cinnamon oil
^i'lry c.c.
0-90 c.c.
OZONE AND HYDROGEN PEROXIDE 917
Consequently, the ozonized oxygen contained 690 c.c. of ozone ; and 6*90 c.c.
of ozone becomes 10"35 c.c. of ordinary oxygen when lieated ; or 2 c.c. of ozone
becomes 3 c.c. of oxygen. In another set of experiments :
Contraction after treatment with turpentine oil . . . 8"0 c.c.
Oxygen absorbed by treatment with potassium iodide . . 3"9 c.c.
Hence 8*0 c.c. of gas gave up 3*9 c.c. of oxygen to potassium iodide solution with-
out change of volume. This means that 8 c.c. of ozone is equivalent to 8'0-3"9 c.c.
of oxygen within the limits of experimental error. L. von Babo and 0. E. Claus,^
in 1863, also found the decrease in the volume of oxygen on ozonization to be equal
to the volume of oxygen calculated from the quantity of iodine liberated by the
action of the ozone on potassium iodide. Hence, it was inferred that three
volumes o£ oxygen produce two volumes of ozone.
J. L. Soret's work was rather crude, but, in 1872, B. C. Brodie ^ repeated the
experiments with cinnamon oil, turpentine, and stannous chloride in such a way
that the above conclusion was the only possible interpretation of the experiment.
A. Ladenburg lo extended J. L. Soret's work to a mixture containing a large propor-
tion of ozone, and calculated the relative density of ozone — assumed unknown —
from the density of a sample of ozonized oxygen, and the amount of iodine liberated
by the gas when in contact with potassium iodide, taken in conjunction with the
fact that the gas undergoes no change in volume during the reaction. A. Ladenburg
tacitly assumed that a molecule of iodine is liberated by a molecule of ozone. The
argument is accordingly fallacious because if the molecule of ozone be represented
^y On-\-nO, each molecule of ozone will liberate n molecules of iodine. The excess
of the density of ozonized oxygen over that of pure oxygen represents the weight
of oxygen available for the liberation of iodine, and this is not necessarily dependent
on the molecular weight of ozone.
Example.— If 100 c.c. of ozonized oxygen are converted into oxygen by passage through
a hot tube, what was the composition of the original mixture if 110 c.c. of oxygen remained ?
The mixture contained x c.c. of ozone, and (100— a;) c.c. of oxygen, and since x volumes
of ozone yield frc c.c. of oxygen, the original mixture contained the equivalent of
fx' + (100 — a:) = 110 c.c. of oxygen. Hence, ic = 20, or the mixture contained 20 per cent,
of ozone and 80 per cent, of oxygen.
M. Otto (1897) 11 determined the density of ozone by weighing a bulb filled with
oxygen, and again when filled with ozonized oxygen. The increase in weight gave
the weight of the active oxygen in ozone. The amount of ozone was determined
by means of an acidified solution of potassium iodide. A. Ladenburg (1901)
measured the total volume of ozone by absorption with turpentine. The mean
of A. Ladenburg's five determinations was 47*78 — with 4:5'3 and 50'4 as extreme
values. J. L. Soret (1867) confirmed his determination of the molecular weight of
ozone by assuming T. Graham's relation between the speed of the diffusion of a
gas and its density. He allowed vessels of chlorine and oxygen to be in communica-
tion for a given time ; and likewise vessels of ozonized oxygen and oxygen. The
amounts of chlorine and ozone which diffused in a given time were as 0'227 : 0*271.
Consequently, if the density of chlorine is 2 '4:9, and Z), the density of the ozone,
air unity, by Graham's law, Z> : 2'49=(0'227)2 : (0-271)2 ; hence, D is nearly I'S;
A. Ladenburg obtained the density 1*3698 by Schilling's apparatus. All these
numbers are in accord with a molecular weight 48 (oxygen 32) for ozone.
The frequent formation of dimeric (doubled) ozonides at low temperatures,
suggested to C. D. Harries the possibility that ozone may be itself dimeric, (03)2 ;
at low temperatures and monomeric, O3, at higher temperatures : (03)2^203 ;
nitrogen peroxide behaves similarly, (N02)2^2N02.
Newth's experiment.- — Many neat ways of illustrating the volume relations of oxygen
and ozone have been devised.^'' G. S. Newth's apparatus (1896), slightly modified, consists
918
INORGANIC AND THEORETICiVL CHEMISTRY
of two concentric tubes. Fig. 11. The inner tube has a hoUow stopper ground to fit the
outer tube ; it contains dilute sulphuric acid. The inner tube has two httle projections, A,
and the outer tube has three projections, B, in such a position that a sealed thin glass
tube containing cinnamon oil can be broken, when desired, by twisting the stopper
of the inner tube. The outer tube is fitted with a three-way cock, D, connected with a
manometer charged with concentrated sulphuric
a,cid. The apparatus is placed in a cylinder con-
taining, say, ice and water. The annular space
between the two tubes is filled with oxygen, via
the cocks E and D. The manometer is then put
in communication with the annular space between
the two tubes. Note the level of liquid in the
manometer. Pass a current from an induction coil,
so as to ozonize the oxygen sufficiently to give,
say, a one-centimetre contraction on the mano-
meter. Note the contraction. Give the stopper a
twist so as to break the glass tube containing the
cinnamon oil; the contraction which occurs v/ill be
twice the former contraction, namely 2 cm. more.
It may be advisable to level the liquid in the mano-
meter, after the first contraction, by admitting
either air or oxygen, before breaking the capillary
tube. The same or a similar apparatus can be
employed for showing the contraction which occurs
when ozone is treated with potassium iodide by
using a tube C with this substance in place of
cinnanion oil.
Mano/t)
Fig. 11.- — Newth's Apparatus
(Modified).
The graphic or constitutional formula of
ozone. — If it be assumed that all three atoms of
oxygen are bivalent, the only possible formula
for ozone is the ring structure indicated in formula I. Owing to the peculiar oxidizing
qualities of the odd oxygen atom, some consider that the three oxygen atoms cannot
0 0
oAo
I
oAo
n
o<
o
m
o
oAo
IV
o4o
V
be symmetrically placed in the molecule, and this has given rise to other suggestions.
For example, M. Traube 13 assumes that two of the oxygen atoms in the ozone mole-
cule are tervalent and one bivalent, as indicated in formula II. The evidence for the
tervalency of oxygen is very weak since there are few compounds in which such an
assumption can be accepted ; nor does it correspond with the position of oxygen
in the periodic system. A. Angeli has drawn attention to the relationship between
ozone, O3, and azoimide, N3H ; both compounds are endothermal, explosive,
poisonous, and react with unsaturated compounds forming addition products —
ozone gives ozonides ; azoimide gives triazoles :
6 6
\/
o
Ozonide.
HC-CH
N NH
\/
N
Triazole.
From observations on the molecular refraction of hydrogen peroxide, J. W.
Briihl 1* inferred that this compound contains quadrivalent oxygen, and he also
assumed that ozone has one or more quadrivalent oxygen atoms as illustrated in
formula} III to V ; J. W. Briihl favoured formula IV or V. Formula V is related
to I in the symmetrical arrangement of its atoms. J. W. Briihl favoured IV, but
gave no particularly strong evidence in its support. It will be observed that even
though one atom in the ozone molecule appears to behave differently from the
others, this does not prove a dissymmetrical molecule, for the trinity is possibly
unstable because it is overloaded with atoms, and as soon as anr/ one atom has been
OZONE AND HYDROGEN PEROXIDE 919
ejected, the remaining pair is stable. On this hypothesis the greater chemical
activity of the odd oxygen atom is not due to its being oriented differently from the
other two, but rather to the molecule containing one atom too many for stability.
There is no objection to the assumed quadrivalency of oxygen. In fact, it is
highly probable that oxygen, like its companions — sulphur, selenium, and tellurium —
in the periodic system can be bi-, quadri-, or sexi-valent. 0. Wolkowicz ^^ concludes
that ozone has the constitution 0=0=0, indicated in formula III, by analogy
with sulphur dioxide 0=S=0. This would make ozone the anhydride of ozonous
acid, H2O4, analogous with sulphurous acid, H2SO3 ; and the so-called tetroxides
— e.g. K2O4 — analogous to the sulphites :
Ozone. Potassium ozouate. Sulphur dioxide. Potassium sulpliite.
Both series of salts reduce the permanganates. There is no decisive evidence in
favour of any one formula for ozone. The question cannot be answered by far-
fetched analogies, or by pen- and paper-abstractions. Accordingly, the graphic
or structural formula for ozone is still suh judice.
The quadrivalency of oxygen. — The idea that oxygen may be quadrivalent
was suggested by A. Naquet i^ in 1864 because of the analogy between oxygen and
the elements of the sulphur family — sulphur, selenium, and tellurium — where the
compounds SCI4, SeCl4, Tel4, etc., show that these elements are quadrivalent. Two
years later H. Buff applied the idea to hydrogen and barium dioxides, for he explained
the composition of these compounds by the formulae H2=0=0 and Ba=0=0
respectively. A. W. Williamson (1869) suggested that oxygen is quadrivalent in
carbon monoxide, C^O ; and about the same time, S. M. Jorgensen assumed that in
virtue of the quadrivalency of oxygen in water, the H20-group can act as a bivalent
radicle analogous to the bivalency of the NHs-group — nitrogen quinquevalent.
J. Thomsen (1873) also explained the constitution of periodic acid, by assuming
that oxygen is quadrivalent, and S. M. Jorgensen's idea was employed by W. A.
Tilden (1876) to explain the constitution of crystallized zinc sulphate, ZnS04.7H20 ;
and by A. Wurtz (1879), the constitution of potassium magnesium sulphate,
K2Mg(S04)2.
The investigation of C. Friedel,i7 in 1875, on the compounds of methyl ether,
(0113)20, with hydrogen chloride, sulphur dioxide, SO2, methyl iodide, CH3I,
established the existence of a series of compounds :
CHg-^^ CHa-^^^Cl CHs^^-^^O CHg-^^^I
in which it appears highly probable that in addition to the two ordinary valencies
of oxygen, two others are wakened into activity. In an analogous manner,
A. A. T. Cahours found that methyl sulphide, (0113)28, reacts with methyl iodide,
OH3I, to form a compound — trimethylsulphonium iodide with quadrivalent
sulphur, (CH3)2S.0H3l,
CH3>^ CH3>^'^I
The oxygen compounds are called oxonium salts on account of their analogy with
ammonium salts, say ammonium chloride, H4NOI, formed by the direct coupling
of ammonia, NH3, and hydrogen chloride, HOI. Instead of 0. Friedel's formula for
the hydrochloride, A. Wurtz suggested the alternative (0H3)20 : Cl"H, where the
oxygen is still quadrivalent ami the chlorine tervalent. 0. Friedel cited Rose's
quadrantoxides, Ag40 and OU4O, as further evidence of a quadrivalent oxygen.
In 1888, J. F. Heyes,i8 like H. Buff, in 1866, argued that one of the oxygen atoms
in the dioxides of barium, lead, manganese, etc., corresponds with quadrivalent
oxygen, and he attributed the ready polymerization of the aldehydes, cyanates,
920 , INORGANIC AND THEORETICAL CHEMISTRY
and metaphosphates to the presence of oxygen potentially quadrivalent. J. F.
Heycs also referred the formation of compounds with water of crystallization to
the same cause. A ,c;reat number of organic compounds have been discovered in
which the most satisfactory formula) are based on the quadrivalency of oxygen. ^^ In
1899, J. N. Collie and T. Tickle 20 showed that hydrochloric acid unites with
dimethylpyrone forming dimethylpyrone hydrochloride :
»=c<ch:S)>«+«^'-^«=^<ch:S)>^<ci
Dimethylpyrone. Dimethylpyrone liydrochioride.
This compound acts as an easily dissociated salt of a strong base and a weak acid ;
it contains a quadrivalent oxygen atom which must play a similar part to that of
the nitrogen atom in the salts of dimethylpyridone, and this quadrivalent oxygen
/-v_p^CH.C(CH3)^
^-^'^CH.C(CH3)>^-^
imparts a basic character to the salts in question. By analogy with ammonium,
phosphonium, sulphonium, and iodonium bases, he supposed these salts to be
derivatives of an hypothetical oxonium hydroxide, H3O.OH. A. von Baeyer
and V. Villiger (1901) ^^ further argued that the simple oxygen atom, in every form
in which it appears in organic chemistry, can form salts with the proper acids under
suitable conditions. While a few oxygen compounds give well-characterized salts
with the simpler acids, it is usually better and surer to work with the complex
acids — e.g. hydroferricyanic, hydroferrocyanic, hydrocobalticyanic, phosphotungstic,
chloroplatinic, or, above all, perchloric acid. In these salts there is usually little
room for doubting the interpretation of their constitution based upon the quadri-
valency of oxygen. The same influences which increase or diminish the basicity
of nitrogen similarly affect the basicity of oxygen, and also, but to a limited extent,
the basicity of sulphur :
NH3, base OH2, neutral SHg, acid
N(C2H5)3, strong base 0(C2H5)2, weak, base unknown S(C2H5)2, neutral
N(C2H5)4, OH, very strong base S(C2H5)30H, base
The entrance of a positive alkyl radicle group renders the oxygen of Heutral water
basic although this is not the case with negative groups like phenyl, CgHr, — e.g. in
triphenylamine, N(C6H5)3, the basic character of ammonia has virtually disappeared.
Oxygen is an amphoteric element in that it can form both acids and bases. The
base-forming qualities of oxygen are comparable with those of nitrogen, phosphorus,
sulphur, and iodine, for when united with certain base-forming organic groups, the
resulting compound has well-defined basic qualities. Oxygen has not so strong a
tendency as nitrogen to pass from a lower to a higher valency.
When the tendency of oxygen to act as a quadrivalent element was generally
recognized, and chemists ceased to be restricted to a bivalent oxygen, many com-
pounds previously represented by graphic formula with oxygen bivalent were
^e^lodelled on the assumption of the higher valency. For example, following
A. W. AVilliamson's suggestion that oxygen is quadrivalent in carbon monoxide,
F. Goldschmidt (1904) 22 used the formula R-CezO-OH in place of R-CO-OH
for the organic acids and esters ; and for ammonium sulphate, J. C. Cain (1904)
used :
^nPO : NH3 . , f ,, 1 QH ^O . NH4
*^^2^n • "Nrn ^^ place of the usual ^^o'^n NH
Ammonium hydroxide 23 is also regarded as a compound H3N : OHo rather than
the usual H4N.OH.
E. H. Archibald and D. Mcintosh (1904) 24 studied the compounds formed by
the liquid hydrogen halidea with ether or acetone, and assumed that the valency of
OZONE AND HYDROGEN PEROXIDE 921
oxygen increases as the temperature diminishes. While ethyl oxide, {Coiir,)^), forms
a compound (C2H5)20.HI with hydrogen iodide, HI, analogous with the oxonium
salts, where the oxygen is quadrivalent, they assume that oxygen is sexivalent
in C3H70H.2HBr, and dodecavalent in the compound C3H7OH.5HCI. There is here
no reason for assuming these abnormally high valencies for oxygen if the halogen be
taken tervalent. J. I. Kanonnikoff (1901) thought that the spectrometric constants
of certain organic compounds containing oxygen better agreed with sexivalent
oxygen than with either quadri- or bi-valent oxygen ; and F. Flavitzky 25 used
sexivalent oxygen atoms to explain the union of water of crystallization of a salt.
If the tendency of water to crj^stallize in union with salts as water of crystallization
is to be ascribed to the tendency of oxygen to pass from the bi- to the quadrivalent
condition, other substances, built on the water type with bivalent oxygen, might
be expected to act in a similar manner. Examples are common. Methyl alcohol,
CH3.OH, unites with calcium chloride as alcohol of crystallization to form
CaCL.^CHsOH. Similarly, the unsaturated character of the oxygen in water,
H2O, and of nitrogen in ammonia, NH3, is employed to explain how ammonia
and water are frequently interchangeable in chemical compounds; thus, CuSO^.SHoO
and CUSO4.5NH3, etc."
References.
1 C. F. Schonbein, Pogg. Ann., 50. 616, 1840 ; G. W. Osann, ib., 75. 386, 1818 ; A. P. Dubrun-
faut, Compt. Bend., 70. 159, 1870.
2 J. C. G. de Marignac and A. de la Rive, Compt. Rend., 20. 808, 1291, 1845 ; J. J. BerzeUus,
Pogg. Ann., 67. 142, 1845; C. F. Schonbein, Journ. prakt. CJiem., (1), 42. 383, 1847;
A. VV. Williamson, Journ. Chem. Sac, 2. 395, 1845 ; Liehig's Ann., 54. 127, 1845 ; 61. 13, 1847 ;
E. Fremy and E. Becquerel, Ann. Chim. Phys., (3), 35. 62, 1852; G. Baumert, Pogg. Ann., 89.
38, 1853 ; T. Andrews, Phil. Trans., 146. 1, 1856.
3 T. Andrews and P. G. Tait, Phil. Trans., 149. 113, 1860; Proc. Roy. Sac, 8. 498, 1857;
9. 006, 1859.
* A. Houzeau, Bnll. Soc. Chim., (2), 6. 340, 1866; Ann. Chim. Phys., (4), 22. 150, 1871;
ComjJt. Rend., 70. 1286, 1870 ; 74. 256, 1280, 1871 ; L. von Babo, Liehig's Ann. Suppl, 2. 265,
1863 ; J. L. Soret, Compt. Rend., 38. 445, 1854 ; 57. 604, 1863 ; C. F. Schonbein, Memoires sur
V ozone. Bale, 1849.
* W. Odiing, A Manual of Chemistry, London, 1861.
« J. Tyndall, Heat a Mode of Motion, London, 33, 1 863.
' J. L. Soret, Ann Chim. Phys., (4), 7. 113, 1866 ; (4), 13. 247, 1868 ; B. C. Brodie, Proc.
Roy. Soc, 20. 272, 1872 ; R. Wolilenstein, Pogg. Ann., 139. 320, 1870.
8 L. von Babo and C. E. Glaus, Liebig's Ann. Suppl, 2. 297, 1863.
» B. C. Brodie, Proc. Roy. Soc, 20. 472, 1872 ; Phil. Tran-^., 162. 435, 1872.
" A. Ladenburg, Ber., 31. 2830, 1898 ; 32. 221, 1899 ; 33. 2282, 1900 ; W. Staedel, ib., 81.
3143, 1898 ; M. Groger, ib., 32. 3174, 1899 : 0. Brunei?, ib., 33. 1832, 2999, 1900.
11 M. Otto, Compt. Rend., 124. 78, 1897; Ber., 34. 1118, 1901 ; A. Ladenburg, ib., 34. 631,
1834, 1901 ; J. L. Soret, Compt. Rend., 64. 104, 1867 ; Ann. Chim. Phys., (4), 13. 257, 1868 ;
A. Ladenburg, Ber., 31. 2508, 1898.
12 W. A. Shenstone and J. T. Gundall, Journ. Chem. Soc, 51. 625, 1887; G. S. Newth, ib.,
69. 1298, 1896.
13 M. Trail be, Ber., 26. 1876, 1893 ; A. Angeli, Atti Accad. Lincei, (5), 20. i, 625, 1911.
1* J. W. Briihl, Ber., 28. 2864, 1895.
li* A. Wolkowicz, Zeit. anorg. Chem., 5. 264, 1893.
1^ A. Naquet, Compt. Rend., 58 381, 675, 1864 ; H. BnfF, Gruvdlehrender theoretischen Chcmie,
Etrangen, 1866 ; A. W. Williamson, Journ. Chem. Soc, 22. 360, 1869 ; S. M. Jorgenscn, Zeil.
anorg. Chem., 7. 327, 1894 ; Journ. pralt. Chem., (2), 29. 419, 1884 : J. Thomsen, Ber., 6. 6, 433,
1873 ; L. Meyer, ib., 6. 101, 1873 ; W. A. Tilden, Introduction to the Study of Chemical Philosophy,
London, 1876 ; A. Wurtz, La theorie atomique, Paris, 1879.
17 C. Friedel, Bidl. Soc Chim., (2), 24. 166, 241, 1875; J. H. van't Hoff, Ansichten iib.r die
organische Chemie, Braunschweig, 1. 56, 1878 ; F. Zochini, Zeit. phys. Chem., 19. 431, 1896 ;
J. P. Kuenen, ib., 37. 485, 1901 ; F. Jiittner, ib., 38. 56, 1901.
18 J. F. Heyes, Phil. Mag., (5), 25. 221, 297, 1888.
i» R. Meldola, Phil. Mag., (5), 26. 403, 1888 ; E. Bamberger, Ber., 24. 1761, 1891 ; P. Walden,
ib., 34. 4185, 1901 ; 35. 1764, 1902; N. S. KurnakoflF, Journ. Russian Phys. Chem. Soc, 25.
726, 1893 ; J. L Kanonnikoff, ib., 32. 639, 1900 : 33. 61,95, 197, 1001 ; J. Schmidt, Ueber die
basische EigenscJuiflcn des Sauerstojfs und Kohlcnstoffs, Berlin, 1904 ; A. Rosenhein and W. Stell-
mann, Ber., 34. 3377, 1901 ; A. Werner, ib., 34. 3300, 1901.
20 J. N. CoUie and T. Tickle, Journ. Chem. Soc, 75. 710, 1899.
922 INORGANIC AND THEORETICAL CHEMISTRY
" A. von Baeycr and V. Villiger, JScr., 34. 2079, 3612, 1901.
«« F. Goklscbmidt, Zeit. Ekktrochem., 10. 221, 1904; J. C. Cain, Mem. Manchester Phil. Soc,
48. 14, 1904.
23 A. R. Hantzsch and W. B. Davidson, Ber., 31. 1616, 1898.
»* E. H. Archibald and D. Mcintosh, Journ. Chcm. Soc, 85. 919, 1904.
"* F. Flavitzky, Jour7t. Mu^siaji Phys. Chem. Soc, 23. 125, 1891 ; Journ. prakt. Chem., (2), 46.
67, 1892.
§ 8. The Modes o! Formation and Preparation o£ Hydrogen Peroxide
At first sight hydrogen peroxide, H2O2, is related to water much as ozone is related
to oxygen ; while the latter can be regarded as oxidized oxygen, so the former can
be regarded as oxidized water. Consequently, the term eau oxygenee, applied to
this compound by its discoverer L. J. Thenard in 1818, is singularly appropriate.
Just as ozone is obtained from oxygen by the expenditure of energy equivalent to
30 Cals. per gram-molecule of O3, so is hydrogen peroxide formed from water by
the expenditure of 21*5 Cals. per gram-molecule of H2O2. As in the case of ozone,
the various methods of preparing hydrogen peroxide may be classed as physical
or chemical — in the one case, energy is added in the form of heat, electricity, or ultra-
violet radiations ; and in the other case, energy is added indirectly or through the
mediation of a chemical reaction.
(1) Thefortmtion of hydrogen peroxide hy the action of heat. — Hydrogen peroxide
can be formed by passing a current of moist oxygen through a tube at about 2000°
and rapidly chilling the issuing
gases. In H. St. C. Deville's 1
hot and cold tube— -tube chaud
etfroid — method of conducting
the experiment, a narrow silver
or platinum tube is kept cool
by a current of cold water. This
tube is placed in the centre of a
porcelain tube. Fig. 12. A
current of gas is placed along
This arrangement is placed in a furnace
The products of decomposi-
Fia. 12.— Deville's Tube Chaud et Froid
the annular space between the two tubes.
so that the gas is heated to a very high temperature.
tion are suddenly chilled by the cold tube and partially prevented from recombining
as they are carried out of the hot zone. The products of many high temperature
reactions can thus be examined at ordinary temperatures. According to W. Nernst,
the formation of hydrogen peroxide cannot be observed by passing a mixture of
steam and oxygen through a hot platinum or iridium tube and cooling the products
rapidly, presumably because of the extremely rapid rate of decomposition which
W. Nernst found to be nearly the same as with ozone. The formation and decom-
position of hydrogen peroxide is a balanced reaction 2H20+02^2H202 ; and
W. Nernst (1905). estimates that the percentage amount of hydrogen peroxide
which can coexist in equilibrium with steam and oxygen under 0*1 atm. pressure,
at different temperatures, is as follows :
Temperature . . . 650° 867° 1220° 1881° 2511°
Per cent, of hydrogen peroxide . 0'00036 0*0032 0 028 0 24 0 66
Hence, the rate of cooling must be exceedingly fast if hydrogen peroxide formed at
a high temperature is to survive undecomposed. For example, W. Nernst prepared
hydrogen peroxide by spraying water on to a glowing Nernst's filament, F. Fischer
and 0. Ringe made it by passing steam, at 40 mm. of mercury pressure, through
a tube of fused magnesia nearly white hot. A large block tin condenser was placed
as near as possible to the hottest zone. The condensed liquid contained 0*0045 per
OZONE AND HYDROGEN PEROXIDE 923
cent, of hydrogen peroxide. The magnesia tube was 0'6 mm. diameter, and the
yield was very much reduced with tubes 0*3 mm. or I'O mm. diameter. F. Fischer
and 0. Ringe also obtained hydrogen peroxide by blowing steam from a sloping
quartz capillary tube into a flame of hydrogen 4 to 6 cm. in height. The jet of
steam, at 7 mm, pressure, was between 4 to 6 mm. from the flame, and so directed
that the products were driven into the neck of a tin condenser. Although the con-
densation was incomplete owing to the velocity of the gaseous stream, 0'067 per
cent, of hydrogen peroxide was found in the condensed liquid.
(2) The formation of hydrogen peroxide hy exposure to ultraviolet radiations. —
Water confined in a quartz vessel, permeable to ultraviolet rays, is decomposed by
exposure to ultraviolet radiant energy furnished by a mercury lamp ; hydrogen
and hydrogen peroxide are formed : 2H20=H202-hH2 ; and, according to K. V.
CharitschkofE,2 hydrogen peroxide can be detected in water, with a little oxygen
in solution, which has been exposed to bright sunlight at 9°-22° for 12 days ; at
19°-31°, after 8 days ; and at 20°-41°, after 7 days. If the oxygen be replaced by
air, no hydrogen peroxide was observed. According to A. Tian (1915), ^ ultraviolet
rays of short wave-length (A2500-3000) decompose solutions of hydrogen peroxide
at a measurable rate, but water is decomposed only by the rays in the extreme
ultraviolet (less than A1900). The water is decomposed in accord with the
reversible reaction : 2H20=H202+H2 ; the hydrogen peroxide is subsequently
decomposed in accord with the equation H202==^H20+0 ; and after the lapse
of sufficient time, the hydrogen and oxygen evolved by these two reactions are
in the proportions which would occur if the water alone were directly decomposed
into its elements. If the water contains dissolved oxygen it may unite with the
hydrogen evolved in the primary reaction and form hydrogen peroxide, some of
the dissolved oxygen forms ozone which in turn reacts with the hydrogen peroxide.
No reaction between water and ozone has been observed under the influence of
ultraviolet light. The conditions which favour the formation of hydrogen peroxide
are (i) the use of ultraviolet light rich in rays of very short wave-length : (ii) the
exposure of thin layers of water ; and (iii) eliminating the conditions which favour
the decomposition of hydrogen peroxide. Radioactive barium bromide produced a
negative result in darkness, but a positive result in light, hence K. V. CharitschkofE
assumes that it is not the radium emanation per se which produces the hydrogen
peroxide, but that the radium salt acts as a catalytic agent.
(3) The formation of hydrogen peroxide hy electrolysis. — The water which collects
about the anode or positive electrode during the electrolysis of water acidulated
with sulphuric acid possesses oxidizing properties which are usually attributed to
the presence of hydrogen peroxide. * The formation of hydrogen peroxide by
electrolysis is favoured by high acid concentration, low temperature, strong current,
and small electrode surface. Its formation is prevented if the acid be dilute and
the temperature exceeds 60°. According to M. Traube,^ hydrogen peroxide is
formed in small quantities at the cathode, not the anode, during the electrolysis
of aqueous solutions, provided oxygen is bubbled about the cathode from which
hydrogen is being evolved during the electrolysis of dilute sulphuric acid. If air
or oxygen be carefully excluded from the cathode, no hydrogen peroxide is formed.
According to M. Traubc, this shows that hydrogen peroxide can be regarded as an
intermediate product in the reduction of oxygen, where the end-product is water.
Indeed, M. Traube draws the extraordinary conclusion that hydrogen peroxide is
always a product of the reduction of ^nolecular oxygen, and is never produced hy the
oxidation of water. M. Traube (1887) explains the formation of hydrogen peroxide
by the electrolysis of water by assuming that the hydrogen liberated at the cathode
immediately unites with the oxygen of the air to form the compound in question,
and that this compound is reduced to water by the further action of hydrogen.
F. Richarz,6 on the contrary, shows that hydrogen peroxide is also formed at the
anode, and assumes with M. Berthelot that persulphuric acid is primarily formed
during the electrolysis of water acidulated with sulphuric acid, and hydrogen
924 INORGANIC AND THEOKETICAL CHEMISTRY
peroxide is formed by the secondary reaction between the water and the acid.
Persulphuric acid appears during the eh^ctrolysis of sulphuric acid more concentrated
than H2SO4.2H2O ; and when this acid reacts with water, a certain amount of
hydrogen peroxide is formed. According to M. Traube, if a plate of hydrogenized
palladium be used as anode, the oxygen developed is absorbed without forming
a trace of hydrogen peroxide.
(4) The formation of hydrogen peroxide hy the electric discharge. — Although
W. Nernst ^ failed to obtain hydrogen peroxide by passing a stream of electric
sparks through a mixture of steam and oxygen, F. Fischer and 0. Ringe found that
hydrogen peroxide is formed under these conditions provided the stream of gas
passes quickly enough to allow the escaping gases to be rapidly cooled. W. Nernst,
however, did fmd that if electric sparks are passed through liquid water, hydrogen
peroxide is formed, for the rate of cooling is then fast enough to prevent its complete
decomposition.
When a mixture of hydrogen and oxygen is exposed to a brush electrical discharge,
while the gas is passing through a U-tube cooled by licjuid air, a yield of about 2"5 per
cent, of hydrogen peroxide is obtained, but none is formed if the experiment be
conducted at ordinary temperatures. The mixed gases must be under reduced
pressure — say 3 cm. of mercury — to prevent explosion. At ordinary pressures
when the mixed gases contain less than about 5 per cent, or more than 95 per cent,
hydrogen, a yield of 87 per cent, has been obtained at the temperature of liquid
air ; and at 22°, a yield of 6'4 per cent, resulted. F. Fischer and 0. Ringe obtained
traces of hydrogen peroxide by subjecting steam to the brush discharge of an
ozonizer at a temperature of 130°, so as to prevent the condensation of water within
the instrument. Better results were obtained with a mixture of steam and oxygen
or air. Hydrogen peroxide is also formed when the Tesla brush discharge,^ or the
discharge in an ozonizer, is passed through a mixture of oxygen and water vapour
at pressures between 385 and 770 mm. ; no hydrogen peroxide was observed with
water vapour alone ; and nitrogen oxides are simultaneously produced if moist
air be used in place of moist oxygen. Hydrogen peroxide is also said to be formed
by blowing moist air against a spark or arc discharge — this may be purely a
temperature effect.
(5) The formation of hydrogen peroxide during oxidation processes. — Both
hydrogen peroxide and ozone have been detected in the flame of burning
hydrogen by 0. Loew (1870) and by W. Manchot (1909). The latter considers
that these substances do not take part in the processes of combustion, but are an
effect of the high temperature of the flame. Hydrogen peroxide is produced during
the explosion of hydrogen with an excess of oxygen,^ and it has been recognized
among the products of various types of combustion in air — e.g. hydrogen, carbon
monoxide, cyanogen, etc. 10 Since the temperature of the hydrogen flame exceeds
2000°, it is natural to expect that this flame w^ill contain appreciable amounts of
hydrogen peroxide ; and this was demonstrated by M. Traube, who allowed a jet
of burning hydrogen to impinge on the surface of cold water in which ice was floating
or on ice itself, and detected the hydrogen peroxide in the water. ^I. Traube thus
obtained an aqueous solution containing 0'74 per cent, of hydrogen peroxide.
This compound has also been reported in the water produced during the combustion
of hydrogen in air ; in the flames of alcohol, ether, coal gas, or carbon disulphide.
It might be asked why no trace of hydrogen peroxide was detected by F. Fischer
and F. Brahmer (1906) in the products obtained when hydrogen, etc., is burned
under liquid air {vide supra). It is supposed that the formation of ozone is preceded
by the dissociation 02^^=^0-1-0 ; that the speeds of formation of ozone and hydrogen
peroxide are approximately equal, and therefore in the presence of a large excess
of oxygen, more atoms of oxygen will be oxidized to ozone than water molecules to
hydrogen peroxide ; and further, that at the moment of cooling the ozone reacts
with the relatively small quantity of hydrogen peroxide- 03-hH202->H20-|-202.
Hydrogen peroxide is formed when ozonized oxygen or air is passed through water
OZONE AND HYDROGEN PEROXIDE 925
on the surface of which a little ether floats. If a little water is placed in a beaker
containing ether, and the latter is burnt by placing a spiral of hot platinum wire
just over the surface of the liquid, hydrogen peroxide can be detected in the water
after all the ether has burnt away. It is supposed that the ozone first produced
forms a peroxide with the ether and that this is decomposed by the water forming
ether and hydrogen peroxide. According to this view, the ether acts as a catalytic
agent. K. V. CharitschkofE and M. AmbardanofE i^ reported the formation of
hydrogen peroxide in water containing oxygen, during exposure to bright sunlight.12
According to A. Richardson, hydrogen peroxide is produced by exposing
urine to direct sunlight, and also by exposing ether, amyl alcohol, and certain
organic acids — e.g. oxalic acid — to sunlight. W. R. Dunstan and T. S. Dymond,
however, were unable to detect hydrogen peroxide in pure ether — dry or moistened
with water — ^after exposure to sunlight. Some specimens of less pure ether— e._^.
methylated ether— did develop hydrogen peroxide in light. Hydrogen peroxide
is also formed when turpentine or other oils containing terpenes are oxidized by
air or ozone in the presence of water. C. T. Kingzett (1878) claims to have made
a solution of hydrogen peroxide of 2-vol. strength in this way. Moist bone-black,
when exposed to air and light, rapidly forms appreciable quantities of hydrogen
peroxide.
Hydrogen peroxide is often formed when metals, and particularly their amalgams,
are slowly oxidized in the presence of water — e.g. by shaking zinc amalgam with
water — a better yield is said to be obtained if an alkaline earth be present. Hydrogen
peroxide, says M. Traube (1893) ,1^ is formed only when the oxidation is effected
by ordinary oxygen ; oxidizing agents, other than ozone, do not give rise to this
compound — e.g. hydrogen peroxide is obtained when finely divided zinc, magnesium,
aluminium are shaken with water, but not if all traces of free oxygen are excluded.
In 1859 E. von Gorup-Besanez 1* reported that an oxidizing substance, probably
hydrogen peroxide, is formed during the evaporation of water, but N. Smith could
detect this compound only when the evaporation took place in the presence of the
metal zinc. When zinc, copper, or lead is shaken up with air and dilute sulphuric
acid (1:55), the reactions symbolized: Zn+2H20+02=Zn(0H)2+H202 ; and
Zn(0H)2+H2S04 = ZnS04+2H20 occur. It will be observed that twice as
much oxygen is required for the oxidation process as is actually consumed in
oxidizing the zinc : Zn-[-02==Zn0+0 ; H204-0=H202. One half of the oxygen
is said to be used in the primary process and the other half in the secondary reaction.
The reaction is a concurrent or side reaction, but since half a molecule of oxygen is
used in each, the two concurrent reactions are not independent of one another.
This particular type of reaction is known as auto-Oxidation. There is a consider-
able difference of opinion as to the mechanism of auto-oxidation. The oxygen
used in the secondary reaction— formation of hydrogen peroxide — is said to be
" rendered active " by the primary reaction. The formation of ozone during the
oxidation of phosphorus is another example. W. R. Dunstan, H. A. D. Jowett,
and E. Goulding '^^ consider that hydrogen peroxide is an intermediate stage in the
rusting of iron.
In certain cases, if a substance undergoing slow oxidation at ordinary temperature
be mixed with another substance which is not oxidized when alone, both substances
are simultaneously oxidized. This phenomenon was noticed by C. F. Schonbein i^
in 1858. For example, (1) Ozone is formed during the oxidation of phosphorus ;
(ii) hydrogen peroxide is formed during the oxidation of zinc, lead, etc. ; (iii) indigo
blue is simultaneously oxidized to colourless isatin when benzaldehyde or turpentine
is oxidized ; sodium arsenite is likewise oxidized in the presence of oxidizing sodium
sulphite, etc.
That part of the oxygen which unites with the substance undergoing oxidation
is sometimes called bound oxygen, while the oxygon which is consumed in the
formation of ozone, hydrogen peroxide, is called active oxygen, and the oxygen
is said to be activaled or rendered active during the process of oxidation. C. Engler
926 INORGANIC AND THEORETICAL CHEMISTRY
calls the substance undergoing oxidation the autoxidizer, and the substance which
unites simultaneously with the active oxygen, the acceptor.
C. F. Schonbein still further demonstrated that just so much oxygen is rendered
active as is consumed by the oxidizing substance ; or, in all slow oxidations the same
amount of oxygen is required for the oxidation of the substance as is consumed in
the formation of hydrogen peroxide from water, ozone from oxygen, etc. The
hydrogen peroxide is generally decomposed into water and oxygen, so that an
exact proof of the above deduction can be obtained only under favourable conditions.
C. F. Schonbein obtained a confirmation of the hypothesis by the slow oxidation
of lead amalgam in the presence of dilute sulphuric acid. Almost the same amount
of oxygen was rendered active in the form of hydrogen peroxide, as is used in the
formation of lead sulphate. Similar results were obtained by M. Traube in the
oxidation of zinc in the presence of water, whereby zinc hydroxide and hydrogen
peroxide are formed in equi-molecular proportions. The reaction was symbolized :
„ , HOH , O „ .OH , HO
^^+HOH+0=^^<OH+Hb
C. F. Schonbein's observation has also been verified by J. H. van't HofE, W. P.
Jorissen, and by C. Engler and his co-workers.i7
(6) The formation of hydrogen peroxide in chemical reactions. — If a peroxide,
MO2, be treated with such an acid that the base radicle is precipitated as a salt
of the acid, hydrogen peroxide will remain in solution. For example, if concentrated
sulphuric acid be allowed to react with barium peroxide, ozonized oxygen is evolved ;
if the acid is of moderate concentration, ordinary oxygen gas is evolved :
2Ba02+2H2S04=2BaS04+2H20+02; while if dilute acid be used, hydrogen
peroxide is formed : Ba02+H2S04=BaS04+H202. Barium peroxide is the usual
starting point for the preparation of hydrogen peroxide ; for example, in one process :
Gradually add barium peroxide to ice-eold water through which a stream of carbon
dioxide is passing. The insoluble barium carbonate is precipitated, and a dilute aqueous
solution of hydrogen peroxide remains: Ba02 + C02+H20=BaC03 + H202. If an
excess of carbon dioxide be used, the yield of hydrogen peroxide is low and an insoluble
barium percarbonate, BaC04, is precipitated. ^^
In the modifications of this process, the barium peroxide is mixed with a little ice-
cold water and gradually added to ice-cold dilute hydrochloric,!^ sulphuric,20 hydro-
fluosilicic acid,2i hydrofluoric,22 or phosphoric acid.23 A barium salt — chloride,
sulphate, fluosilicate, or phosphate — and hydrogen peroxide are formed. In the
first case, the barium chloride is soluble. It can be removed by adding just
sufficient silver sulphate to precipitate insoluble barium sulphate and silver chloride :
BaCl2+Ag2S04=BaS04+2AgCl. This method is mainly of historical interest,
because L. J. Thenard employed a similar process when he discovered hydrogen
peroxide in 1818. The process with sulphuric acid, as employed by R. Wolffenstein,
is as follows :
Gradually add barium peroxide, suspended in a little water, to a mixture of equal
volumes of water and sulphuric acid (cooled by a freezing mixture of ice and salt) until
the solution is just barely acid. If too much barium peroxide has been added, a little more
sulphuric acid is needed. Keep the solution in a freezing mixtvire for about a day. Filter
off the insoluble matter, and evaporate the liquid on a water-bath, at about 70°, in a smooth
platinum or porcelain basin until signs of effervescence appear. This will occur when the
solution contains about 45 per cent, of hydrogen peroxide. Cool the solution quickly.
Concentrated solutions soon decompose if they are not kept cold.
By treating a cold aqueous solution of sodium peroxide with dilute and cold
hydrochloric acid, a solution of hydrogen peroxide in sodium chloride is obtained :
Na202+2HCl=2NaCl+H202 — hydrofluoric acid has been recommended in place
of hydrochloric acid ; and by treating potassium peroxide with tartaric acid in the
cold, an aqueous solution of hydrof^en peroxide contaminated with a little potassium
OZONE AND HYDROGEN PEROXIDE 927
tartrate is obtained. Most of the potassium tartrate separates from the cold solution.
Hydrofluosilicic acid and potassium peroxide has been recommended. 2*
A. von Baeyer and V. Villiger, G. Adolph and A. Pietzsch, and L. Lowenstein
found that if persulphates 25 be treated with dilute acids hydrogen peroxide is
obtained ; R. Wolffenstein and E. Merck used percarbonates ; and F. Jaubert
found that if a mixture of a perborate with an equivalent amount of a dry solid
organic or inorganic acid, or an acid salt be moistened with water, hydrogen peroxide
is formed. Hydrogen peroxide is prepared commercially from potassium persulphate
or persulphuric acid.
The concentration of solutions of hydrogen peroxide.— The concentration
of solutions of hydrogen peroxide has been effected in many ways. A 3 per
cent, solution of hydrogen peroxide, which is free from alkaline compounds,
from traces of salts of the heavy metals, and from suspended solids, can be con-
centrated on the water-bath at 75° to a 50*7 per cent, solution with a loss of about
36 per cent., and to a 66*6 per cent, solution with a loss of about 72 per cent. Further
concentration by evaporation is impracticable because of the decomposition. The
solution can also be further concentrated by evaporation over concentrated sulphuric
acid in vacuo, or rather under reduced pressure until one volume of the liquid
gives 475 volumes of oxygen gas.
By agitating the liquid with 10 to 12 times its volume of ether, decanting off
the ethereal liquid, and removing the ether by evaporation on a water-bath, a 73*8
per cent, solution can be obtained from a 48 per cent, aqueous solution,26 A. Houzeau
concentrated dilute solutions by freezing out the water ; and by this means
M. Hanriot obtained a residual liquid such that one volume of liquid gave 70 volumes
of oxygen.
The fractional distillation of hydrogen peroxide. — ^A solution of hydrogen
peroxide decomposes rapidly when heated to 100° — even if the solution be
dilute — -hence, for a long time, the concentration of an aqueous solution of
hydrogen peroxide by fractional distillation was thought, to be impracticable.
M. Hanriot (1885) concentrated the liquid by distillation under reduced pressure,
and obtained a liquid of such a concentration that one volume furnished 267 volumes
of oxygen. In 1894, R. Wolffenstein discovered that anhydrous hydrogen peroxide
can be readily distilled under reduced pressure without undue decomposition ; and
it also can be distilled at temperatures below 85° in a very rapid stream of an inert
gas. W. Spring observed a violent explosion in concentrating hydrogen peroxide
by vacuum distillation.
The aqueous sokition, containing about 45 per cent, of hydrogen peroxide, can be
distilled under reduced pressure in the following manner. Fit up the apparatus indicated in
Fig. 13. Transfer the solution to a round-bottomed, thick- walled litre flask A, fitted with
a receiver B, and a thermometer 1' passing through a one-hole rubber stopper. The receiver
B is placed over a funnel so that cold water can be sprayed on the receiver, and run off to
the sink through rubber tubing attached to the stem of the fmmel. The side neck of the
receiver is connected, by pressure tubing, with a manometer ; which in turn is connected
with a 3-way stop-cock E, a water trap G, and a filter pump D. The flask A is heated by
an oil bath F, and Bimsen's burner. When the manometer shows a pressure of about 15 mm.
and the thermometer a temperature between 35° and 40°, a dilute aqueous solution of
hydrogen peroxide in water distils into the receiver. The temperature rises gradually to
about 70°, when a very concentrated solution of hydrogen peroxide remains in the distilling
flask A. Further concentration is best effected by placing a beaker containing some of
the hydrogen peroxide solution in a mixture of solid carbon dioxide and ether. The whole
mass freezes. Drop a little of the frozen solid into a portion of the concentrated hydrogen
peroxide solution. At between —8° to —10° small needle-shaped crystals separate. Drain
away the mother liquid from the crystals ; melt the crystals and cool the mass, so that
another crop of crystals is obtained. Repeat the operations. The solution remaining in
the distilling flask will serve for most experiments where concentrated solutions of hydrogen
peroxide are required. If desired, it can be concentrated a little more by evaporation over
sulphuric acid under reduced pressure.
Concentrated hydrogen peroxide begins to attack the glass distilling flask at
INOKGANIC AND THEORETICAL CHEMISTRY
about 80°. R. Wolffenstein, by fractional distillation, obtained a solution of
90 per cent, hydrogen peroxide at 8r-85° at 68 mm. pressure ; and by repeatedly
T
oi
[
1
a]
UJ
Fig. 13. — Distillation of Hydrogen Peroxide under Reduced Pressure.
redistilling the product he got a liquid containing 99 per cent, of hydrogen peroxide
and boiling at 84°-85° C. at 68 mm. pressure.
References.
* H. St. C. Deville, Legons sur la dissociation, Paris, 307, 1804 ; H. W. Schroder, Pogg. Ann.,
129. 481, 1866 ; W. Nernst, Zeit. Elektrochem., 11. 710, 1905 ; Zeit. anorg. Chem., 45. 126, 1905 ;
F. Fischer and 0. Ringe, Ber., 41. 945, 1908.
2 K. V. CharitschkofF, Journ. Russian Phys. Chem. Soc, 42. 900, 1910 ; W. R. Dunstan and
T. S. Dymond, Journ. Chem. Soc, 57. 574, 1890.
' A. Tian, Transjorm/itions et iquilihre chimiques de Veau et des solutions de peroxyde d'hydroglne
a la lumiere uUra-violette, Marseille, 1915 ; E. Warburg, Ber. deut. phys. Ges., 17. 194, 1915.
* H. Meidinger, Liebig's Ann., 88. 57, 1853 ; R. Bunsen, Pogg. Ann.. 91. 621, 1854 ; C. F.
Schonbein, ib., 65. 161, 1845 ; C. Hoffmann, ib., 132. 607, 1867 ; A. Rundspaden, ib.. 151. 306,
1874 ; M. le Blanc, Compt. Bend., 75. 170, 1872.
6 M. Traube, Ber., 15. 2434, 1882 ; Sitzber. Akad. Berlin, 1041, 1887.
« M. Berthelot, Compt. Rend., 86. 71, 1878 ; F. Richarz, Wied. Ann., 31. 912, 1887.
' W. Nernst, Zeit. Elektrochem., 11. 710, 1905 ; F. Fischer and O. Ringe, Ber., 41. 945, 1908.
8 W. Nernst, Zeit. Elektrochem., 11. 710, 1905 ; A. Findlay, ib., 12. 129, 1906.
» A. SchuUer, Wied. Ann., 15. 289, 1882 ; A. R. Leeds, Journ. Ainer. Chem.. Soc, 6. 3, 1885 ;
K. Finchk, Zeit. anorg. Chem., 45. 118, 1905.
i» H. Struve, Bull. Acad. St. Petersburg, 15. 325, 1870; M. Traube, Ber., 18. 1890, 1885;
C. Engler, ib., 33. 1109, 1900 ; H. B, Dixon, Journ. Chem. Soc, 49. 94, 1886 ; L. I. de N. Ilosva,
Bull. Soc. Chim., (3), 2. 360, 1889 ; A. Bach, Compt. Rend., 124. 951, 1897.
11 K. V. Charitschkoff and M. Ambardanoflf, Journ. Russian Phys. Chem.. Soc, 42. 904, 1910.
12 A. Richardson, Journ. Chem. Soc, 59. 51, 1891 ; 63. 1110, 1893 ; 65. 450, 1894 ; A. Richard-
son and K. C. Fortey, ib., 69. 1.349, 1896 ; W. R. Dunstan and T. S. Dymond, ib., 57. 574, 1890;
C. T. Kingzett, Chem. Neivs, 38. 224, 1878.
i» M. Traube, Ber., 26. 1471, 1893.
1* E. von Gorup-Besanez, Liebigs Ann., 111. 232, 1859; N. Smith, Journ Chem. Soc, 89.
481, 1906.
15 W. R. Dunstan, H. A. D. Jowett, and E. Moulding, Journ. Chem. Soc, 87. 1548, 1905.
i« C. F. Schonbein, Jour7i. prakt. Chem., 75. 99, 1858; 77. 137, 1859 ; 78. 6<), 1859 ; 79. 87,
1860 ; 93. 25, 1864 ; 105. 226, 1868.
1' J. H. van't Hoff, Zeit. phys. Chem., 16. 411, 1895; W. P. Jorissen, Zeit. phys. Chem., 23.
667, 1897 ; Ber., 29. 1951. 1896 ; 30. 1051, 1897 ; C. Engler and W. Wild, Ber., 30. 1669, 1897 ;
33. 1109, 1900 ; C. Engler and J. Weissberg, Ber., 31. 3046, 3055, 1898 ; 33. 1090, 1097, 1900 ;
Krilische Studien iiber die Autoxydaiionsvorgange, Braunschweig, 1903 ; C. Engler and W. Frank-
enstein, Ber., 34. 2933, 1901 ; C. Engler, ib., 30. 2358, 1897; 36. 2642, 1903; C. Engler and
T. Ginsberg, ib., 36. 2645, 1903 ; M. Traube, ih., 26. 1471, 1893.
18 F. Duprey, Cornjit. Rend., 55. 736, 1862: A. J. Balard, ib., 55. 758, 1862 ; G. Lunge, Zeit.
angew. Chem., 4. 3, 1890; L. Mond, Ber., 16. 980, 1883; E. Merck, Cerman Pat., D.R.P.
152173, 1903.
OZONE AND HYDROGEN PEROXIDE 929
^» L. J. Thenard, Ann. Chim. Phys., (2), 8. 306, 1818; (2), 9. 51, 94, 314,441, 1818 ; (2), 10.
1 14, 335, 1819 ; (2), 11. 85, 208, 1819 ; (2), 50, 80, 1832. '
20 A. Gawalowsky, Apoth. Ztg., 4. 530, 1889 ; J. Thomsen, Ber., 7. 73, 1874.
21 0. Linder, Monit. Scient., (3), 15. 818, 1885; A. Bourgougnon, Joum. Amer. Cham.
Soc, 12. 64, 1890.
2 2 J, Pelouze, J. J. Berzeliits'' Lehrbuch der Chemie, Dresden, 1. 411, 1835; M. Hanriot, Compt.
Rend., 100. 57, 172, 1885 ; Bull. Soc. Chim., (2), 43. 468, 1885.
23 T. Mann, Monit. Sctent, (4), 2. 1455, 1888 ; O. Linder, ib., (3), 5. 818, 1876; G. E. Davis,
Chem. News, 39. 221, 1879 ; A. H. Mason, Pharm. Joum., (3), 11. 704, 1880.
2^ H. Osann, Chem. Centr., 97, 1862; C. Hoffmann, Liehig's Ann., 136. 188, 1865; R. de
Forcrand, Compt. Rend., 129. 1246, 1899 ; P. L. HuUn, German Pat., D.R.P. 132090, 1901.
25 A. von Baeyer and V. Villiger, Ber., 34. 856, 1001 ; G. Adolph and A. Pietzsch, German
Pat., D.R.P. 241702, 1910; 233856, 256148, 1911; L. Lowenstein, ih., 249893, 1910; R.
Wolffenstein, Ber., 41. 278, 1908; E. Merck., German Pat., D.R.P. 179771, 179826, 1904;
F. Jaubert, Compt. Rend., 139. 796, 1905.
26 W. Nagel, Pharm. Ztg., 43. 536, 1898 ; P. Schiloff, Joum. Rusman Phys. Chem. Soc, 25. 3,
1893 ; A. Houzeau, Compt. Rend., 66. 314, 1868 ; M. Hanriot, Compt. Rend., 100. 57, 172, 1885 ;
R. Wolffenstein, Ber., 27. 3307, 1894; 28. 2265, 1895; 34. 2430, 1901; J. W. Briihl, ib., 28.
2847, 1895 ; 30. 162, 1897 ; 33. 1709, 1900 ; W. Spring, Zeit. anorg. Chem., 8. 424, 1895 ; Bull.
Acad. Roy. Belgique, (3), 29. 363, 1895; H. R. Talbot and H. R. Moody, Joum. Anal. Chem.,
6. 650, 1893.
§ 9- The Physical Properties of Hydrogen Peroxide
Hydrogen peroxide is a viscid transparent liquid which does not wet solids so
easily as water. In thick layers, it has a blue colour ; and a metre long column
has the same tint as a 1"8 metre long column of water. There is always a greenish
tinge along with the blue, and W. Spring i attributes this to the presence of minute
bubbles of oxygen. The Specific gravity is 1'453, according to L. J. Thenard ;
1-4996, according to W. Spring; and, according to J. W. Briihl,2 1-4584: at 0°,
1-4375 at 2°, and 14378 at 20°, referred to water at 4°. J. W. Bruhl says that the
specific gravity is das scharfste Kriterium of the purity and dryness since the smallest
amount of water reduces the specific gravity in a marked degree, thus, the presence
of 0-52 per cent, of water lowers the value of this constant to 1'4094 (0°) ; and a
45-9 per cent, solution, according to H. T. Calvert, has a specific gravity 1'144.
I. Traube has investigated the molecular volume of hydrogen peroxide. The
liquid has no smell ; the vapour has a smell recalling that of nitric acid. Dilute
solutions have a bitter metallic taste. When a drop of the liquid peroxide comes
in contact with the skin, it forms a white blister. According to W. Spring, the
surface tension of the liquid of specific gravity 1-4996 is 3-5374 (10°), when water
at the same temperature has the value 7*750.
According to J. L. Thenard, the liquid does not freeze at 30°, but it freezes in a
mixture of ether and solid carbon dioxide, and if a trace of the solid is put into the
liquid cooled to —8° to — 10°, transparent needle-like prisms are formed ; the re-
crystallized peroxide has the melting point —2° ; hydrogen peroxide evaporates
slowly at ordinary temperatures and pressures, and he has studied the volatility of
aqueous solutions of hydrogen peroxide at different temperatures. The boiling
point of the anhydrous liquid at different pressures is
21
29
33
38
44
47
68 mm.
iling point
. 62-8°
69-7''
73-7°
76-7° ,
79.70
80-2°
84-85°
H. C. Jones, J. Barnes, and E. P. Hyde ^ have measured the lowering of the freezing
points of solutions of various acids and salts in hydrogen peroxide and found this
solvent to have a greater ionizing power than water, vide infra. The specific
heat of hydrogen peroxide calculated * by W. Spring from the specific heat of
the 34-25 per cent, solution is 0-6208 — this number is to be regarded as an upper
limit since other numbers are obtained with more dilute solutions,
VOL. I. 3 o
930 INORGANIC AND THEORETICAL CHEMISTRY
M. Berthelot, J. Thomsen, and R. de Forcrand ^ have published thermocheinical
data for this compound, from which it follows that the heat of formation
HgOiiq +Ogas=H202iiq -22-16 Cals. ; and H2gas+02gas=H202iiq.+46-84 Cals.
R. de Forcrand found that freshly distilled solutions of hydrogen peroxide with the
composition and formula indicated below have the heats of solution :
Percent H,0, . • • 35-31 4266 5516 65-38 85-93
H.O,+wHAwhenn= . 3-46 253 1-54 1-00 03
H^ate of solution . . . 0-071 0-093 0-099 0-310 0403 Cal.
By Bjctrapolation the heat of solution of the anhydrous peroxide : H202iiq.+Aq.
=H202soiution+0"46 Cal. The solution is distinctly acid ; the heat of neutralization
is 270 Cals.
The oxidation potential of the oxygen electrode is diminished by the addition
of hydrogen peroxide. Attempts to measure the electrode potential by W. Nernst,
F. Haber and S. Grinberg give material for calculating a value for the free energy
of hydrogen peroxide, but, as with ozone, G. N. Lewis and M. Randall believe
that there is no satisfactory evidence that the electromotive force is due to a single
definite reversible reaction, and they calculated a value for the free energy from
the dissociation pressure of barium dioxide and the equilibrium between hydrated
barium dioxide and water. The free energy of the reaction H2an.+02aq.=H202aq.
at 298° K. is —30,970 cals. for the formation of the HO2' ion on the assumption that
hydrogen peroxide in solution is an acid which ionizes H202=H*H-H02'; G. N.
Lewis and M. Randall give JH2+02=H02' ; for the free energy of the reaction
H202aq.=H20iiq.+J02, which mcasurcs the tendency of aqueous hydrogen peroxide
to decompose, they give at 298° K. —25,650 cals. W. Nernst from e.ni.f. measure-
ments obtained —17,100 cals., a difference of 8500 cals. For the heat of vaporiza-
tion H202iiq.=H202gas> Calculated from J. W. Briihl's and R. WolfEenstein's
measurements of the vapour pressure of pure hydrogen peroxide, G. N. Lewis
and M. Randall find 12,300 cals. per gram-molecule, or the free energy at 298° K. is
3500 cals. The same calculation makes the vapour pressure of hydrogen peroxide
at 25° to be 21 mm. ; the boiling point, 144° ; and Trouton's constant, 29 "5. This
high value shows that hydrogen peroxide is a very abnormal liquid. If F. M. Raoult's
laws of ideal solutions were followed by aqueous hydrogen peroxide, the vapour
pressure of hydrogen peroxide would be 2 '1/56*5 =0*037 mm., but W. Nernst's
distillation experiments give about one-fourth or one-fifth the value corresponding
with F. M. Raoult's law. If this value be approximately 0 01 mm., the free energy
of the reaction H20^aq =H202gas at 298° K. is RT log (760/0-01) =6700 cals. ; for
H2+02=H202iiq. at 298° K., —27,770 cals. ; and for H2+02=H202gas, at 298° K.,
—24,270 cals. From the heat of vaporization, and J. Thomsen's and R. de Forcrand's
thermochemical data, the heat of the reaction at 291° K. is —32,600 cals.
G. N. Lewis and M. Randall assume that the thermal capacity of hydrogen
peroxide may, as a first approximation, be taken as being the same as that of the
tetratomic gas NH3, and accordingly Cj,=7'5+0'0042r ; this in conjunction with
Cp=6'5+0-0009T for hydrogen, and C^=6-5H-0-0010T for oxygen, gives the free
energy of hydrogen peroxide, H2-f 02=H202gas as -31,100+5*5^ log T— 0*00115^2
— S'OSr, when the integration constant is evaluated from the free energy at 291° K.
From this equation it would appear to be impossible to obtain spontaneously
appreciable amounts of hydrogen peroxide from the two elemental constituents at
a temperature below 1000°. The free energy of the reaction H20gas+i02— H202ga8
was also found to be 26,310+4-56T log J— 0*0028T2-j-0-00000027T3_ll-80J.
From this equation, the pressure of hydrogen peroxide in equilibrium with water
vapour at atmospheric pressure is 10~7 atm. at 2000° K., and 3xl0~^ atm. at
3000° K. The experiment of M. Traube in which a considerable yield of H2O2 is
obtained by the rapid cooling of an oxy hydrogen flame cannot, therefore, be explained
by assuming that H2O2 is largely present in the gases in the hottest portion of the
oxy hydrogen flame. It must be explained rather by assuming that in the colder
OZONE AND HYDROGEN PEROXIDE 931
parts of the flame, probably between 500° and 1000° C, hydrogen and oxygen
combine directly to form hydrogen peroxide.
Thermochemistry of hydrogen peroxide. — While the formation of a gram-
molecule of liquid water from hydrogen and oxygen is attended by the evolution
of 684 Cals., the oxidation of water to hydrogen peroxide in aqueous solution would
be attended by an absorption of 23*1 Cals., and consequently, it is not to Jbe expected
that this compound will be formed during the combustion of hydrogen in oxygen —
except possibly by arresting the reaction under special conditions on the assumption
that hydrogen and oxygen first formed hydrogen peroxide as a transient intermediate
product, before forming water. To prepare hydrogen peroxide an indirect process
is employed. Oxygen unites with barium oxide with the evolution of 5*9 Cals.
per molecule of barium peroxide, Ba02, ^^^ ^his compound in turn is decomposed
by hydrochloric acid with the evolution of 11*0 Cals. Again, in the remarkable
reaction whereby hydrogen peroxide reacts with silver oxide forming water, silver,
and oxygen, H202-|-A.g20->2Ag+H20+02, one school of chemists says that the
attraction of oxygen atoms for one another in the two different compounds upsets
the unstable silver oxide and hydrogen peroxide ; but since the heat of formation
of silver oxide is 3'5 Cals., and of hydrogen peroxide, 11 '2 Cals., the reaction will
evolve 7 '7 Cals., an amount sufficient to account for the reaction apart from any
alleged molecular attraction. Still further, the powerful oxidizing effects produced
by hydrogen peroxide have been attributed to the effects of nascent oxygen in
the atomic condition. There is no particular need for this assumption because,
when hydrogen peroxide decomposes into water and oxygen : 2H202=2H20+02
+44*8 Cals., the heat evolved by the reaction, if confined to the products of the
reaction, would suffice to raise the temperature nearly 1000°, and this amply
suffices to explain the marked oxidizing properties of hydrogen peroxide over those
of oxygen. The superior oxidizing properties of ozone can be explained in a
similar manner.
The index of refractions (20-4°) is 1-40379 for the Li-line; 1-40421 for the
Ha-line ; 1'40624 for the Na-line ; 1-41100 for the H^-line ; and 1-41494 for the
Hy-line. According to J. W. Briihl, the specific refraction by Lorentz and Lorenz's
formula is 6-1708 for the Ha-line and 0-1742 for the Hy-line ; the specific dispersion
is therefore 0-0039. According to P. Drude and H. T. Calvert, the dielectric
constant of the anhydrous peroxide is 92-8 when the value for water is 81. This
high value for the dielectric constant indicates that hydrogen peroxide is a very
abnormal liquid. J. Dewar and J. A. Fleming investigated the effect of temperature
on the dielectric constant. H. T. Calvert observed that hydrogen peroxide did
not present the so-called anomalous electrical absorption.
L. J. Thenard found that when hydrogen peroxide is electrolyzed, it decomposes
like water with hydrogen at the negative and oxygen at the positive pole, but the
proportion of oxygen is much greater than with water. M. Berthelot, E. Schone,
M. Hanriot, and S. Tanatar found that with feeble currents and concentrated
solutions, the hydrogen peroxide is reduced by the hydrogen at the cathode, and no
gas is there given off, while oxygen is given off at the anode : 2H2O2— 2H2O+O2.
The specific electrical conductivity of a 4-5 per cent, solution is 2-89x108. H. C.
Jones, J. Barnes, and E. P. Hyde ^ could not determine the electrical conductivity
of aqueous solutions because they could find no satisfactory material for the
electrodes which would resist chemical action. H. C. Jones and his co-workers
found that the salts, potassium chloride, sodium nitrate, potassium nitrate, etc.,
lower the freezing point of solutions of hydrogen peroxide less than they do water,
possibly because of the formation of compounds analogous with KF.H2O2,
Na2SO4.9H2O.H2O2, etc., isolated by S. Tanatar ; while acids, like sulphuric, oxalic,
or acetic acid, lower the freezing point of solutions of hydrogen peroxide more than
they do water, presumably because hydrogen peroxide has a stronger ionizing power
than water. The action of bases on hydrogen peroxide, studied by G. Bredig and
H. T. Calvert, is analogous with the action of an acid on a base, for peroxides are
932 INORGANIC AND THEORETICAL CHEMISTRY
formed : 2NaOH+3H202=Na204+4H20, and it therefore follows that hydrogen
peroxide has acid properties ; indeed, G. Carrara and A. Bringhenti believe that it
is a monobasic acid which ionizes in solution H202=f^H*4-H02'.
Solubility. — Hydrogen peroxide dissolves in water in all proportions. Accord-
ing to J. W. Briihl, it is insoluble in petroleum ether, ^ and exerts no chemical action
on that menstruum. When aqueous solutions are shaken out repeatedly with ether
much of the hydrogen peroxide can be extracted. An ethereal solution of hydrogen
peroxide is more stable than the aqueous solution. The strength of aqueous solutions
is represented commercially by the number of volumes of oxygen which 100 c.c. of
the solution will furnish on decomposition. Thus 100 c.c. of a real \0-volume
solution should give 10 times its own volume of oxygen when decomposed ; as a
matter of fact, a 3 per cent, solution of hydrogen peroxide by weight is very nearly
a 10-volume solution ; a 5 per cent, hydrogen peroxide solution is nearly 20-volume
strength, and so on. The most concentrated solution on the market is called
perhydrol, and it contains about 30 per cent, of hydrogen peroxide corresponding
with a concentration of 100 volumes. There is some ambiguity in this mode of
expressing the concentration of the solutions, because if the solution be decomposed
by potassium permanganate, 2KMn04+5H202+3H2S04=502+2MnS04+K2S04
-J-8H20, half the oxygen comes 'from the permanganate, and in that sense, a
10-volume solution would furnish but five volumes of oxygen derived from the
peroxide, and five from the permanganate. In addition to perhydrol, there is also
on the market a 3 per cent, solution intended for general purposes, and a purer
3 per cent, solution for medicinal purposes. Besides, there is the so-called solid
hydrogen peroxide — known in commerce as hyperol — a white crystalline product
containing equimolecular parts of urea and hydrogen peroxide, which is made ^ by
dissolving urea in perhydrol ; it furnishes a solution of hydrogen peroxide when
treated with citric or other acids.
The coefficient of distribution between water and ether at 17*5° is, according
to K. Osikoff and S. Popoff,
Volume of ether : volume of water 0"5 1 2 5 7 9 10
Con. in ether : Cone, in water . 0-0575 0-0596 0-050 0-060 0-074 0-070 0072
The presence of sodium chloride has no influence on the coefficient, but sodium and
potassium carbonates lower the solubility of the peroxide in ether.io Hydrogen
peroxide is soluble in many organic solvents. ^ The partition coefficient (concentra-
tion in solvent: concentration in water) for ethyl acetate is 1 : 215; nitrobenzene, 1 : 200 ;
acetophenone, 7:4; amyl acetate, 1:8; ethyl isovalerianate, 1 : 40 ; isoamyl pro-
pionate, 1 : 12 ; chloroform, 1 : 600 ; benzene, 1 : 200 ; isobutyl alcohol, 1:3;
propyl formate^ 1:8; isobutyl butyrate, 1 : 50 ; propyl butyrate, 1 : 30 ; phenol (25°),
1 : 45 ; aniline (25°), 1:4; and quinoline (25°), 1 : 0*276. Quinoline is thus a very
good solvent for hydrogen peroxide, and when equal volumes of quinoline and
an aqueous solution are agitated together, the quinoline layer contains more than
half the peroxide.
References.
1 W. Spring, Zeit. anorg. Chcm., 8. 424, 1895.
2 J. W. Bruhl, Ber., 28. 2855, 1895 ; H. T. Calvert, Ann. Physik, (4), 1. 483, 1900 ; W. Staedel,
Zeit. angew. Chem., 15. 642, 1902 ; J. Traube, Ber., 40. 138, 1907.
' R. Wolfifenstein, Ber., 27. 3311, 1894; M. Traube, ib., 22. 1528, 1889; J. W. Bruhl, ih.,
28. 2853, 1895; H. C. Jones, J. Barnes, and E. P. Hyde, Amer. Chem. Journ., 27. 22, 1902;
H. C. Jones and C. Murray, ib., 30. 205, 1903 ; H. C. Jones and G. Carroll, ib., 28. 284, 1902.
* W. Spring, Zeit. anorg. Chem., 9. 205, 1895.
6 M. Berthelot, Compt. Rend., 90. 331, 897, 1880 ; R. de Forerand, ib., 130. 1250, 1620, 1900
J. Thomsen, Thermochemische Untprsuchnn^en, Leipzig, 1888; Pogg. Ann., 151. 194, 1874
W. Nemst, Zeit. phys. Chem., 46. 720, 1903"; F. Haber, Zeit. Elektrochem., 7. 441, 1043, 1901
F. Haber and S. Crinberg, Zeit. anorg. Chem., 18. 37, 1898 ; G. N. Lewis and M. Randall, Journ,
Amer. Chem. 8oc., 36. 1986, 1914 ; F. Kuspert, Naiur vnd Schule, 171, 1903.
OZONE AND HYDROGEN PEROXIDE 933
Hr « J. W. Briihl, Ber., 28. 2859, 1895 ; G. Carrara, AlH Accad. Lincei, (5), 1. 19, 1892 ;
^KW. Clayton, Trans. Faraday Soc, 11. 164, 1916 ; H. T. Calvert, Ann. Physik, (4), 1. 483, 1900 ;
^■f P. Drude, ib., (4), 1. 483, 1900 ; Zeit. phys. Chem., 23. 267, 1897 ; J. Dewar and J. A. Fleming,
^m Proc. Roy. Soc, 62. 250, 1897.
■l ' H. C. Jones and C. G. Carroll, Amer. Chem. Journ., 28. 284, 1902; H. C. Jones and
K G. Murray, ib., 30. 205, 1903 ; H. C. Jones, J. Barnes, and E. P. Hyde, ib., 27. 22, 1902 ; S. Tanatar,
m Z^-it- <^'^org. Chem., 28. 255, 1901 ; G. Bredig and H. T. Calvert, Zeit. Elektrochem., 7. 622, 1901 ;
m^Zeit. phys. Chem., 38. 513, 1901 ; G. Carrara and A. Bringhenti, Gazz. Chim. Hal, 33. 362, 1903 ;
^KM. Hanriot, Compt. Bend., 100. 172, 1885 ; M. Berthelot, ib., 95. 8, 1882 ; E. Schone, Liebig's
^m.Ann., 197. 137, 1897 ; S. Tanatar, Ber., 36. 199, 1903.
■r 8 R. Bottger, Journ. praJcL Chem., (1), 80. 58, 1859 ; J. W. Briihl, Ber., 28. 2855, 1895 ;
f K. Osikoff and S. Popoflf, Journ. Russian Phys. Chem. Soc, 35. 637, 1903.
' S. Tanatar, Journ. Russian Phys. Chem. Soc, 376. 40, 1908.
1" L. Crismer, Bull. Soc Chim., (3), 6. 24, 1893.
11 T. H. Walton and H. A. Lewis, Journ Amer. Chem. Soc, 38. 633, 1956, 1916.
§ 10. Quantitative Application of the Law of Mass Action
Chemical phenomena must be treated as if they were problems in mechanics. —
L. Meyer (1868).
I. Kant ^ has said that in every department of physical science there is only so
much science as there is mathematics ; and as our knowledge of natural phenomena
grows more clear and precise, so does it become more and more possible to employ
mathematical methods. Owing to the absence of all mathematical treatment in
chemical phenomena in his time, I. Kant denied to chemistry the name of science.
The most simple type of chemical reaction is one in which individual molecules
are involved in the change ; more complex reactions are concerned with the mutual
action of two or more molecules. For example, in the decomposition of nickel
carbonyl, Ni(CO)4->Ni+4CO, the individual molecules of nickel carbonyl are
independently concerned in the change — this type of reaction is called a unimolecular
reaction; with a reaction of the type, H20+C0Cl2->2HCl+C02, the mutual
action of two molecules is necessary for the reaction, and this is accordingly called
a bimolecular reaction ; and in the formation of ozone, 3O2— >203, the mutual action
of three molecules of oxygen is necessary and this is accordingly called a termolecular
reaction. The back reaction, in the preceding bimolecular reaction, is CO2+2HCI
->C0Cl2+H20, which is a termolecular reaction. The terms uni-, bi-, ter-, and
multi-molecular, or what is equivalent, mono-, di-, tri-, poly-molecular reaction,
were introduced by J. H. van't Hoff 2 to 'indicate the number of molecules con-
cerned in a reaction. Eeactions involving more than two molecules are not very
common. This is easily understood if we assume that bimolecular reactions are
caused by the collision of two molecules, termolecular reactions by the simultaneous
collision of three molecules, etc. The probability of a simultaneous collision between
three molecules is very much less than between two molecules, and the greater
the number of molecules taking part in a given transformation, the more likely is
the reaction to proceed by some other path than by the simultaneous collision of
a large number of reacting molecules.
The decomposition of hydrogen peroxide in light. — A solution of hydrogen
peroxide decomposes when it is exposed in a quartz vessel to the rays of light
from a mercury lamp. The decomposition ceases when the light is extinguished.
If the amounts of hydrogen peroxide in the solution exposed for various periods of
time be determined, the rate of decomposition can be calculated. It is found that
if a represents the initial concentration of the solution expressed in gram-molecules
per litre, and x the amount decomposed at the time t, the solution will then
contain a—x gram-molecules of the compound in question. Let L. Wilhelmy's
hypothesis, op. cit., be now tested. The velocity of the reaction at any time t
must be equal to k{a—x). If the symbol dx be employed to denote the amount
934 INORGANIC AND THEORETICAL CHEMISTRY
of peroxide decomposed in the minute interval of time dt, the velocity of the
reaction, the amount of substance decomposed in unit time, at the moment t, will
be represented by
— =Jc(a~-x) : .'. 7 log =k . . . (1)
dt ^ ' * , t ^ a—x ^ '
The passage from the equation on the left to that on the right involves a
very simple mathematical operation. The expression a-^t{a—x) measured at
different intervals of time must be a constant, h, if the reaction progresses so that
only otie moleaale of hydrogen peroxide is concerned in the process H202->H20+0.
Selecting a few measurements by J. H. Mathews and H. A. Curtis (1914:),3 we get
Time (t) .... 0 100 160 220 310 432
H.Oj per cent, (x^ . . . 1-58 106 083 0-63 044 0-26
k . . . ' . . . 00040 00041 00042 00041 • — 00042
The values of k are computed by the substitution of a=l'58, and the corresponding
values of x and t in the second of the above equations. The constancy of the
different values of k is quite consistent with the hypothesis. However, suppose
that the decomposition were to be represented by the usual equation, 2H2O2
->2H20+02, implying that two 7nolecules of hydrogen peroxide mutually react pro-
ducing water and oxygen molecules. Then the velocity of the reaction must be
represented by
^=k^-x){a-x>; .-.l^^^h . . .(2)
With the same data as before, the values of ki are no longer even approximately
constant. Hence, it is inferred that the decomposition of hydrogen peroxide in
light is a unimolecular reaction, H202->H204-0, and not really a bimolecular
reaction, 2H202->2H20+02, even though the last-named equation is conventionally
used to represent the process in order that attention may be focussed on the initial
and end products of the reaction. The unimolecular reaction is slow enough to be
readily measured. The atoms of oxygen from two different molecules of hydrogen
peroxide unite to form molecular oxygen, 0+0=02, ^^^ too quickly to influence
the measurement of the unimolecular change. This may be illustrated ^ by the
following analogy :
The time occupied in the transmission of a telegraphic message depends both on the
rate of transmission along the conducting wire, and on the rate of progress of the messenger
who delivers the telegram ; but it is obviously this last slower rate that is of really practical
importance in determining the time of transmission.
Hence the following rule : If a chemical reaction takes place in two stages, one of
which is considerably faster than the other, the observed order of the whole reaction will
be determined by the order of slower change.
The decomposition oi hydrogen peroxide in contact with platinum. — It has
been found by G. Bredig and M. von Berneck (1900) ^ that while the catalytic
decomposition of dilute solutions of hydrogen peroxide— say below -—th gram-
molecule per litre — by colloidal platinum is undoubtedly unimolecular, H2O2
-»H204-0, more concentrated solutions — say above the \ gram-molecule per litre —
decompose bimolecularly, 2H202->2H20+02 ; and that with intermediate con-
centrations, both types of reaction prevail. For instance, with a concentration
of 00034 gram-molecule of hydrogen peroxide per litre, G. Dyer and A. B. Dale
(1913) 6 find the following values of the constant k :
Unimolecular reaction . . 0014 0016 0*015 0-013 0015 0015
Bimolecular reaction . . 0*0036 0-0047 0-0048 0-0052 00073 00090
The constancy of the values of ^ in the first case is satisfactory, but not in the second
OZONE AND HYDROGEN PEROXIDE 935
case. Hence it is inferred that the decomposition of hydrogen peroxide by colloidal
platinum is a uni- not a bi-molecular reaction. Again, with a concentration of 0145
gram-molecules of hydrogen peroxide per litre, the values of the constant k are :
Unimolecular reaction . . 0*0075 0'0068 0'0062 0*0057 0-0051 0*0054
Bimolecular reaction . . 0*0015 0*0015 0*0015 0*0015 0*0014 0*0016
Here the fluctuations in the value of the so-called constant show that the decom-
position of the hydrogen peroxide is undoubtedly a bi- and not a uni-molecular
process. The decomposition of hydrogen peroxide by heat similarly follows the
bimolecular law. To summarize, the decomposition of hydrogen peroxide in light,
and when stimulated by colloidal platinum in dilute solutions, is a unimolecular
process ; and when decomposed by heat, or by colloidal platinum in concentrated
solutions, it is a bimolecular process. It has also been found that the velocity of
the photochemical decomposition of hydrogen peroxide is proportional to the
radiant energy absorbed. The energy absorbed during the decomposition of a
gram-molecule of hydrogen peroxide is nearly equal to that given out by the decom-
position of the substance in darkness.
The decomposition of steam by red-hot iron. — ^Let the method just developed
be applied to the reaction of steam on red-hot iron previously described ; and
let Co, Ci, (^2, C3 respectively denote the concentrations of the iron, steam, hydrogen,
and iron oxide at any time t.
3Fe+4H20 =4H2+Fe304
Cq Ci G2 C3
From Guldberg and Waage's law, the velocity of the decomposition of steam will
be proportional to the product of the concentrations of each of the reacting mole-
cules. There are presumably three molecules of iron and four of steam. Hence,
the velocity of the decomposition of steam=A;Co^Ci* ; and, similarly, the velocity
of the oxidation of hydrogen=A;'(72*C3. The condition of equilibrium when these
two velocities are equal must therefore be kCo^Ci^=kV2^C^. The condition of
equilibrium, however, is independent of the concentrations of the two solids ; and
hence, kCo^ must be a constant number, say ki ; and likewise, k'C^ must be another
constant number, say k^. The condition of equilibrium can accordingly be written
kiCi'^=k2C2!^. The concentrations of the two gases, hydrogen and steam, must be
proportional to their partial pressures pi and p^ respectively. Accordingly, the
preceding equation can be written :
^= Constant
since the fourth root of a constant is itself constant. In an experiment by G. Preuner
(1904),7 at 200°, when the partial pressure of steam pi was 4*6 mm. of mercury,
that of hydrogen was 95*9. Hence, the value of the constant is nearly 0*048. In
another experiment at the same temperature, the partial pressure of hydrogen
Pi was 195"3, then that of steam p2 iiiust have been 0*048 xl95*3=9'3 — the
observed value was 9*7. The value of the constant at 440° was 0'176 ; at 900°,
0*69 ; at 1025°, 0-78 ; and at 1150°, 0'86, showing that the ratio of steam : hydrogen
approaches unity with a rise of temperature.
Examples.- — (1) If p^ denotes the partial pressure of steam, p^ that of hydrogen, and
P3 that of oxygen, show that if A* is a constant, then, for the reaction 2H20^2H2-l-02,
(2) When barium peroxide is heated, it decomposes : 2Ba02r=^2BaO + 02. Show that
for any given temperature, jo= constant, where p denotes the partial pressure of oxygen.
References.
^ I. Kant, Metaphysischen Anfangsgrilnden der Naturwisscnschaften, 1786.
2 J. H. van't Hoff, Etudes de dynamique chimique, Amsterdam, 13, 1884.
936 INORGANIC AND THEORETICAL CHEMISTRY
3 J. H. Mathews and H. A. Curtis, Jour7i. Phys. Chem., 18. 101, 521, 1914.
* J. Walker, Proc. Roy. Soc. Edin., 22. 22, 1898.
* G. Bredig and M. von Borneok, Zeit. phys. Chem., 31. 289, 1900.
« G. Dyer and A. B. Dale, Proc. Chem. Soc, 29. 55, 1913.
' G. Preuner, Zeit. phys. Chem., 47. 385, 1904; H. St. C. DeviUe, Compt. Rend., 70. 1105,
1201, 1870 ; 71. 30, 1870 ; H. Debray, ib., 88. 1241, 1879.
§ 11. The Chemical Properties of Hydrogen Peroxide
Solutions of hydrogen peroxide are not very stable, and readily decompose into
oxygen and water. Similar remarks apply to the anhydrous peroxide. If the
liquids are free from other substances they are moderately stable at ordinary
temperatures. J. W. Bruhl ^ found that after anhydrous peroxide had been kept 50
days protected from atmospheric dust, it had lost only one-half per cent, of peroxide
by decomposition. According to R. Wolffenstein, aqueous solutions keep very well
if they are free from alkaline substances, salts of the heavy metals, and from particles
of alumina and silica. R. Bottger, M. Berthelot, and P. Sabatier found that the
presence of acids increases the stability of aqueous solutions. A 3 per cent, solu-
tion suffered no appreciable change after it had been kept for a year. The fact was
well known to L. J. Thenard, who considered that the acid combines chemically
with hydrogen peroxide. He said :
With phosphoric, sulphuric, hydrochloric, hydrofluoric, nitric, oxalic, citric, and acetic
acids hydrogen peroxide forms compounds in which it is less easily decomposable than when
alone. In these compounds, the acid was at first regarded as existing in a higher state of
oxidation. The comparatively weak carbonic and boracic acids do not give stability to
peroxide of hydrogen. . . . The evolution of oxygen gas from these mixtures takes place
less easily and more slowly than from the pure peroxide of hydrogen ; but when the acid
is neutralized by an alkali, the former facility of decomposition is restored. The greater
the quantity of acid mixed with the peroxide, the more does the affinity of the acid for
that compound interfere with its decomposition by elevation of temperature, or by contact
with most of the bodies mentioned below. If any of the acids just enumerated be added
to hydrogen peroxide which has begun to evolve gas, the escape of gas ceases ; it recom-
mences at a higher temperature, but the whole of the oxygen is not driven off, even by
half an hour's boiling. It is remarkable that although gold decomposes the pure peroxide
much more rapidly than bismuth does, yet the quantity of acid required to stop the action
of the gold is smaller than that which must be added to prevent the action of the bismuth.
Hydrogen peroxide brought into a state of effervescence by gold, palladium, or rhodium,
is restored to tranquillity by the addition of a single drop of dilute sulphuric acid.
L. J. Thenard also knew that alkaline solutions do not keep very well. G. Lemoine
attributes the retarding effects of acids to their affinity for water which counter-
acts the catalytic action of the water ; he attributes the accelerating effects of the
caustic alkalies to the cyclic formation and decomposition of alkali peroxides.
Hydrogen peroxide solutions corrode glass vessels faster than water, and the liquid
becomes alkaline. Hydrogen peroxide is for preference kept in paraffin or paraffin-
lined glass bottles, or in quartz-glass vessels. The rate of decomposition of solu-
tions of hydrogen peroxide prepared with ordinary tap -water is said by W. Clayton
to be fifty times the rate with highly purified water.
A little platinum black dropped into the solution may cause an explosion ; in
any case, it causes rapid decomposition. Similar remarks apply to finely divided
gold, silver, and similar metals, as well as to powdered manganese dioxide. The
action appears to be catalytic since the manganese dioxide, etc., remains at the
end of the action unchanged in composition. 2
J. L. Thenard's classical observations on the action of various substances on
eau oxygenee are worth quoting :
Substances which induce the evolution of oxygen without themselves undergoing any
alteration : — A violent action occurs with charcoal (forming carbon dioxide), silver, gold,
platinum, palladium, rhodium, iridium, and osmium. The action is the more vigorous
OZONE AND HYDROGEN PEROXIDE 937
the finer the state of subdivision of the metal. A moderate action occurs with mercury,
lead filings, powdered bismuth, and powdered manganese. The action is slight with
copper, nickel, cobalt, and cadmium. A violent reaction occurs with manganese dioxide,
manganese and cobalt sesquioxides, and lead monoxide. The reaction is moderate with
ferric, potassium, sodium, magnesium, and nickel hydroxides. The reaction is mild with
ferric, nickel, copper, bismuth and magnesium oxides ; and feeble with the magnetic oxide
of iron, and with uranium, titanium, cerium, and zinc oxides, and the hydrated dioxides
of calcium, strontium, and barium. The reaction is very feeble with sodium carbonate,
potassium hydrogen carbonate, manganous, zinc, ferrous, and copper sulphates ; with
potassium, sodium, barium, calcium, antimony, ammonium, and manganous chlorides ;
and with manganous, copper, mercurous, and silver nitrates. The fibrin of blood acts
violently.
Substances which induce the evolution of oxygen but at the same time give up their
own oxygen by reduction : — The oxides of platinum, gold, silver, and mercury are reduced
to the metallic state ; lead dioxide and red lead are reduced to lead monoxide. The action
is in all cases violent.
Substances which allow some of the oxygen of the peroxide to escape as a gas and them-
selves absorb the remainder of the gas to form oxides : — Examples are — selenium forms
selenic acid ; arsenic or arsenious oxide forms arsenic acid ; molybdenum, molybdic acid ;
tungsten, tungstic acid ; and chromium, chromic acid. The metals potassium and sodium
are violently oxidized ; zinc forms zinc oxide ; barium hydroxide forms barium dioxide ;
copper oxide forms a yellow peroxide ; manganic oxide forms manganese dioxide ; cobalt
and iron monoxides form sesquioxides. The sulphides of arsenic, molybdenum, antimony,
lead, iron, and copper are vigorously oxidized to sulphates ; bismuth and stannic sulphides
are slowly converted into sulphates ; mercury and silver sulphides are not oxidized ; and
barium iodide probably forms the iodate.
The following oxides take the whole of the oxygen they require from hydrogen peroxide
without liberating any gas — sulphur dioxide forms the trioxide ; hydrosulphuric acid gives
water, sulphur, and a little sulphuric acid ; hydriodic acid forms iodine and water ; the
peroxides are precipitated from lime, strontia, or baryta water ; and stannous oxide forms
stannic oxide.
No action was observed with antimony ; tellurium ; tin ; iron ; alumina ; silica ;
tungstic acid ; chromium sesquioxide ; antimonious and antimonic oxides ; stannic oxide ;
sodium phosphate ; potassium, sodium, calcimn, barium, or strontium sulphate ; alum ;
turbite ; potassium chlorate ; potassium, sodium, barium, strontium, or lead nitrate ;
zinc, stannic, or mercuric chloride ; white of egg — liquid or coagulated ; glue ; and urea.
Hydrogen peroxide is not decomposed perceptibly faster with organic substances like
potassium oxalate or acetate, alcohol, camphor, olive oil, sandarac, woody fibre, starch,
gum, sugar, mannite, and indigo than when it is alone, but in some cases, the gas evolved
is mixed with carbon dioxide — e.g. with starch or sugar.
J. H. Walton and D. 0. Jones 3 found that the metal salts which catalytically
decompose hydrogen peroxide in aqueous solutions, act similarly if amyl alcohol,
amyl acetate, isobutyl alcohol, or quinoline be substituted for water. The reaction
with manganese acetate in a solution of quinoline with 2 per cent, of water is
bimolecular, and unimolecular if the quinoline be saturated with water. A small
trace of some of the extremely finely divided metals- — colloidal platinum, colloidal
gold, etc. — can accelerate the decomposition of an indefinitely large amount of the
peroxide. The action, though different, has been compared with that of yeast on
a solution of sugar, and these colloidal metal solutions have been styled inorganic
ferments. According to G. Bredig, a gram-atom of colloidal platinum diluted to
approximately 70 million litres, can distinctly accelerate the decomposition of more
than a million times this amount of hydrogen peroxide. The reaction in neutral
and acid solutions is unimolecular, and is irreversible and complete, H202=H20-l-0,
not 2H202=2H20+02 ; with organic ferments, the reactions are not usually
complete. Under similar circumstances, in alkaline solutions, one gram-atom of
colloidal manganese diluted to 10,000,000 litres ; colloidal cobalt or copper to one
million litres ; and colloidal lead to 100,000 litres, can act in a similar way ; since
their action is more or less retarded or paralyzed by traces of certain other sub-
stances, so that the inorganic ferments are said to be poisojied by these agents. *
The catalysis of hydrogen peroxide by colloidal platinum, and the poisoning of the
catalyst has been studied by G. Bredig and his co-workers, J. H. Kastle and A. S.
Loevenhart, E. H. Neilson and 0. H. Brown, etc. The following act as poisons
in retarding the activity of colloidal platinum : arsine, phosphine, phosphorus,
938 INORGANIC AND THEORETICAL CHEMISTRY
carbon disulphide, mercuric chloride, sulphide, or cyanide ; hydrocyanic acid ;
cyanogen iodide ; bromine ; iodine ; hydrogen sulphide ; sodium thiosulphate,
nitrate, and sulphite ; carbon monoxide ; aniline ; hydroxylamine ; hydrochloric
acid ; oxalic acid ; arsenious acid ; phosphorous acid ; nitrous acid ; hydrofluoric
acid ; amyl nitrite ; pyrogallol ; nitrobenzene ; and ammonium chloride and
fluoride. The decomposition is accelerated by hydrazine, dilute nitric acid, and
formic acid; and it is not affected by potassium chlorate, ethyl alcohol, amyl
alcohol, ether, glycerol, turpentine, and chloroform. G. Phragmen studied the
effect of sodium phosphate and of the hydroxide on the decomposition of hydrogen
peroxide.
G. Bredig and W. Reinders investigated the influence of colloidal gold on the
decomposition of hydrogen peroxide in alkaline solutions, and the poisoning of the
catalytic agent by potassium chloride, sodium phosphate, potassium cyanide,
sodium sulphide, thiosulphate, and sulphite. Mercuric chloride stimulates the
activity of the catalyst probably because that salt is reduced to colloidal mercury,
which itself acts catalytically. In feebly alkaline solutions, the effect of 0*0003
mgrm. of colloidal gold is perceptible per c.c. of solution. G. Bredig and his co-
workers have investigated the action of colloidal palladium under similar conditions.
The catalytic agent is activated by hydrogen. Hydrogen cyanide, hydrogen
sulphide, arsine, iodine, and mercuric chloride act as poisons ; while carbon monoxide
acts first as a positive and then as a negative catalyst. G. A. Brossa investigated
the catalytic action of colloidal iridium ; F. Ageno, colloidal horon ; and G. Bredig
and A. Marck, colloidal manganese dioxide. C. Doelter investigated the effect of
a number of minerals ; G. Lemoine, the effect of wood charcoal ; and E. B. Spaer,
the effect of pressure on the decomposition of hydrogen peroxide.
The decomposition of hydrogen peroxide by blood, hcemoglohin, animal or
plant extracts, etc., has been studied by A. Bechamp, G. Senter, etc.^ F. L. Usher
and J. H. Priestley, A. Heffter, J. Dewitz, K. Togami, and E. J. Lesser ^ have studied
the catalysis of enzymes ; A. Bach, by yeast catalase ; H. van Laer, by diastase;
A. Renard, by milk ; and J. J. Ford, by starch. In a general way, the agents
which retard the activity of colloidal platinum also retard the activity of proto-
plasmic catalysts, but not all those which retard the activity of the latter
interfere with the activity of the former.
Charcoal or magnesium mixed with a trace of manganese dioxide ignites
immediately in contact with hydrogen peroxide. With finely powdered iron
or lead, hydrogen peroxide remains quiescent, but if a trace of manganese
dioxide be present, the iron burns. A few drops of liquid hydrogen peroxide
on a piece of cotton wool will make the cotton inflame, although the peroxide
can be filtered through gun-cotton. Similar results are obtained with aqueous
solutions of hydrogen peroxide, but the action is much less vigorous. Rough
surfaces have a disturbing effect on the stability of hydrogen peroxide — a
concentrated solution is decomposed when poured on a ground-glass surface.
W. Clayton (1916) considers that the chief factor in the decomposition of aqueous
solutions of hydrogen peroxide is colloidal organic matter ; he doubts if the nature
of the surface of the vessel is really so active as is generally supposed ; and he
further attributes the observed effects to variations in the amount of colloidal
organic matter which is present. The presence of small quantities of some sub-
stances 7 — e.g. alcohol, glycerol, ether, naphthalene, sodium pyrophosphate, oxalic
acid, pyrogallol, acetanilide (1 : 2000) ; magnesium silicate ; etc.- — act as preserva-
tives and make the solutions more stable, and these agents have been called anti-
catalysts or negative catalysts. The use of many of these preservatives has been
patented. The use of sodium or calcium chloride as a preservative is preferred to
sulphuric or phosphoric acid for medicinal hydrogen peroxide.^ Light is a factor
m the decomposition of hydrogen peroxide ; an eight per cent, solution was decom-
posed completely after ten months' exposure, while a similar solution in darkness
was but little affected.^ H. Thiele found that exposure to ultraviolet light from
OZONE AND HYDROGEN PEROXIDE 939
a mercury lamp hastened the decomposition. i® H. A. Curtis also showed that the
oxidizing power of hydrogen peroxide, as manifested in bleaching dyes, is hastened
in a similar manner. The effect cannot be duplicated by substituting oxygen for
hydrogen peroxide so that the result is not due to the formation of ozone. The
presence of radium bromide increases the speed of decomposition of the peroxide.i^
The compound is also decomposed when heated.
The chemical reactions with hydrogen peroxide fall into five types :
(1) The hydrogen peroxide is decomposed and the second compound is
reduced. With permanganates, for example, both substances are simultane-
ously reduced, and the resulting oxygen comes from both the permanganate
and the peroxide.
(2) The hydrogen peroxide is decomposed, and the second compound is
oxidized by the oxygen derived from the peroxide, as was the case with ozone.
Sulphur dioxide, for instance, changes into sulphuric acid. There are numerous
other similar oxidations.
(3) Certain acids form special addition products with the hydrogen per-
oxide, thus sulphuric acid gives persulphuric acid ; molybdic acid, permolybdic
acid ; chromic acid, perchromic acid ; etc.
(4) Certain bases may react by double decomposition whereby the hydrogen
or part of the hydrogen of the peroxide is replaced by a metal. In this case,
the hydrogen peroxide has the character of an acid.
(5) Hydrogen peroxide unites with many organic and inorganic salts
much in the way of water of crystallization, and it is then called hydrogen
peroxide of crystallization,!^ e.g. NH4Cr05.H202 ; (NH4)2S04.H202.
H. A. Curtis has shown that the oxidizing power of hydrogen peroxide, as
manifested in the bleaching of dyes, is increased by exposing the reaction mixture
to light of short wave-lengths. This effect cannot be duplicated by substituting
oxygen for hydrogen peroxide in the reaction mixture, indicating that the result
is not due to formation of ozone.
Concentrated solutions of hydrogen peroxide in water form 13 a crystalline mono-
hydrate, H2O2.H2O, and dihydrate, H2O2.2H2O. For the action of ozone on
hydrogen peroxide, vide ozone.
The halogens, chlorine, bromine, and iodine, act slightly on hydrogen peroxide
solutions producing the haloid acids and oxygen,!^ e.g. E. Schone gives with chlorine,
H20.0+OH2+Cl2=2HCl+02+H20 ; orH202+Cl2=2HCl+02. C. F. Schonbein
found that bromine gives oxygen, and hydrogen bromide. Iodine, in the presence
of alkali carbonates, is transformed into hydrogen iodide. E. Lenssen found
that hydrogen chloride gives oxygen and the free halogen or chloric acid and
water ; hydrogen bromide or iodide gives oxygen the free halogen. The affinity
of the halogen acids for hydrogen peroxide varies as the affinity of the halogen for
oxygen. 15 The reaction depends on the order of mixing. If hydriodic acid be
added to the peroxide the reaction is more energetic than if the peroxide is added
to the acid ; the reverse obtains with the other haloid acids. It is probable the
peroxide first liberates the haloid acids from the haloid salts and then decomposes
the acid. According to J. Brode,!^ ferrous or cupric sulphate and molybdenum
or tungsten trioxide accelerate the decomposition of hydriodic acid catalytically.
The presence of acids affects the reaction with the different catalytic agents in a
different way. Copper sulphate alone is not very active, but it stimulates the
catalytic effects of ferrous sulphate. Molybdenum and tungsten trioxides act more
vigorously than ferrous sulphate. Hsrpochlorous acid is reduced by hydrogen
peroxide, H202-fH0Cl=H20-|-HCl-}-02 ; chloride of lime reacts similarly, the
reaction is quantitative ; one molecule of oxygen is obtained for each molecule of
hydrogen peroxide employed. Hence, the reaction can be employed for the
quantitative determination of either hydrogen peroxide or chloride of lime.i^
Hydrogen peroxide has no action on the alkali chlorates or perchlorates or per-
chloric acid ; the periodates and periodic acid are reduced to iodates or iodic acid
940 INORGANIC AND THEORETICAL CHEMISTRY
respectively. The iodates decompose hydrogen peroxide catalytically without
themselves being affected by this agent. Bromic acid is reduced to bromine and
hydrobromic acid with the liberation of oxygen. According to S. Tanatar,i8 potas-
sium fluoride forms monoclinic prisms, KF,H202. According to C. F. Schonbein,
potassium iodide in alkaline solution forms free iodine and potassium hydroxide,
2KI+H202=2KOH4-l2 ; so also, in acid solutions, or in the presence of ferrous
sulphate. In the decomposition of iodides by hydrogen peroxide in acid solution,
the liberated iodine may oxidize to iodic acid if the iodine be kept in solution by
the addition of, say, hydriodic acid. The reaction takes place only in the presence
of hydrochloric or hydrobromic acid. V. Auger (1911) represents the course of
the reaction by the equations : 2HC1+H202=2H20+Cl2 ; l2+3Cl2=2ICl3 ; 5ICI3
+9H20^3HI03+l2+15HCl. According to G. Meissner,i® in neutral solutions,
potassium iodide is not decomposed, but the peroxide is decomposed catalytically.
E. Schone represents the reaction by a set of quite imaginary equations : With a
feebly acid solution, iodine is liberated, the liquid becomes alkaline, and oxygen
is evolved. The alkaline reaction disappears in a few days and the coloration by
the free iodine decreases. E. Pechard assumes that an iodate and a periodate are
formed as intermediate products, but periodates are at once reduced to iodates by
hydrogen peroxide ; and iodates and perchlorates have no action on hydrogen
peroxide. More probably, a hypoiodite is the intermediate product : H2O2+KI
=H20+KI0; KI0+H202=KI+H20+02. The corresponding alkali chlorides
and alkali bromides are very slowly attacked in a similar manner.
According to V. Auger, the sodium periodate, Na2H3l06, is very slowly decom-
posed by hydrogen peroxide forming sodium iodate, and liberating more oxygen
than corresponds with the equation because of the catalytic decomposition of
the peroxide. S. Tanatar found acid or alkaline solutions of periodic acid are
reduced to iodic acid in acid or alkaline solutions, while iodic acid is stable ; but
V. Auger found that the results with periodic acid vary according to the conditions
with dilute solutions, the acid is quickly and completely reduced to iodic acid, with
the separation of very little iodine. In concentrated solutions the reaction is incom-
plete and much iodine is formed. Cold solutions of hydrogen peroxide decompose
solutions with less than 0*6 per cent, of iodic acid and iodine is set free ; if over
0*8 per cent, of iodic acid is present the solution remains colourless, owing to the
fact that the reaction l2-f5H202=2HI03+4:H20 proceeds faster than 2HIO3
+5H202=l2+6H20+502. According to S. Tanatar, hydrogen peroxide has no
action on acid or alkaline solutions of perchloric acid or chloric acid; and
bromic acid is reduced to hydrogen bromide with a little bromine, and the
hydrogen bromide is oxidized by the peroxide.
According to M. Kleinstiick, silver chloride suspended in a solution of potassium
hydroxide is quickly reduced by hydrogen peroxide : 2AgCl+H202+2KOH
=2Ag+02 + 2KCl+2H20. A. P. H. Trivelli has studied the action of hydrogen
peroxide on silver subbromide. H. T. Calvert ^^ found that hydrogen peroxide acts
in darkness on a photographic plate as if it had been exposed to light, and it is
thought that the activity of certain metals and organic compounds in darkness on
photographic plates, observed by W. J. Russell, is due to the formation of hydrogen
peroxide by the action of moisture on these substances. L. Graetz found that the
effect produced by hydrogen peroxide is not prevented by shielding the plate with
paper, ebonite, or metal foil. 0. Dony-Henault investigated the hypothesis that
a kind of radioactivity is induced during the decomposition of hydrogen peroxide
into water and oxygen. The activity is lessened by lowering the temperature ; it
is not accelerated by platinum foil, although this metal accelerates the decomposition
of the peroxide ; additions of sulphuric acid, alcohol, or dilute alkalies decrease the
effect. J. Precht and C. Otsuki do not believe that the photographic activity of
hydrogen peroxide is a radiation phenomenon at all ; rather it is a consequence of
the direct reducing action of the vapour of hydrogen peroxide on the gelatinized
silver bromide of the plate. M. Padoa found that if a substance capable of
OZONE AND HYDROGEN PEROXIDE 941
decomposing hydrogen peroxide, e.g. platinum black, or manganese dioxide — be
inserted between the peroxide and the photographic plate, no action occurs.
Hydrogen peroxide oxidizes selenium with the formation of selenic acid. Colloidal
tellurium is attacked by dilute solutions of the peroxide, while the crystalline
element reacts but slowly with 60 per cent, peroxide at l(X)°.2i According to
T. Fairley, the oxidizing power of hydrogen peroxide is singularly dormant in the case
of pure hydrogen sulphide, for if alkaline or other salts be absent, the two substances
may remain in contact a considerable time with no more decomposition than would
have occurred with a solution of hydrogen peroxide alone. Hydrogen sulphide,
however, is very slowly oxidized with the deposition of sulphur and the formation
of sulphuric acid ; but the decomposition is very swift with hydrogen selenide.22
Sulphurous acid is oxidized to sulphuric acid, but according to L. Marino, selenious
acid is not oxidized ; the sulphides, hyposulphites, and tetrathionates are also
oxidized to sulphates. A. Gutbier found that tellurium and tellurium dioxide, in
alkaline solutions, are oxidized 23 to telluric acid, H2Te04. Concentrated sulphuric
acid forms Caro's acid, HO.OSO3H ; and when the latter is treated with water, it
forms hydrogen peroxide and sulphuric acid.^^ Hydrogen peroxide transforms
nitrous acid quantitatively into nitric acid ; 25 nitric oxide, NO, furnishes a product
which blues starch and potassium iodide paper very quickly, and hence, said C. F.
Schonbein, the product cannot be nitric acid ; and it is thought to be a compound
of the two components. Ammonia is quickly oxidized to the nitrite and nitrate.
If hydrogen peroxide be added to a large excess of ammonia dissolved in ether,
and the solution cooled to —48°, a crystalline deposit of (NH4)202.H20 is formed ;
the crystals rapidly decompose at ordinary temperatures. ^6 Hydroxylamine
hydrochloride at 50° forms some nitric acid, and gives ofi a mixture of nitrogen
and oxygen.27 Hydroxylamine sulphate, (NH20H)2H2S04, is quantitatively
oxidized by hydrogen peroxide at 40°: (NH20H)2.H2S04+6H202=H2S04
-)-2HN03+8H20 ; in alkaline solutions, nitrous and nitric oxides, nitrous and
nitric acid are formed. 28 According to T. Weyl, hydrogen peroxide solutions— 6
to 30 per cent. — at 60°, convert yellow phosphorus into phosphine, phosphorous
and phosphoric acids. The peroxide acts more vigorously on Schenck's phosphorus.
The reaction has been represented 29 by the equations: 3H202+2P=2P(OH)3,
followed by 4P(OH)3=PH3+3PO(OH)3.
Arsenic furnishes arsenic acid without giving off oxygen. F. Raschig found
that antimony sulphide dissolves in an ammoniacal solution of hydrogen peroxide
forming ammonium antimoniate and ammonium sulphate.^o With bismuth nitrate
in a warm, slightly alkaline solution, K. Hasebrock reported that hydrogen peroxide
gives 3^ellowish-brown bismuthic anhydride ; bismuthous hydroxide gives the same
product.31 K. Hasebrock found that the reaction is quantitative.
Some metals which appear to be either insoluble or, but sparingly soluble in
acids often dissolve in the cold dilute acid if hydrogen peroxide be present.32 With
hydrochloric acid, the action is due to the generation of free chlorine, and accord-
ingly all the metals, except those which form insoluble chlorides, are dissolved. Thus,
T. Fairley found that a mixed solution of ferric chloride and hydrogen peroxide
dissolves gold, and adds that the dissolution is accelerated by heat, and is sometimes
followed by a reprecipitation of the gold. With dilute sulphuric acid, copper,
silver, mercury, nickel, and bismuth are soluble in the presence of hydrogen peroxide,
while tin, lead, gold, platinum, and antimony are not attacked. A mixture of
glacial acetic acid with hydrogen peroxide dissolves copper, silver, lead, mercury,
and bismuth in the cold, but has no action on tin, nickel, gold, and platinum.
T. Fairley says that gold dissolves readily in a mixed solution of potassium cyanide
and hydrogen peroxide, but only at the surface of a solution of the potassium
cyanide alone. The presence of hydrogen peroxide also hinders the precipitation
of gold by ferrous sulphate or oxalic acid. C. Weltzien found magnesium slowly
reacts with hydrogen peroxide, forming an alkaline liquid, which on evaporation
gives a white mass soluble in water, and which is probably magnesium hydroxide ;
942 INORGANIC AND THEORETICAL CHEMISTRY
likewise also with aluminium. H. B. Baker and L. H. Parker found the reaction
with sodium amalgam is faster with a solution of hydrogen peroxide than with
water. There is probably a peroxidation as indicated below. According to
S. Droste, a 3 per cent, solution of hydrogen peroxide slowly dissolves aluminium ;
250 c.c. dissolved 0*2 grm. of the metal in 45 days, forming white insoluble aluminium
hydroxide, Al(OH)3.H20 ; the soluble or colloidal hydroxide does not appear to
be formed. T. Okaya has studied the rhythmic decomposition of hydrogen
peroxide by mercury (q.v.).
Powdered silver is a powerful catalytic agent in the decomposition of hydrogen
peroxide. L. J. Thenard found that in the presence of nitric acid silver oxide is
partly reduced and partly dissolved. According to M. Berthelot, there is a cyclic
series of reactions in which the metal is alternately peroxidized and reduced. 33 Silver,
Ag20, is reduced to metallic silver by a reaction which B. C. Brodie symbolizes :
H202+Ag20->2Ag+H20+02, so that half the oxygen is derived from the silver
oxide and half from the hydrogen peroxide. M. Berthelot has shown that metallic
silver is not exclusively formed, since some of the reduced silver is peroxidized.
W. Manchot, and A. von Baeyer and V. Villiger have studied the action of hydrogen
peroxide on silver. The finely divided silver, formed during the reaction between
silver oxide and hydrogen peroxide, acts catalytically on the latter, so that a mixture
of an excess of hydrogen peroxide on metallic silver always gives off more oxygen
than is represented by the above equation, and there is no need for assuming the
formation of a silver peroxide. The catalytic action of finely divided silver, gold,
platinum is most vigorous in alkaline solutions, weakest in acid solutions ; and inter-
mediate in neutral solutions. T. Fairley assumes that unstable oxides are formed
in alkaline solutions, and more stable salts are formed in acid solutions. Hence,
most metals dissolve in dilute acids in the presence of hydrogen peroxide.
According to E. Mulder, the action of hydrogen peroxide on silver oxide, dioxide,
carbonate, nitrate, and peroxynitrate, is catalytic. Gold oxide is similarly reduced :
Au203+3H202->2Au+3H204-302. Curiously enough, in these reactions the
reducing agent is itself reduced ; usually the reducing agent is oxidized during the
reduction.
A series of peroxides are formed with solutions of the hydroxides or salts of the
alkalies, alkaline earths, or metals. These reactions show that hydrogen peroxide
behaves like a monobasic or a dibasic acid. V. Macri noted that when ammonia
is added to hydrogen peroxide in the presence of calcium chloride, calcium dioxide,
Ca02, is precipitated. Hydrogen peroxide is decomposed catalytically by carbon ;
there is no appreciable oxidation ; a mixture of carbon, magnesium, and manganese
dioxide takes fire in hydrogen peroxide. 0. Masson 34 found that potassium cyanide
at ordinary temperatures gives potassium cyanate, potassium carbonate, and
ammonium carbonate. No oxygen is given off so long as any potassium cyanide
remains unoxidized. With titanium salts 35 hydrogen peroxide gives an orange-
yellow coloration supposed to be due to the formation of pertitanic anhydride, TiOs,
by a reaction symbolized: Ti02+H202=H20+Ti03. In alkaline solutions, salts
of the type Na2O.TiO3.3H2O are formed. The particular tint depends upon the
amount of titanium present, and hence the reaction is used for the determination
of the amount of titanium in various materials. The tint of a solution containing
an unknown amount of titanium is compared with that of similar solutions con-
taining a known quantity of titanium ; and the amount in the unknown solution
determined by simple rule of three. The reaction is also used as a test for hydrogen
peroxide. It is said that one part of titanium in 1800 parts of water gives a deep
yellow coloration, and one part in 180,000 a light yellow coloration. Cerium and
vanadium salts give a brick-red coloration ; molybdenum salts, the intense yellow
of permolybdic acid — H2M02O8 ; uranium salts, a bluish coloration due to the
formation of peruranic acid, U04(H202)2 ; in alkaline solutions yellow peruranates
are formed. With tungsten salts, pertungstates are obtained in a similar way.
J. R. Cain and J. C. Hostetter find that vanadic acid is reduced by hydrogen
OZONE AND HYDROGEN PEROXIDE 943
peroxide. Zirconia and cerium oxide give peroxides analogous to pertitanic oxide ;
and thorium oxide ^^ gives a peroxide Th207. Lead monoxide is converted into the
puce-coloured dioxide, Pb02, by an alkaline solution of peroxide, and the lead
dioxide is then decomposed forming the monoxide. ^7 Lead dioxide, obtained when
red lead is digested with dilute nitric acid, dissolves very slowly, but if a few drops
of hydrogen peroxide be added, all the lead dioxide dissolves in a few moments.
The lead dioxide is reduced to lead monoxide by the hydrogen peroxide, Pb02+H202
=PbO +1120+02, and the product dissolves immediately in the dilute acid. This
method is generally employed to hasten the solution of red lead in dilute acid
prior to analysis. Thallium oxides behave similarly. C. F. Schonbein found that
thallium is oxidized to TIO(OH) ; with an excess of hydrogen peroxide, the oxidation
products are oxygen, water, and thallous hydroxide. The latter is not affected by
hydrogen peroxide. Mercuric oxide is reduced to mercurous oxide in alkaline solu-
tions. G. Bredig and A. AntropofE ^8 obtained an explosive peroxidized compound,
Hg02, by the action of hydrogen peroxide on mercuric oxide ; copper sulphide forms
the sulphate. The precipitation of oxides by hydrogen peroxide from alkaline or
ammoniacal solutions of copper, silver, mercury, and bismuth salts, with the evolu-
tion of oxygen, is the basis of several processes for the separation -of a number of
metals from one another by W. R. E. Hodgkinson and A. H. Coote, etc. C. F. Schon-
bein found cupric hydroxide changes from blue to green when treated with hydrogen
peroxide, and forms an unstable peroxide, H2CUO3 ; ^^ on the contrary, in alkaline
solutions, cupric salts are reduced to cuprous salts.^o StamiOUS salts are oxidized
to stannic salts without loss of oxygen. In neutral or acid solutions' ferrous salts
are oxidized to ferric salts ; and in alkaline solutions ferric hydroxide and alkaline
ferrates are reduced to ferrous salts. Ferric salts are not affected. H. Colin and
A. Senechal have studied the action of ferric chloride on hydrogen peroxide. Alka-
line solutions of potassium ferricyanide were reported by E. Lenssen to be reduced
to the ferrocyanide, 2K3FeCy6+2KOH+H202--2K4FeCy6+2H20+02, and ferro-
cyanides to be oxidized to ferricyanides. C. Weltzien said the opposite is true.
J. Quincke recommends a process based on the reduction of ferricyanide to ferro-
cyanide for the volumetric determination of hydrogen peroxide, or of potassium
ferricyanide. In neutral or acid solutions, the reaction is reversed, and potassium
ferrocyanide is oxidized to the ferricyanide. M. Prud'homme ^^ assumes that when
solutions of potassium ferricyanide and hydroxide are boiled, equilibrium is at-
tained in the reversible reaction, 2K3FeCy6+2KOH=2K4FeCy6+H202, because
(i) potassium ferricyanide and hydroxide are formed when a solution of hydrogen
peroxide is added to one of the ferrocyanide ; (ii) the addition of an excess of
hydrogen peroxide to a solution of potassium ferricyanide and sodium hydroxide
forms the ferrocyanide with the evolution of oxygen ; and (iii) indigotin is bleached
more rapidly by hydrogen peroxide in the presence of alkali hydroxide than in acid
solutions. E. S. Barralet, C. F. Schonbein, E. Schone, and W. Wobbe have examined
the sensitiveness of the reaction as a test for hydrogen peroxide. According to
T. Bayley, in alkaline solutions, cobaltous hydroxide forms a black peroxide ;
nickel hydroxide is not changed, but the sesquioxide forms nickelous hydroxide,
with the evolution of oxygen. G. Watson has studied the action of hydrogen
peroxide on ammoniacal solutions of nickel sulphate. Iron sulphide forms the
sulphate ; molybdenum sulphide forms sulphuric and molybdic acids ; brown lead
sulphide forms the white sulphate ; bismuth sulphide and tin sulphide are only
attacked slightly ; mercury and silver sulphides are still less attacked. Metallic
iron is but slightly attacked, ^^ tungsten and molybdenum respectively form tungsten
and molybdenum trioxides. According to V. Maori, hydrogen peroxide prevents
the precipitation of ammonium phosphomolybdate when solutions of ammonium
molybdate and phosphoric acid are mixed. B. KurilofE and W. Stadel found that
zinc oxide is but slightly attacked ; zinc hydroxide is peroxidized.^^ Similarly with
magnesium hydroxide and with cadmium hydroxide. The oxides of yttrium,
didymium, lanthanum, and samarium furnish oxides approximating R4O9. With
942 INORGANIC AND THEORETICAL CHEMISTRY
likewise also with aluminium. H. B. Baker and L. H. Parker found the reaction
with sodium amalgam is faster with a solution of hydrogen peroxide than with
water. There is probably a peroxidation as indicated below. According to
S. Droste, a 3 per cent, solution of hydrogen peroxide slowly dissolves aluminium ;
250 c.c. dissolved 0*2 grm. of the metal in 45 days, forming white insoluble aluminium
hydroxide, Al(OH)3.H20 ; the soluble or colloidal hydroxide does not appear to
be formed. T. Okaya has studied the rhythmic decomposition of hydrogen
pero2dde by mercury (q.v.).
Powdered silver is a powerful catalytic agent in the decomposition of hydrogen
peroxide. L. J. Thenard found that in the presence of nitric acid silver oxide is
partly reduced and partly dissolved. According to M. Berthelot, there is a cyclic
series of reactions in which the metal is alternately peroxidized and reduced, 33 Silver,
Ag20, is reduced to metallic silver by a reaction which B. C. Brodie symbolizes :
H202+Ag20->2Ag+H20+02, so that half the oxygen is derived from the silver
oxide and half from the hydrogen peroxide. M. Berthelot has shown that metallic
silver is not exclusively formed, since some of the reduced silver is peroxidized.
W. Manchot, and A. von Baeyer and V. Villiger have studied the action of hydrogen
peroxide on silver. The finely divided silver, formed during the reaction between
silver oxide and hydrogen peroxide, acts catalytically on the latter, so that a mixture
of an excess of hydrogen peroxide on metallic silver always gives ofi more oxygen
than is represented by the above equation, and there is no need for assuming the
formation of a silver peroxide. The catalytic action of finely divided silver, gold,
platinum is most vigorous in alkaline solutions, weakest in acid solutions ; and inter-
mediate in neutral solutions. T. Fairley assumes that unstable oxides are formed
in alkaline solutions, and more stable salts are formed in acid solutions. Hence,
most metals dissolve in dilute acids in the presence of hydrogen peroxide.
According to E. Mulder, the action of hydrogen peroxide on silver oxide, dioxide,
carbonate, nitrate, and peroxynitrate, is catalytic. Gold Oxide is similarly reduced :
Au203+3H202->2Au-}-3H20+302. Curiously enough, in these reactions the
reducing agent is itself reduced ; usually the reducing agent is oxidized during the
reduction.
A series of peroxides are formed with solutions of the hydroxides or salts of the
alkalies, alkaline earths, or metals. These reactions show that hydrogen peroxide
behaves like a monobasic or a dibasic acid. V. Macri noted that when ammonia
is added to hydrogen peroxide in the presence of calcium chloride, calcium dioxide,
Ca02, is precipitated. Hydrogen peroxide is decomposed catalytically by carbon ;
there is no appreciable oxidation ; a mixture of carbon, magnesium, and manganese
dioxide takes fire in hydrogen peroxide. 0. Masson 3* found that potassium cyanide
at ordinary temperatures gives potassium cyanate, potassium carbonate, and
ammonium carbonate. No oxygen is given off so long as any potassium cyanide
remains unoxidized. With titanium salts 35 hydrogen peroxide gives an orange-
yellow coloration supposed to be due to the formation of pertitanic anhydride, TiOa,
by a reaction symbolized: Ti02+H202=H20+Ti03. In alkaline solutions, salts
of the type Na2O.TiO3.3H2O are formed. The particular tint depends upon the
amount of titanium present, and hence the reaction is used for the determination
of the amount of titanium in various materials. The tint of a solution containing
an unknown amount of titanium is compared with that of similar solutions con-
taining a known quantity of titanium ; and the amount in the unknown solution
determined by simple rule of three. The reaction is also used as a test for hydrogen
peroxide. It is said that one part of titanium in 1800 parts of water gives a deep
yellow coloration, and one part in 180,000 a light yellow coloration. Cerium and
vanadium salts give a brick-red coloration ; molybdenum salts, the intense yellow
of permolybdic acid— H2Mo20g ; uranium salts, a bluish coloration due to the
formation of peruranic acid, U04(H202)2 ; in alkaline solutions yellow peruranates
are formed. With tungsten salts, pertungstates arc obtained in a similar way.
J. R. Cain and J. C. Hostetter find that vanadic acid is reduced by hydrogen
OZONE AND HYDROGEN PEROXIDE 943
peroxide. Zirconia and cerium oxide give peroxides analogous to pertitanic oxide ;
and thorium oxide ^^ gives a peroxide Th207. Lead monoxide is converted into the
puce-coloured dioxide, Pb02, by an alkaline solution of peroxide, and the lead
dioxide is then decomposed forming the monoxide. 37 Lead dioxide, obtained when
red lead is digested with dilute nitric acid, dissolves very slowly, but if a few drops
of hydrogen peroxide be added, all the lead dioxide dissolves in a few moments.
The lead dioxide is reduced to lead monoxide by the hydrogen peroxide, Pb02+H202
=PbO+H204-02, and the product dissolves immediately in the dilute acid. This
method is generally employed to hasten the solution of red lead in dilute acid
prior to analysis. Thallium oxides behave similarly. C. F. Schonbein found that
thallium is oxidized to TIO(OH) ; with an excess of hydrogen peroxide, the oxidation
products are oxygen, water, and thallous hydroxide. The latter is not affected by
hydrogen peroxide. Mercuric oxide is reduced to mercurous oxide in alkaline solu-
tions. G. Bredig and A. Antropoff ^8 obtained an explosive peroxidized compound,
Hg02, by the action of hydrogen peroxide on mercuric oxide ; copper sulphide forms
the sulphate. The precipitation of oxides by hydrogen peroxide from alkaline or
ammoniacal solutions of copper, silver, mercury, and bismuth salts, with the evolu-
tion of oxygen, is the basis of several processes for the separation of a number of
metals from one another by W. R. E. Hodgkinson and A. H. Coote, etc. C. F. Schon-
bein found cupric hydroxide changes from blue to green when treated with hydrogen
peroxide, and forms an unstable peroxide, H2CUO3 ; ^9 on the contrary, in alkaline
solutions, cupric salts are reduced to cuprous salts.^o Staimous salts are oxidized
to stannic salts without loss of oxygen. In neutral or acid solution* ferrous salts
are oxidized to ferric salts ; and in alkaline solutions ferric hydroxide and alkaline
ferrates are reduced to ferrous salts. Ferric salts are not affected. H. Colin and
A. Senechal have studied the action of ferric chloride on hydrogen peroxide. Alka-
line solutions of potassium ferricyanide were reported by E. Lenssen to be reduced
to the ferrocyanide, 2K3FeCy6+2KOH+H202=2K4FeCy6+2H20+02, and ferro-
cyanides to be oxidized to ferricyanides. C. Weltzien said the opposite is true.
J. Quincke recommends a process based on the reduction of ferricyanide to ferro-
cyanide for the volumetric determination of hydrogen peroxide, or of potassium
ferricyanide. In neutral or acid solutions, the reaction is reversed, and potassium
ferrocyanide is oxidized to the ferricyanide. M. Prud'homme *^ assumes that when
solutions of potassium ferricyanide and hydroxide are boiled, equilibrium is at-
tained in the reversible reaction, 2K3FeCy6+2KOH=2K4FeCy6-|-H2025 because
(i) potassium ferricyanide and hydroxide are formed when a solution of hydrogen
peroxide is added to one of the ferrocyanide ; (ii) the addition of an excess of
hydrogen peroxide to a solution of potassium ferricyanide and sodium hydroxide
forms the ferrocyanide with the evolution of oxygen ; and (iii) indigotin is bleached
more rapidly by hydrogen peroxide in the presence of alkali hydroxide than in acid
solutions. E. S. Barralet, C. F. Schonbein, E. Schone, and W. Wobbe have examined
the sensitiveness of the reaction as a test for hydrogen peroxide. According to
T. Bayley, in alkaline solutions, cobaltous hydroxide forms a black peroxide ;
nickel hydroxide is not changed, but the sesquioxide forms nickelous hydroxide,
with the evolution of oxygen. G. Watson has studied the action of hydrogen
peroxide on ammoniacal solutions of nickel sulphate. Iron sulphide forms the
sulphate ; molybdenum sulphide forms sulphuric and molybdic acids ; brown lead
sulphide forms the white sulphate ; bismuth sulphide and tin sulphide are only
attacked slightly ; mercury and silver sulphides are still less attacked. Metallic
iron is but slightly attacked,^^ tungsten and molybdenum respectively form tungsten
and molybdenum trioxides. According to V. Maori, hydrogen peroxide prevents
the precipitation of ammonium phosphomolybdate when solutions of ammonium
molybdate and phosphoric acid are mixed. B. Kuriloff and W. Stadel found that
zinc oxide is but slightly attacked ; zinc hydroxide is peroxidized.'*^ Similarly with
magnesium hydroxide and with cadmium hydroxide. The oxides of yttrium,
didymium, lanthanum, and samarium furnish oxides approximating R4O9. With
944 INORGANIC AND THEORETICAL CHEMISTRY
alkaline solutions, hydrogenperoxide transforms the chromic oxides ^4 into chromates ;
with neutral or acid solutions of chromic acid, H2Cr04, hydrogen peroxide forms a
blue solution which immediately begins to decompose with the evolution of oxygen.
The solutions of both hydrogen peroxide and of chromic acid are comparatively
stable in the cold ; when mixed they simultaneously decompose — half the oxygen
comes from the peroxide and half from the chromic acid. It is therefore inferred
that an unstable compound of both is formed — possibly 3H202.2Cr03 — which
breaks up with the evolution of oxygen leaving behind chromic sesquioxide, CrgOs,
which immediately dissolves in the acid solution. The transient intermediate
compound has been isolated by operating at a low temperature. The blue-coloured
peroxide, whatever it be, is much more soluble and stable in ethereal than in aqueous
solutions, so that if a solution of chromic acid and hydrogen peroxide in a test-tube
be shaken with ether, a blue ethereal solution of the peroxide will float on the
surface of the aqueous layer. The compound decomposes when the ether is
evaporated. This reaction is used for the detection of chromates. Add hydrogen
peroxide to the neutral or alkaline solutions containing a chromate, and then
acidify with dilute sulphuric acid. The presence of a chromate is indicated by a
blue coloration. If but small quantities of chromate be present, shake up the
solution with 2 or 3 c.c. of ether. The separation of a blue ethereal layer indicates
chromic acid. The necessary modification of the process for the detection of
hydrogen peroxide will be obvious. It is said that this method will indicate one
part of hydrogen peroxide in 80,000 parts of water. If chromic acid be added to
the solution of hydrogen peroxide mixed with hydrogen sulphide, V. Macri found
that the latter makes no difierence to the reaction.
The action o! hydrogen peroxide on manganese compounds. — ^Hydrogen
peroxide transforms manganous hydroxide in neutral or alkaline solutions into the
dioxide which catalyticalTy decomposes the hydrogen peroxide ; .if the solution is
acid, manganese dioxide is reduced to manganous oxide : Mn02+H202+2HC1
=MnCl2+02+2H20. The reducing action of the hydrogen peroxide is only
apparent. According to B. C. Brodie (1872), ^^ the oxides of silver, manganese,
etc., have an atom of oxygen which is readily disengaged from its combination.
Similarly, hydrogen peroxide readily parts with its odd atom of oxygen. Conse-
quently, the atom of oxygen in hydrogen peroxide is supposed to oxidize the odd
oxygen atom in the metallic peroxide. According to C. Weltzien, a neutral solution
of potassium permanganate is reduced to potassium hydroxide and hydrated man-
ganese dioxide, which catalytically decomposes the peroxide. In the presence of
sulphuric or nitric acid, the peroxide reduces the permanganate to a manganous
salt : 5H202+2KMn044-3H2S04=2MnS04+K2S04+8H20+502, so that the solu-
tion of potassium permanganate, acidified with sulphuric acid, is rapidly reduced
and decolorized by hydrogen peroxide. The reaction is quantitative and is used
in the volumetric determination of hydrogen peroxide.
According to C. F. Schonbein, the presence of a millionth part of hydrogen
peroxide in a solution can be detected by its decolorizing action. Consequently,
if an acidified solution, containing a known amount of potassium permanganate,
be run from a burette into a known volume of a solution of hydrogen peroxide
until the pink colour of the permanganate is no longer discharged, it follows, from
the equation, that every two molecules of KMn04 correspond with five molecules
of H2O2 ; or 2x158 (the approximate molecular weight of KMn04) grams of
potassium permanganate correspond with 5x34 (the approximate molecular
weight of H2O2) grams of hydrogen peroxide ; otherwise expressed, one gram of
potassium permanganate represents 0'5382 gram of hydrogen peroxide.
Example. — 45 c.c. of a standard solution of potassium permanganate containing 20
grams of KMnO, per litre were decolorized by 25 c.c. of a solution of hydrogen peroxide.
What amount of HgOj is present in a litre of the hydrogen peroxide ? Here 1000 c.c.
of the standard solution contain 20 grams of KMn04 ; hence, 1 c.c. contains 0"02 gram ;
or 45 c.c. contain 09 gram ; but from the equation, one gram of KMn04 represents 0'5382
OZONE AND HYDROGEN PEROXIDE 945
gram of HaOg ; hence, 25 c.c. of hydrogen peroxide has 0-5382 X 09 = 0-4844 gram of
H2O2. Hence, a litre will have 19-4 grams of H^Og.
Less permanganate is required for titrating a mixture of hydrogen peroxide
and sulphuric acid than if a mixture of permanganate and sulphuric acid is used for
titrating hydrogen peroxide alone. According to T. M. Lowry and J. H. West,
this is due to the formation of persulphuric acid when sulphuric and hydrogen
peroxide are mixed together, and the slowness of the reaction between persulphuric
acid and potassium permanganate.
Higher hydrogen peroxides.— M. Berthelot (1880) ^^ noticed that when potassium
permanganate is titrated with hydrogen peroxide at a low temperature, say 12°, the
permanganate is decolorized without liberating oxygen, and hence he concluded that
this is due to the formation of a compound H2O3, or, as A. Bach suggests, H2O4, which
is stable only at low temperatures. For instance, with hydrogen peroxide, H2O2,
one gram-molecule of oxygen is liberated for every molecule of the peroxide decom-
posed: 2KMn04+5H202+3H2S04=K2S04+2MnS04-f8H20+502; with hydrogen
tetroxide, if it exists, and if it reacts in an analogous manner, each molecule
requires just as much potassium permanganate as hydrogen peroxide, but twice as
much oxygen would be liberated : 2KMn04+5H204+3H2S04-K2S04+2MnS04
4-8H2O+IOO2. Similar remarks would apply to Berthelot's hypothetical H2O3.
According to A. Bach (1897),*'^ also, the oxidation products of nascent hydrogen
from palladium hydride oxidize indigo solutions more rapidly than hydrogen
peroxide, and he therefore inferred that a higher peroxide than H2O2 is formed
during the slow oxidation of hydrogen. Again, there is nothing to show that the
rubidium and potassium tetroxides have the respective formulae : Rb204 and K2O4 ;
analysis alone gives a percentage composition corresponding with RO2 and KO2 ;
but, just as sodium peroxide is represented by the formula Na202 on account of
its relation to hydrogen peroxide, known to have a molecular formula H2O2, so the
peroxides of rubidium and potassium are assumed to be derivatives of a hypothetical
hydrogen peroxide, H2O4.
A. Bach (1900) sought for the supposed higher hydrogen peroxides (1) in the
oxidation products of nascent hydrogen ; (2) in the product derived from the action
of dilute acids on sodium peroxide, Na202 ; and (3) on potassium tetroxide, K2O4 ;
and (4) in the oxidation product of Caro's acid on potassium permanganate. It
was found that the corresponding amounts of oxygen obtained from each of these
products by the permanganate titration is :
Bydrogen
(1) Oxidation products
(2) Sodium
(3) Potassium
(4) Caro'8
peroxide.
of nascent hydrogen.
peroxide.
tetroxide.
acid.
1
1-07
1-17
1-28
1-65
These results might be caused by the presence of hydrogen trioxide, H2O3, or hydro-
gen tetroxide, H2O4. A. M. Clover (1903) failed to verify A. Bach's conclusion.
During the permanganate titration of A. Bach, the reaction, 2KMn04+3H2S04
-(-5H202=K2S044-2MnS04+8H20+502 occurs, and the hydrogen in the solution
can be estimated from (1) the amount of standard permanganate used in the
titration, or (2) from the volume of oxygen evolved. Bach found from 25 to 34 per
cent, more oxygen was evolved than corresponded with the permanganate required
for the titration, and he concluded that a higher peroxide than H2O2 must have
been present in the solution. H. E. Armstrong (1900) and W, Ramsay (1901) *^
tried to explain Bach's results by assuming that hydrogen peroxide was consumed
in a secondary reaction — namely, the formation of persulphuric acid, or Caro's acid,
by the interaction of hydrogen peroxide with the sulphuric acid in the solution ;
but A. von Baeyer and V. Villiger (1900) showed that neither Caro's acid nor persul-
phuric acid rapidly affected potassium permanganate, so that titration with perman-
ganate does not give the strength of a solution of hydrogen peroxide in sulphuric
acid ; more peroxide is present than is represented by the amount of permanganate
consumed. This criticism cannot be valid because in his permanganate titrations,
VOL. I. 3p
946 INORGANIC AND THEORETICAL CHEMISTRY
A. Bach does not appear to have used sufficient acid, and some manganese dioxide
was in consequence precipitated ; this acted catalytically, decomposed the hydrogen
peroxide remaining in the solution with the evolution of oxygen ; hence, more oxygen
gas was formed than corresponded with the hydrogen peroxide actually decomposed
by the permanganate. This conclusion is confirmed (i) by the results obtained when
sufficient acid is present to keep the manganese oxide in solution, and (ii) solutions
of sodium peroxide also give an excess of oxygen if an insufficient amount of acid
is present during the titration. At present, therefore, the evidence in support of the
higher hydrogen peroxides is not satisfactory.
According to M. Kleinstiick,*^ carbonyl chloride and phenyl carbonate react with
alkaline hydrogen peroxide, and so does a saturated solution of potassium hydrogen
carbonate in a pressure bottle at 100°. The distillate obtained by passing steam
into the product reduces ammoniacal silver oxide, and is therefore said to contain
formaldehyde, H.COH. M. Kleinstiick therefore suggests that possibly the assimi-
lation of carbon dioxide by plants proceeds : 2H2C03+2H202==2H.COH-f 2H2O
-J-3O2. Hydrogen peroxide oxidizes many organic compounds particularly in
the presence of an inorganic salt as catalytic agent — e.g. it converts sugars into
ozones, and benzene into phenol in the presence of ferrous sulphate ; many organic
alkaloids are converted into new crystalline bases which are often coloured ; thus
quinine turns lemon-yellow ; nicotine, blood-red ; etc. With potassium cyanide
it forms potassium cyanate, KCy-|-H202=KCyO-f H2O, and according to H. Cook,
ammonia and potassium formate are simultaneously produced. The mono-
hydric alcohols are not attacked, but the polyhydric alcohols— glycol, glycerol,
mannite, etc. — are oxidized to the corresponding aldehydes, particularly in presence
of ferrous sulphate. OxaUc acid is converted into carbon dioxide ; tannin, gallic
acid, and P3^0gallol are not browned by hydrogen peroxide ; indigo solution is
slowly bleached, and more rapidly if ferrous sulphate is present. Tincture of
guaiacum is turned blue. White of egg in a solution of lactic acid, and the serum
of blood, become insoluble at 40°. Fibrin and blood act catalytically on the decom-
position of hydrogen peroxide.
The uses of hydrogen peroxide. — Hydrogen peroxide bleaches many organic
colouring agents — e.g. litmus and indigo solutions. Dilute solutions of hydrogen
peroxide are used for bleaching silk, feathers, straw, hair, ivory, teeth, etc., where
more violent bleaching agents — e.g. chlorine — would injure the material. Instead
of hydrogen peroxide an acidified solution of sodium peroxide is sometimes employed.
The actions are similar. Since the products of the decomposition of hydrogen per-
oxide— water and oxygen — are harmless, it is also used medicinally as an antiseptic,
etc. Numerous mixtures of hydrogen peroxide with disinfectants have been regis-
tered, and they are sold under various trade names — e.g. perhydrol, dioxogen, hydro-
zone, glycozone, pyrozone, peroxal, etc. M. Pettenkofer's proposal is to use hydrogen
peroxide for cleaning oil paintings which have been darkened by the action of hydro-
gen sulphide — sometimes present in the air of towns — upon the lead compounds in
the paint. The brownish-black coloured lead sulphide is transformed into white
lead sulphate. According to reports, the treatment is sometimes satisfactory and
sometimes it spoils the picture. Ethereal solutions of hydrogen peroxide are used
in photography for intensifying negatives. Hydrogen peroxide is also used in analy-
tical work for the oxidation of sulphites to sulphates ; arsenites to arsenates ;
chromic salts to chromates ; ferrous to ferric salts ; nitrites to nitrates ; etc.
References.
1 J. W. BruhJ, Ber., 28. 2847, 1895 ; 30. 162, 1897 ; 33. 1709, 1899 ; R. Bottger, Dingier' s
Journ., 209. 157, 1873 ; M. Berthelot, CompL Bend., 90. 897, 1880 ; P. Sabatier, Bull. Soc. c'him.,
(2), 44. 169, 1885 ; G. Ixmoine, Compt. Bend., 161. 47, 1915 ; W. Clayton, Trans. Faraday Soc,
11. 164» 1916.
2 L. J. Thcnard, Ann. Chim. Phys., (2), 9. 441, 1818; (2), 10. 114, 335, 1819; (2), 11. 85,
1819.
OZONE AND HYDROGEN PEROXIDE 947
3 J. H. Walton and D. 0. Jones, Journ. Amer. Chem. Soc, 38. 1956, 1916.
* G. Bredig and M. Fortner, Ber., 37. 798, 1904 ; G. Bredig and J. Weinmayr, Zeit. phya. Chem.,
42. 601, 1903 ; G. Bredig and R. Muller von Beraeck, ib., 31. 258, 1899 ; G. Bredig and K. Ikeda,
ib., 37. 1, 1901 ; G. Bredig and W. Reinders, ib., 37. 323, 1901 ; G. Bredig and A. Marck, Bemme-
len'fi Festschrift, 342, 1911 ; T. S. Price and A. D. Denning, ib., 46. 89, 1904 ; C. H. Neilson and
0. H. Brown, Amer. Jour7i. Physiol, 10. 225, 335, 1904 ; T. S. Price and J. A. N. Friend, Journ.
Chem. Soc, 85. 1526, 1904; C. Liebermann and W. Genersich, Pfluger's Archiv., 104. 119, 155,
1904 ; G. Senter, Proc. Roy. Soc, 74. 566, 1905 ; G. A. Brossa, Zeit. phys. Chem., 66. 162, 1909 ;
F. Ageno, Atti Accad. Lincei, (5), 19. i, 381, 1910 ; C. Doelter, Monatsh., 30. 179, 1909 ; E. B,
Spaer, Journ. Amer. Chem. Soc, 30. 195, 1908; G. Lemoine, Compt. Rend., 144. 357, 1908;
J. H. Kastle and A. S. Loevenhart, Amer. Chem. Journ., 29.397, 563, 1903; G. Phragmen,
Med. Vet Nobel-Inst., 5. 22, 1919.
5 A. Bechamp, Compt. Rend., 94. 1720, 1882 ; 95. 925, 1882 ; P. Bergengriin, Centrb. Physiol,
689, 1889 ; R. Robert, Pfliiger's Archiv., 82. 603, 1900 ; C. Liebermann, ib., 104. 176, 201, 1904 ;
J. Ville and J. Moitessier, Bull Soc Chim., (3), 27. 1003, 1902 ; S. Cotton, ib., (3), 25. 255, 1901 ;
A. Jolles and M. Oppenhein, Virchow's Archiv., 180. 185, 1905 ; A. Gottstein, ib., 133. 295,
1893 ; G. Senter, Zeit. phys. Chem., 44. 257, 1903 ; 51. 673, 1905 ; Proc Roy. Soc, 74. 201,
1904 ; A. S. Loevenhart, Amer. Journ. Physiol, 13. 171, 1905 ; W. Lob and P. Mulzer, Biochem.
Zeit., 13. 475, 1908 ; M. E. Pozzi-Escott, Bull Assoc Chim. Sucr. DisL, 21. 1247, 1904 ; C. F.
Schonbeiu, Journ. prakL Chem., (1), 75. 79, 1858 ; (1), 78. 90, 1859 ; G. Hiifner, ib., (1), 10. 156,
1837 ; H. St. C. Deville and H. Debray, Compt. Rend., 78. 1782, 1874 ; F. Hoppe-Seyler, Zeit.
physiol Chem., 5. 395, 1881 ; 11. 566, 1887 ; G. Lockemann, J. Thies, and H. Wichem, ib., 58.
390, 1909; P. Wantig and 0. Steche, ib., 74. 101, 1911 ; 76. 177, 446, 1912; 83. 315, 1913;
Biochem. Zeit, 60. 463, 1914 ; 0. Sulc, Zeit. phys. Chem., 28. 719, 1899 ; A. Herlitzka, Atti
Accad. Lincei, (5), 15. ii, 333, 1906.
« F. L. Usher and J. H. Priestley, Proc Roy. Soc, 77. 369, 1906 ; A. HefiEter, Med. Nat.
Archiv., 1. 81, 1907 ; E. J. Lesser, Zeit Biol, 49. 571, 1907 ; J. Dewitz, Zentr. Physiol, 22. 145,
1908; K. Togami, Berlin Klin. Wochen., ^. 1528, 1908; A. Bach, Ber., S9. 1664,1906;
H. van Laer, Bull Soc Chim. Belg., 19. 337, 1906 ; Zentr. Bakleriol, 17. 546, 1907 ; A. Renard,
Monit. ScienL, (4), 18. 39, 1904 ; J. J. Ford, Journ. Soc Chem. Ind., 23. 414, 1904.
' P. Sabatier, Bull Soc Chim., (2), 44. 169, 1885 ; C. T. Kingzett, Journ. Soc Chem. Ind.,
9. 3, 1890 ; G. E. Davis, Chem. News, 49. 226, 1884 ; W. Clayton, Trans. Faraday Soc, 11.
164, 1916.
8 M. Lorenzen, Pharm. Centrhatle, 47. 478, 1906.
» A. Downes and F. P. Blunt, Nature, 20. 521, 1879 ; R. F. d'Arcy, Phil Mag., (6), 3. 42,
1901.
10 H. Thiele, ZeiL angew. Chem., 22. 2472, 1907 ; Ber., 40. 4914, 1907 ; A. Tian, Compt.
Rend., 151. 1040, 1910 ; V. Henri and R. Wurmser, ib., 157. 126, 284, 1913 ; J. H. Mathews and
H. A. Curtis, Journ. Phys. Chem., 18. 166, 521, 1914 ; Journ. Amer. Chem. Soc, 42. 720, 1920.
11 H. J. H. Fenton, Proc Cambridge Phil Soc, 12. 424, 1904.
12 0. F. Wieder, Ber., 31. 516, 1898 ; R. Willstatter, ib., 36. 1828, 1903 ; S. Tanatar, ib., 32.
1544, 1909 ; H. A. Curtis, Journ. Amer. Chem. Soc, 42. 720, 1920.
13 R. Wolffenstein, Ber., 27. 3311, 1911 ; R. de Forcrand, Compt. Rend., 130. 1630, 1900.
1* T. Fairley, Journ. Chem. Soc, 31. 1, 125, 1875; E. Schone, Liebig's Ann., 196. 239, 1879 ;
E. Lenssen, Journ. prakL Chem., (1), 81. 276, 1860 ; H. Aschoff, ib., (1), 81. 487, 1860.
16 J. Sherber, Schweiss. Apoth. Ztg., 52. 245, 1914.
i« J. Brode, ZeiL phys. Chem., 37. 257, 1901 ; M. Traube, Ber., 17. 1062, 1884 ; P. Planes,
Journ. Pharm. Chim., (6J, 20. 538, 1904 ; G. Magnanini, Gazz. Chim. Ital, 19. 476, 1891.
1' G. Lunge, Ber., 19. 868, 1886 ; H. Aschoff, Journ. prakL Chem., (1), 81. 487, 1860.
18 S. Tanatar, ZeiL anal Chem., 28. 255, 1889 ; Ber., 32. 1015, 1899.
19 N. Schelow and A. Pudofkin, Zeit. Elektrochem., 16. 125, 1910 ; E. Abel, ib., 14. 630, 1908 ;
E. Orloff, Journ. Russian Phys. Chem. Soc, 45. 489, 511, 1913 ; J. Wolf and E. de Stocklen,
CompL Rend., 146. 1415, 1908 ; G. Magnanini, Gazz. Chim. Ital, 19. 476, 1891 ; A. P. H. Trivelli,
Chem. Weekb., 6. 525, 1909 ; G. Meissner, Untersuchungen uber den Sauerstoff, Hanover, 1863 ;
0. Loew, ZeiL Chem., 609, 1870 ; M. Berthelot, CompL Rend., 90. 1863 ; H. Struve, ZeiL anal
Chem., 8. 317, 1869; 11. 28, 1873; E. Schone, Liebig's Ann., 195. 228, 1879; Ber., 13. 627,
1880 ; Chem. News, 43. 149, 249, 1881 ; C. T. Kingzett, ib., 43. 161, 278, 1881 ; E. Pochard,
CompL Rend., 128. 1101, 1899; 130. 1705, 1900; V. Auger, ib., 152. 712, 1911; 153. 1005,
1911; S. Tanatar, Ber., 32. 1013, 1899; M. Traube, ib., 17. 1062, 1884; A. Bach, ib., 37.
3785, 1904 ; G. Bredig and J. H. Walton, ZeiL Elektrochem., 9. 114, 1903 ; ZeiL phys. Chem.,
47. 185, 1904; G. Bredig, ib., 48. 368, 1904; E. Schone, Liebig's Ann., 195. 228, 1879;
M. Kleinstuck, Ber., 51. 108, 1918 ; J. H. Walton, Zeit. physiol Chem., 47. 185, 1901 ; J. Brode,
ib., 37. 257, 1901.
20 H. T. Calvert, Ann. Physik, (4), 1. 483, 1900 ; W. J. Russell, Proc Roy. Soc, 64. 409,
1899 ; L. Graetz, Phys. ZeiL, 2. 688, 1904 ; Ber. deuL phys. Ges., 3. 78, 1905 ; J. Precht and
C. Otsuki, ib., 3. 53, 163, 1905 ; ZeiL phys. Chem., 52. 236, 1905 ; Ann. Physik, (4), 16. 890, 1905 ;
W. Merckens, «6., (4), 16. 667, 1905; ZeiL angeiv. Chem., 18. 489, 1905; A. Kufferath and
W. Merckens, ib., 17. 1095, 1904 ; Stockert, ib., 17. 1671, 1904 ; Ann. Physik, (4), 17. 192, 1905 ;
E. van Aubel, CompL Rend., 138. 961, 1904 ; 0. Dony-Henault, Trav. Lab. VInsL Solvay, 6. 13,
1903 ; K. Kof and H. Hahn, ZeiL phys. Chem., 60. 367, 1907 ; S. Saeland, Ann. Physik, (4), 26.
948 INORGANIC AND THEORETICAL CHEMISTRY
899, 1908 ; 0. and A. Dony, BuU. Chim. Soc. Belg., 22. 224, 19Q8 ; M. Padoa, Atti Accad. Lincei,
(5), 14. ii, 43, 1905 ; C. Otsuki, Jmtrn. Soc. Chem. Ind., 24. 575, 1905.
" G. Schunck, Monntsh., 37. 489, 1916.
*> C. Engler and A. Nasse, Liebig's Ann., 154. 215, 1870; T. Fairley, Journ. Chem. Soc, 31.
1, 125, 1877 ; Ch^m. News, 33. 237, 1876 ; 62. 227, 1890 ; E. Abel, Zeit. Elektrochem., 13. 555,
1907 ; 18. 705, 1912 ; 19. 477, 1913 ; N. Tarugi and H. Vitali, Oazz. Chim. Ital, 39. i, 418,
1909 ; S. Tanatar and E. Burkser, Journ. Russian Phys. Chem. Soc, 45. 1, 1913 ; L. Marino,
Zeit. anorg. Chem., 65. 26, 1909 ; C. Weltzien, Liebig'a Ann., 138. 129, 1866 ; A. Besson, Collegium,
193, 1907.
2» A. Gutbier, Zeit. anorg. Chem., 40. 260, 1904.
»* A. von Baeyer and V. Villiger, Ber., 33. 124, 1900 ; T. S. Price, ib., 35. 291, 1902.
25 A. von Baeyer and V. Villiger, Ber., 34. 755, 1901 ; G. Lunge, ib., 19. 868, 1886 ; W. Weith
and A. Weber, ib., 7. 1745, 1874.
2« P. Melikoff and I,. Pissarjewsky, Journ. Russian Phys. Chem. Soc, 30. 475, 1898.
" S. Tanatar, Ber., 32. 241, 1899.
" C. Wurster, Ber., 20. 2631, 1887.
2« T. Weyl, Ber., 39. 1307, 1906.
" F. Raschig, Ber., 18. 2743, 1885 ; K. Hasebrock, ib., 20. 213, 1887.
" V. Macri, Boll. Chim. Farm., 56. 417, 1917; A. Riche, Bull. Soc Chim., (1), 2. 178, 1800 ;
E. Schone, Liebig's Ann., 192. 257, 1898 ; Ber., 13. 623, 1880 ; M. Berthelot, Cornpt. Rend., 90.
334, 1880; E. Drechsel, Journ. prakt Chem., (2), 18. 303, 1878; G. Tammann, Zeit. physiol.
Chem., 4. 441, 1889.
»2 T. Fairley, Joiirn. Chem. Soc, 31. 1, 1877; W. Eichholz, Zahniirzt. Ztg., 13. 1, 1914;
S. Droste, Chem. Ztg., 37. 1317, 1913; E. Salkowsky, Chem. Ztg., 40. 448, 1916; C. Weltzien,
Liebig's Ann., 138. 129, 1866 ; H. B. Baker and L. H. Parker, Journ. Chem. Soc, 103. 2060, 1913 ;
T. Okaya, Proc Phys. Math. Soc. Japan, (3), 1. 283, 1919.
33 M. Berthelot, Ann. Chim. Phys., (5), 21. 164, 1880 ; (7), 11. 217, 1897; (7), 23. 62, 1901 ;
(7), 25. 78, 1902 ; D. Mcintosh, Journ. Phys. Chem., 6. 15, 1902 ; L. J. Thenard, Ann. Chim. Phys.,
(2), 9. 316, 441, 1818 ; B. C. Brodie, Proc Roy. Sec, 11. 442, 1861 ; W. Manchot, Ber., 42. 3948,
1909 ; A. von Baeyer and V. Villiger, ib., 4. 743, 2769, 1901 ; E. M'llder, Rec Trav. Chem.
Pays-Bas, 22. 388, 1903 ; T. Bayley, PhU. Mag., (5), 7. 126, 1879 ; T. Fairley, Journ. Chem. Soc,
31. 1, 125, 1877; M, Kimura, Mem. Coll. Eng. Kyoto Unit:, 5. 253, 1913; V. Kohlschiilter and
E. Eydmann, Liebig's Ann., 398. 26, 1913; C. Weltzien, ib., 142. 105, 1866; T. Bayleigh, Phil.
Mag., (5), 7. 126, 1879.
3* 0. Masson, Journ. Chem. Soc, 91. 1449, 1907.
'6 E. Schone, Zeit. anal. Chem., 9. 41, 1870 ; A. Weller, Ber., 15. 2592, 1882 ; P. Melikoff
and L. Pissarjewsky, ib., 30. 2902, 1897 ; 31. 953, 1898; Journ. Russian Phys. Chem. Soc, 30.
479, 1898 ; P. Melikoff and P. Kasanecky, Zeit. anorg. Chem., 28. 242, 1901 ; G. Deniges, Compt.
Rend., 110. 1007, 1890 ; E. Pechard, Ann. Chim. Phys., (6), 28. 536, 1893 ; E. Knecht and
E. Hibbert, Ber., 38. 3318, 1905; A. Piccini, Gazz. Chim. Hal., 12. 151, 1882; 13. 57, 1883 ;
E. Jackson, Chem. News, 47. 157, 1883 ; T. Fairley, ib., 62. 227, 1890 ; J. F. Aloy, Bull. Soc Chim.,
(3), 27. 735, 1902 ; J. R. Cain and J. C. Hostetter, Journ. Amer. Chem. Soc, 34. 274, 1912.
3« P. T. Cleve, BuU. Soc Chim., (2), 43. 53, 1885; S. Tanatar, Zeit. anal. Cliem., 28. 255,
1889.
37 E. Schone, Liebig's Ann., 196. 58, 1879; T. GigU, Chem. Ztg., 17. 186, 1893; V. Zotier,
Bull. Soc Chim., (4), 13. 61, 1903 ; (4), 15. 402, 1914 ; (4), 21. 241, 1917 ; T. Bayley, Phil. Mag.,
(5), 7. 126, 1879.
38 G. Bredig and A. von Antropoff, Zeit. Elektrochem., 12. 581, 1906; A. von Antropoff,
Journ. prakt. Chem., (2), 77. 273, 1908 ; W. R. E. Hodgkinson and A. H. Coote, Chem, News,
92 38 1905
3» T. Bayley, Phil. Mag., (5), 7. 126, 1879 ; J. Sherber, Schweiz. Apoth. Ztg., 53. 717, 1915.
" C. F. Schonbein, Journ. prakt. Chem., (1), 79. 67, 1860; (1), 81. 276, 1860;
(1), 92. 150, 1863; Zeit. anal. Chem., 1. 10, 442, 1862; 4. 116, 1865; E. Lenssen, Journ.
prakt. Chem., (1), 81. 276, 1860 ; C. Weltzien, Liebig's Ann., 138. 133, 1866 ; I. Quincke, Zeit.
anal. Chem., 31. 1, 1892 ; M. Prud'homme, Bull. Soc Chim., (3), 29. 1010, 1903 ; H. Colin and
A. Sen6chal, Compt. Rend., 153. 76, 1911 ; E. Schone, Ber., 7. 1693, 1874 ; W. Wobbe, Apoth.
Ztg., 18. 458, 465, 487, 1903 ; W. Manchot and 0. Wilhelm, Liebig's Ami., 325. 105, 1903.
*i M. Prud'homme, Bull. Soc Ind. Mulhouse, 73. 294, 1904; E. S. Barralet, Chem. News,
79. 136, 1899 ; P. Melikoff and J. Pissarjewskv, Ber., 30. 2902, 1897 ; T. Bayley, Phil. Mag.,
(5), 7. 126, 1879 ; G. Watson, Chem. News, 46. 9, 1882.
*2 V. Macri, Boll. Chim. Farm., 56. 417, 1916 ; W. Stiidel, Zeit. angew. Chem., 15. 642, 1902.
*3 R. de Forcrand, Compt. Rend., 134. 601, 1902; W. Stadel, Zeit. anorg. Chem., 15. 642,
1902 ; B. Kuriloff, Chem. Ztg., 14. 114, 1890.
** E. Lenssen, Journ. prakt. Chem., (1), 81. 276, 1860; C. F. Schonbein, ih., (1), 81. 276,
1860; M. Martinon, Bull. Soc Chim., (2), 45. 862, 1886; M. Berthelot, Compt. Rend., 108. 24,
167, 477, 1889 ; H. Moissan, ib., 97. 96, 1883 ; L. C. Barreswill', ib., 16. 1085, 1843 ; Ann. Chim.
PhyM., (3), 20. 364, 1847 ; B. C. Brodie, Proc Roy. Soc, 11. 442, 1861 ; 0. F. Wiede, Ber., 30.
2178, 1897; 31. 516, 1898; V. Macri, Boll. Chim. Farm., 56. 417, 1916; E. Pechard, Compt.
Rend., 113. 39, 1891 ; C. Hausermann, Journ. prakt. Chem., (2), 48. 70, 1893 ; K. A. Hofmann
and H. Hiendlmaier, Ber., 37. 1663, 3405, 1904 ; 38. 3059, 3066, 1905 ; E. H. Riesonfeld, H. E.
OZONE AND HYDROGEN PEROXIDE 949
Wohlers, and W. A. Kutsch, ib., 38. 105, 1905 ; E. H. Riesenfeld, ih., 38. 3380, 3578, 4068, 1905 ;
44. 147, 1911 ; Zeit. anorg. Chem., 74. 48, 1912 ; E. Spitalsky, ih., 53. 184, 1907 ; 56. 72, 1907 ;
Journ. Russian Phys. Chem. Soc, 42. 1085, 1910 ; Ber., 43. 3187, 1910.
45 B. C Brodie, Phil Trans., 140. 769, 1850; Proc. Roy. Soc, 11. 442, 1862; Journ, Chem.
Soc, 16. 320, 1863 ; C. Weltzien, Liebig's Ann., 138. 133, 1866 ; G. F. Soh6nbein,ZetY. anal. Chem.,
1. 12, 1862: I.. Swiontkowskv, Liebig's Ann,, 141. 205, 1867; P. Thenard, CompL Rend., 75.
177, 1872 ; A. Gorgeu, ib., 110, 958, 1890 ; M. Berthelot, Ann. Chim. Phys., (5), 21. 176, 1880 ;
A. von Baeyer and V. Villiger, Bet., 33. 2488, 1909 ; A. Bach, ib., 34. 3851, 1910 ; L. Marino,
Zeit. anal. Chem., 65. 25, 1909 ; C. Porlezza and G. Norzi, Atli Accad. Lincei, (5), 22. i, 238,
1913 : E. Schone, Liebig's Ann., 196. 58, 1879 : T. Bayley, Phil. Mag., (6), 7. 126 1879 ;
M. Martinon, Bull Soc Chim., (2), 43. 355, 1885 ; R. Engel, ib., (3), 6. 17, 1891 ; T. Gigli, Chem.
Ztg.,n. 186,1893.
48 M. Berthelot, Ann. Chim. Phys., (5), 21. 176, 1880.
47 A. Bach, Ber., 33. 1506, 3111, 1900 ; 34. 3851, 1901 , 35. 158, 1902.
" H. E. Armstrong, Proc Chem. Soc, 16. 134, 1900 ; W. Ramsay, Journ. Chem. Soc, 79. 1224,
1901 ; A. von Baeyer and V. Villiger, Ber., 33. 2488, 1900 ; A. M. Clover, Amer. Chem. Journ.,
29. 463, 1903.
4» M. Kleinstuck, Ber., 51. 108, 1918.
§ 12 The Qualitative and Quantitative Determination of Ozone and
Hydrogen Peroxide
The detection of ozone and hydrogen peroxide is complicated by the fact that
while many reagents produce marked colorations with these two substances, yet
other substances like nitrogen peroxide, chlorine, and bromine give similar colora-
tions, so that the results with this group of reagents are merely characteristic of an
oxidizing gas.
In 1842, C. F. Schonbein ^ used test papers soaked in a solution of starch and
potassium iodide ; these are coloured blue by ozone. The iodine must be free from
iodate or a blue colour will be obtained by the carbon dioxide which decomposes the
iodate.2 Even in the absence of iodates, the blue coloration is produced by other
oxidizing agents — chlorine, nitrous acid, vapours of ethereal oils, hydrogen peroxide,
etc. 3 Ozone with potassium iodide gives potassium hydroxide and iodine, and it
is the latter which gives the blue coloration with starch. Attention has been directed
to mixing reagents with the potassium iodide which are sensitive to the alkali instead
of the iodine. Thus, A. Houzeau (1868) ^ made test papers by soaking them in a
mixture of potassium iodide and wine-red litmus. These papers are coloured blue
by ozone, ammonia, etc. A. R. Leeds recommended phenolphthalein in place of
litmus ; C. Arnold and C. Mentzel tried rosolic acid, and also fluorescein. The
former gives a red colour with ozone, the latter, with black paper, gives a green
fluorescence. Neither chlorine nor nitrogen oxide gives the red with rosolic acid.
While it is possible to prepare an ozone paper in this way which is not affected by
chlorine or nitrogen oxide, hydrogen peroxide gives the same reactions as ozone. ^
This remark also applies to the substitution of cadmium or zinc iodide in place of
potassium iodide. Nevertheless, the methods in use for the determination of ozone
and hydrogen peroxide are based on the action of these agents upon a neutral solution
of potassium iodide. The results are satisfactory provided other oxidizing agents
are absent.
In the quantitative determination of ozone, a known volume of air is drawn
through such a neutral solution of potassium iodide, when the ozone liberates iodine :
034-2KI+H20=02+l2+2KOH. The liberated iodine is determined by acidifying
the solution, and titrating with standard sodium thiosulphate, Na2S203.5H20.
The reaction is symbolized : l2+2Na2S203=2NaI+Na2S406 ; and accordingly,
every gram-molecule of sodium thiosulphate corresponds with a gram-atom of
iodine, which in turn corresponds with half a gram-molecule of ozone.
Example. — Assuming that ^N sodium thiosulphate solution contains the ^V*^ gram-
molecule of the crystallized salt, NagSgOg-oHoO, per litre, the above equations show that
952 INORGANIC AND THEORETICAL CHEMISTRY
I^eeds, Chcm. News, 38. 224, 1878 ; C. T. Kingzett, ib., 38. 249, 1878 ; C. Arnold and C. Mentzel,
Ber., 35. 1324,1902.
« A. Houzeau, Ann. Chim. Phys., (4), 27. 5, 20, 1872 ; Compt. Rend., 66. 44, 1868 ; F. S. Cloez,
i6., 52. 527, 1861 ; E. Fremy, ih., 70. 61, 1870 ; C. Arnold and C. Mentzel, Ber., 25. 1324, 1892 ;
F. Emich, Monta^h., 22. 670, 1901 ; A. R. Leeds, Chem. News, 38. 243, 1878 ; 43. 161, 1881.
' C. Baskerville and W. A. Hamor, Journ. Ind. Eng. Chem., 3. 378, 1911 ; W. Crozier, Journ.
Amer. Chem. Soc., 34. 1332, 1912; A. R. Leeds, CJiem. News, 38. 224, 1878; R. Bottger, Pol.
Notizhlatt, 35. 95, 1880.
• C. F. Schonbein, Joiirn. prakt. Chem., (1), 105. 219, 1867 ; (1), 53. 69, 1851 ; Pogg. Ann.,
73. 490, 1848; E. Schaer, Ber., 3. 24, 1870; E. Schone, Zeit. anal. Chem., 33. 155, 1894;
A. BoUand, ih., 46. 621, 1907 ; L. I. de N. Ilosva, Bull. Soc. Chim., (3), 2. 347, 1889.
' C. J. Lintner, Journ. prakt. Chem., (2), 34. 378, 1886 ; C. Faulenbach, Zeit. physiol. Chem.,
7. 510, 1882.
8 C. Wurster, Ber., 19. 3195, 1886 ; 21. 921, 1525, 1888.
• L. I. de N. Ilosva, Bull. Soc. Chim., (3), 2. 351, 1889 ; G. V. Chlopin, Zeit. Unters. Nahr.
Genuss., 5. 504, 1902 ; G. Erlwein and T. Weyl, Ber., 31. 3158, 1898 ; C. Arnold and C. Mentzel,
ih., 35. 2902, 1902 ; 39. 1528, 1906 ; F. Fischer and H. Marx, ih., 39. 2555, 1906 ; K. W. Charit-
schoff, Chem. Ztg., 34. 60, 1910.
10 E. H. Kaiser and I^. McMaster, Amer. Chem. Journ., 39. 96, 1908 ; E. Schone, Ber., 7.
1693, 1874 ; C. Engler and W. Wild, ib., 29. 1940, 1896 ; W. Wobbe, Apoth. Ztg., 18. 458, 465,
487, 1903 ; C. F. Schonbein, Journ. prakt. Chem., (1), 79. 67, 1860.
11 C. Engler and W. Wild, Ber., 29. 1940, 1896 ; H. McLeod, Chem. News, 40. 307, 1879 ;
E. H. Kaiser and L. Me Master, Amer. Chem. Journ., 39. 96, 1908.
§ 13. The Composition and Constitution of Hydrogen Peroxide
Rational chemical formulae are a kind of contracted equation ; a compound may have
several rational formulae and that one is best which expresses the greatest number of
reactions.' — C. Gtcbhardt (1856).
In spite of the fact that the composition and molecular weight of hydrogen
peroxide have been determined ; in spite of the simplicity of the resulting formula,
H2O2 ; and in spite of the many ingenious (sometimes far-fetched) arguments
which have been deduced from experiments made to determine the relative dis-
position of its component atoms, the constitution of this compound has not been
yet established by unequivocal experiments.
The empirical formula. — L. J. Thenard (1818) introduced a weighed amount of
the peroxide in a small vial into a graduated cylinder over mercury. The vial
was broken and its contents decomposed either by introducing manganese dioxide,
or by heat ; 17 parts of hydrogen peroxide by weight gave nearly 8 parts by weight
of oxygen, and 17—8=9 parts by weight of water. Otherwise expressed, 34 parts
of hydrogen peroxide give 18 parts of water and 16 parts of oxygen. Hence, the
peroxide contains hydrogen and oxygen in the proportion of 2 atoms of oxygen.
The simplest formula for hydrogen peroxide is therefore HO. There is here
nothing to show whether HO or some multiple of HO, say, HnO^, is the proper
formula for the compound, since the latter has the same percentage composition
as the former.
The molecular formula of hydrogen peroxide. — The instability of hydrogen
peroxide prevents a determination of its vapour density being made in the regular
manner. The molecular weight has been determined by the freezing-point method. ^
The result is nearly 34. This agrees with the formula H2O2 — the generally
accepted value.
The constitutional or graphic formula of hydrogen peroxide — The evidence
for the constitution of hydrogen peroxide deduced from its reactions, or from the
reactions of the analogous peroxides, is somewhat ambiguous because different
lines of argument lead to different conclusions in spite of very positive assertions in
favour of particular formulae. One group of evidence favours the formula HO. OH
with both oxygen atoms bivalent ; another group favours HO : OH with both
oxygen atoms quadrivalent ; and still a third group favours H2 : 0 : 0 with one
OZONE AND HYDROGEN PEROXIDE 953
oxygen atom bi- and the other quadri-valent. The two formulae HO. OH and
HO : OH are similar in type, and with our present knowledge, it does not matter
very much which be favoured ; the evidence for HO : OH against HO. OH is
mainly physical, and special weight was given to the former by J. W. Briihl (1895).
M. Traube (1893) has also argued for a similar constitution. A. Bach (1900),
following C. T. Kingzett (1882), favours H2=0=0 ; and a similar type of formula
has been advocated in order to emphasize the analogy between hydrogen peroxide
— assumed to be H2=0=0 — and ozone — assumed to be 0=0=0. When two
lines of arguments lead to two independent formulae, each of which seems highly
probable if the other be ignored, some one is almost sure to suggest that both
formulae are right, and that one formula represents the equilibrium state under one
set of conditions, and likewise also for the other formula under another set of
conditions. There is thus a labile intra-molecular change from one form to the
other : H2=0=0^H0 : OH. This type of chemical change has been called
tautomerism. 0. Mumm (1907) ^ has assumed that hydrogen peroxide is an illustra-
tion of the phenomenon ; and E. Bose suggested in 1901, that hydrogen peroxide
can exist in two different forms in one of which it acts as an oxidizing agent, and in
the other as a reducing agent. Otherwise expressed, an acid solution favours the
H2=?0=0 formula, and an alkaline solution the HO=OH formula.
1. Evidence for HO— OH or H.O.O.H.—L. Carius (1863) noticed that ethylene,
C2H4, unites directly with hydrogen peroxide to form sehr klein Mengen of ethylene
glycol, C2H4(0H)2, presumably by the equation :
H2=C , 0H_. H2=C-0H
H2=C'^6h"^H2=C-OH
This does not conclusively prove that the constitutional formula is HO. OH since
the glycols are similarly formed by oxidizing one of the olefine series (of which
ethylene is a member) with potassium permanganate. Similarly, the formation of
hydroxides by the action of hydrogen peroxide on metals like zinc :
Zn+H202=Zn<^][^
and on sulphur dioxide forming S02(0H)2. Here the hydrogen peroxide enters
into union as 20H'. The inference that hydrogen peroxide is accordingly
constituted HO. OH is inclusive, because the hydroxides are formed by other
oxidizing agents not containing hydroxyl groups.
The hydrogen of hydrogen peroxide can be indirectly replaced by the ethyl
(C2H5) or benzoyl. (CeHsCO) radicle, to form the corresponding peroxides — viz.
ethyl peroxide, (C2H5)202, and benzoyl peroxide, (C6H5CO)202. It has been
argued that if ethyl peroxide has the formula (C2H5)2 : 0 : 0, it should furnish ethyl
oxide (ether), (€2115)20, when exposed to the reducing action of nascent hydrogen :
(C2H5)202+2H->(C2H5)20+H20 ; and if benzoyl peroxide be constituted
(C6H5CO)2: 0 : 0, it should furnish benzoic anhydride, (C6H5CO)20, under similar
conditions. As a matter of fact, A. von Baeyer and V. Villiger (1900) obtained
neither ethyl oxide nor benzoic anhydride when the respective peroxides were
treated with platinum and hydrogen in the cold ; the actual products were
respectively ethyl alcohol, C2H5OH, and benzoic acid, C6H5.CO.OH. Hence, it is
argued that the reactions are of the type (C2H5)202+2H=2C2H50H, unless a
tautomeric alteration in the relations of the oxygen atoms in the molecule occurs
during the reduction, and it followed that the constitution of ethyl peroxide is
C2H5.O.O.C2H5, and of benzoyl peroxide, CgHsCO.O.O.COCeHs ; and by analogy it
is argued that hydrogen peroxide is probably constituted HO. OH, and not H2 : 0 : 0.
The view that hydrogen peroxide contains two hydroxyl groups is further
supported by the fact, emphasized by M. Traube (1893), that hydrogen peroxide is
formed in many reactions involving the reduction of ox}^genH.H+0 : 0->H0 i OH,
and not by the oxidation of water H2 : 0-l-0->H2 : O : 0 as might be expected if
954 INORGANIC AND THEORETICAL CHEMISTRY
H2 : 0 : 0 represented the formula of this compound. In M. Traube's experiment,
when the electrodes, during the electrolysis of acidulated water, were separated by
a porous cell, no hydrogen peroxide could be detected in the electrolyte ; but when
air was bubbled about the cathode, the hydrogen appears to unite with molecular
oxygen to form hydrogen peroxide. Consequently, he considers that hydrogen
peroxide cannot be regarded as oxidized water since it is never formed as a product
of oxidation, but is always formed by the reduction of molecular oxygen. He
considers the peroxide to be formed by the coupling together of molecular oxygen
and molecular hydrogen, and the oxygen which is given off in reactions with
hydrogen peroxide is not formed from atomic oxygen, but the oxygen atoms,
already paired in the hydrogen peroxide molecule, are liberated. With silver oxide,
for example, the reaction is not altogether catalytic, for part at least of the silver
oxide is reduced by the hydrogen of the peroxide — vide silver oxide.
Ag2:0+H^i02->2Ag+H20+02 ; not Ag2iO+Oi^->2Ag+02+^
Most of the reactions of hydrogen peroxide are similar reduction processes, while
if hydrogen peroxide were built of two hydroxyl groups, it would rather act as an
oxidizing agent. In all its decompositions, said M. Traube, hydrogen peroxide
gives off molecular oxygen. M. Traube rather favoured the formula H.O : O.H,
oxygen tervalent, but this is usually altered to H.O. O.H, oxygen bivalent, or to
H.O : O.H, oxygen quadrivalent.
W. Spring (1895) considers the fact that hydrogen peroxide viewed in thick
layers is rather more deeply coloured than water agrees with the view that it contains
molecular oxygen, and he tries to give this rather feeble argument more weight by
pointing out that ammonium iodide, NH4I, is colourless while the tri-iodide, NH4I.I2,
is green, and the pentaiodide, NH4I.I4, is violet. W. Spring also considers that
the specific heat determinations of the elements H2 and O2 show that only part
of their available free energy is used in forming hydrogen peroxide, and hence,
the molecule has probablv a more complex linking than is shown by the simple
formula H.H+0 : O^HO'^OH.
J. W. Briihl (1896) determined the index of refraction and specific gravity of
hydrogen peroxide purified by R. Wolff enstein's process, and found the molecular
refraction R for the spectrum lines Ha and Hy to be JRa=5*791 and Ry=5'S17 ;
hence the molecular dispersion jRy— 72a=0'136. If corresponding constants for
water and for hydrogen atom be subtracted, a value for hydroxyl OH is obtained :
HOH ....
H
HO . *. . . . 2-59 2-66 0-05
Doubling the value for OH to get the value for HO. OH, there follows the calculated
values : Ra=5;lS, Ry=5'32, and Ry—Ra=0'10. The actual spectrometric
constants are therefore larger than those calculated for the compound HO. OH, and
since the additive rule here employed has been applied to a large number of com-
pounds, the difference is explained by assuming that the two oxygen atoms of hydrogen
peroxide are joined by multiple bonds. Again, the sum of the optical constant of
an atom of oxygen, such as occurs in the molecule of water, is less than the observed
constant for an equivalent of molecular oxygen, and the dispersion, Ry—Ra, of
molecular oxygen is double that for the equivalent oxygen in water :
^«
Ry
Ry-H,
3-69
3-71
0-09
MO
1-05
0-04
Ra
Ry
Ry-Rc
20 in water .
.
. 2-968
3-212
0-036
Molecular oxygen
liquid
gaseous
. 3-958
3-964
4-09
0-069
Again, fewer bonds are supposed to be concerned in uniting the atoms of molecular
Ra
Ry
Ry-Rc
2-968
3-212
0-036
3-591
3-717
0-055
3-958
3-964
0-069
. —
4-09
. —
OZONE AND HYDROGEN PEROXIDE 955
oxygen than is the case with the oxygen atoms in the molecule of hydrogen peroxide,
because the optical constants calculated for hydrogen peroxide are rather less than
for molecular oxygen :
Oxygen in water
Oxygen in hydrogen peroxide
Oxygen molecule {gXeous
M. Traube assumed oxygen to be tervalent in hydrogen peroxide, HO : OH, but
there is no satisfactory evidence to warrant this assumption. The oxygen group
of elements are bi- or quadri-valent, and there is much circumstantial evidence
which warrants the assumption of quadrivalent oxygen. The multiple-bonded
oxygen in hydrogen peroxide is therefore based on a quadrivalent oxygen, and the
formula is written HO [ OH, or HO^OH.
J. W. Briihl ^ assumes that the oxygen in water has two latent valencies, H'O'H,
which are the cause of the tendency of water to form associated molecules. Indeed,
E. Beckmann * found that all associated liquids are of the water type. Thus, the
alcohols and the fatty acids are usually associated,^ and J. T. Hewitt and T. F.
Winmill attribute the association of phenols, R.OH, in the liquid condition to the
coupling of the molecules in virtue of the residual valencies of the contained oxygen.
J. W. Briihl explains the cause of the great ionizing power of water as follows :
Since hydrogen peroxide has even more unsaturated valencies than water, he inferred
that hydrogen peroxide must possess a great power of ionization — three-quarters
perhaps even greater than that of water. The ionizing power is not easily determined
because of the great tendency of hydrogen peroxide to decomposition. Owing to
the parallelism between dielectric constants and ionizing power, it also follows that
the dielectric constant of hydrogen peroxide will be high. This has been demon-
strated by H. T. Calvert, 6 who found the dielectric constant to be 92*8 (18°), when
the value for water is 81 (18°). Again, according to P. Drude,^ all hydroxyl
compounds, with the exception of water, show an anomalous electrical absorption
in that they are poor conductors of electricity and yet they absorb electromagnetic
waves of short wave-length (70 cm.)— normally, substances which conduct electricity
moderately well are non-absorbent. H. T. Calvert also found that hydrogen
peroxide does not show this phenomenon of anomalous absorption. Hence, this
compound is not to be regarded as a di-hydroxyl HO. OH, but, preferably, as a
compound H.OiO.H, "analogous with acetylene, and like acetylene it is formed
endothermally, and is explosive."
2. Evidence for Hq,=0=0. — A. Bach's formula H2=0=0 is regarded by some
as " undoubtedly representing the structure of the molecule correctly." The central
quadrivalent oxygen is considered to be so heavily loaded with atoms that it readily
parts with the two hydrogen atoms in the presence of oxidizing substances, or
gives up the extra oxygen in contact with reducing agents. In the former case,
it acts as a reducing agent, and in the latter, as an oxidizing agent. On this
assumption, the reactions of ozone with hydrogen and barium peroxides are
represented :
H2=0=OH-0=0=0->H20-f202; Ba=0=0+0=0=0->BaO+202
R. Willstatter and E. Hauenstein^ have shown that the reduction of ethyl
and benzoyl peroxides by platinum and hydrogen in the cold furnishes ethyl
alcohol and benzoic acid respectively. It is therefore argued that the formula of
benzoyl peroxide must be
CeHfi.CO.O ^ ^ C6H5.CO^^_^
CeHg.CO.O and not CcHsCO-^^"^
956 INORGANIC AND THEORETICAL CHEMISTRY
since the latter would give rise to benzoic anhydride, not the acid. A. Rius y
Miro claims that this argument is weakened by the fact that the water formed
in the reaction might hydrolyze any aldehyde formed; but the same result is
obtained by reducing the peroxide in a boiling solution by means of yellow
phosphorus. In addition, potassium orthophosphate is the sole product of the
reduction of potassium perphosphate by potassium iodide in acetic acid solution,
or by ferrous or cobalt hydroxides in alkaline solution. It is also shown that
sulphuric acid cannot be oxidized to persulphuric acid by permanganic acid,
plumbic salts, or nickel peroxide. The majority of oxidations due to hydrogen
peroxide are really hydroxylations, but, although this is contrary to the
asymmetrical formula, the case of potassium permanganate shows it does not
necessarily lead to the symmetrical formula. This formula does not explain the
reducing properties of hydrogen peroxide, attributed to weakly bound hydrogen
atoms. A. Rius y Miro accordingly suggests the formula H<x>H for hydrogen
peroxide, and proposes to call peroxides of the type RO2R, anhydrohydwper-
oxides, to distinguish them from A. Baeyer's hydrojperoxides.
References.
1 G. Tammann, Zeit. phys. Cham., 4. 443, 1899; 12. 43], 1893; G. Carrara, Atti Accad.
Lificei, (5), 1. 19, 1892; W. R. Orndorff and J. White, Amer. Chem. Journ., 15. 347,
1893
2*0. Mumm, ZeAt. phys. Chem., 59. 459, 492, 497, 1907 ; E. Bose, ib., 38. 1, 1901.
» J. W. Briihl, Ber., 28. 2866, 1895.
* E. Beckmann, Zeit. phys. Chem., 6. 437, 1890.
6 W. Ramsay and J. Shields, Phil. Trans., 184. 665, 1893 ; K. Auwers, Zeit. phys. Chem., 12.
689, 1893 ; H. Goldschmidt, Zeit.Elektrochem., 10. 221, 1904 ; J. T. Hewitt and T. F. Winmill,
Journ. Chem. Soc, 91. 441, 1907.
• H. T. Calvert, Ann. Physik, (4), 1. 483, 1900 ; Zeit. phys. Chem., 38. 513, 1901 ; J. W.
Bruhl, Ber., 33. 1710, 1900.
' P. Drude, Ber., 30. 940, 1897 ; Zeit. phys. Chem., 23. 308, 1897.
« R. Willstatter and E. Hauenstein, Ber., 42. 1839, 1909; A. Rius y Miro, Helv. Chim. Acta,
3. 327, 1920.
§ 14. Peroxides and Peracids
Hydrogen peroxide has several properties in common with the acids. Eor
example, purified hydrogen peroxide — ^perhydrol — reddens blue litmus before
bleaching the colour ; its acidity can be partly neutralized by the addition of
alkaline solutions ; it can be more readily extracted from its ethereal solution by
alkaline lye than by water ; and hydrogen peroxide displaces the acid radicle from
sodium halides, silicate, borate, metaphosphate, and sulphide, and from potassium
ferro-and ferri-cyanides.i In 1895, W. Spring pointed out that hydrogen peroxide
behaves like a mono- and a di-basic acid in that it contains two hydrogen atoms which
can be replaced singly or in pairs by equivalent radicles. The substitution products
can be regarded as salts. The mono-substituted products are of the type ROOH,
where R denotes a monad radicle — elementary or compound — and the di-substituted
products are of the type ROOR, where the two R's maybe the same or different,
A. von Baeyer and V. Villiger 2 called the former hydroperoxides— ^'..7. ethyl hydro-
peroxide, C2H5.OOH, the latter peroxides— c.^r. diethyl peroxide, C2H5.OO.C2H5.
The peroxides can thus be regarded as salts of the acid, hydrogen peroxide, formed
by the action of this compound on, say, the hydroxide of the alkalies or alkaline
earths. Thus, by the action of sodium ethoxide, C2H5.0Na, on hydrogen peroxide
in alcohol solution, R. Wolff enstein^ prepared sodium hydroperoxide : C2H50Na
-hHOOH->C2H50H+NaOOH, and according to electrical conductivity methods,
OZONE AND HYDROGEN PEROXIDE 957
it is inferred that the salt is really NaOOH, and not NaOONa.H202. J. Tafel
called the salt NaOOH, sodyl hydroxide. E. Schone * prepared
H.OO. -p HOO^p
H.00>^^ jjQQ>ba
Barium. hydroperoxide Calcium hydroperoxide
by the action of an excess of hydrogen peroxide on the hydroxides of the alkaline
earths. If potassium or sodium carbonate be added to hydrogen peroxide, the
corresponding alkaline peroxide is formed and carbon dioxide is evolved, e.g. H2O2
+Na2C03->Na202+C02+H20 ; on the contrary, if the hydrogen peroxide be
added to a solution of the carbonate oxygen is evolved : 2H202+Na2C03->Na2C03
+2H2O+O2. The sodium carbonate in the latter case merely acts as a catalytic
agent. It is not at all uncommon to find reactions progressing differently according
to the way the substances are mixed together. The peroxides, in some cases, can
also be precipitated from solutions of the corresponding salts by the addition of
hydrogen peroxide. For instance, with lead acetate, PbA2, lead peroxide is formed :
PbA2-f-H202->Pb02+2HA. An excess of hydrogen peroxide with lead peroxide
furnished lead monoxide: Pb02+H202=PbO+H20+02. Hydrated peroxides
of the alkaline earths of the type Ba02.8H20 are precipitated from solutions
of the hydrated oxides by hydrogen peroxide ; and conversely, when these
peroxides are dissolved in dilute acids, the corresponding amount of hydrogen
peroxide is set free. In both cases, hydrogen peroxide behaves like an acid, and
the peroxides can accordingly be regarded as salts of hydrogen peroxide just as the
nitrates are salts of hydrogen nitrate — to wit, nitric acid. It can be added that
for similar reasons water liberates hydrogen peroxide from its assumed salts —
the peroxides — -and hence also water has been called an acid — water acid.
The organic radicles can also displace the hydrogen from hydrogen peroxide to
form corresponding hydroperoxides and peroxides. The relations of the alcohols
to the ethers and of the acids — e.g. acetic acid, CH3COOH — to the acid anhydrides —
e.g. acetic anhydride (CH3C0)20 — are analogous to the relations between the hydro-
peroxides and the peroxides — e.g.
CH2.CO.OOH CH3.CO.O
CH3.CO.6
The action of water on the acid anhydrides— e.^f. (CH3.CO)20+HOH->2CH3.COOH
— recalls the action of water on the peroxides. For instance, with acetyl peroxide,
peracetic and acetic acids are formed :
CH^!cO.O + bH "^ CH3.CO.OOH+CH3COOH
Acetyl peroxide Peracetic acid Acetic acid
with persulphuric acid, monopersulphuric acid or Caro's acid and sulphuric acid
are formed :
HO.SO2.O , H _^^ .OGH , c^o ^^H
HG.SG2.6+6h"^^^2<oh +^^2<oH
and with sodium peroxide, sodium hydroperoxide and hydroxide are formed :
^^^j^^+5jj->NaGGH+NaOH
The formation of the organic acids by the action of water on the acid chlorides
is analogous with the formation of the peracids by the action of hydrogen peroxide
on the acid chlorides : thus, CHsCO.Cl+HOOH-^CHaCO.OOH+HCl ; or HO.SO2.CI
+H00H->H0.S02.00H+HC1. Similarly, just as the acid anhydride is formed
by the action of an acid hydrate and acid chloride : CH3CO.OH+CICO.CH3
-^CHsCO.O.CO.CHs-j-HCl, so does hydrogen peroxide, and its acid chloride — viz.
958 INORGANIC AND THEORETICAL CHEMISTRY
hypochlorous acid, HOCl — form the corresponding acid anhydride, H.OO.OH,
which, being unstable, decomposes into water and oxygen :
HOOiH+Cl;OH->HOO.OH+HCl->HCl+H20+02
D. I. Mendeleefi (1881) 5 subdivided the oxides of the type RO2 into two classes
— ^the superoxides and the polyoxides — depending upon the valency of the element
united to the oxygen atoms : the one class was considered to be constituted on the
hydrogen peroxide type, the other on the condensed water type.
Superoxides, peroxides, or true peroxides. — Those oxides in which the oxygen
atom or atoms, over and above those required to form the basic oxide, are singly
linked to the metal and to the other oxygen atoms, so as to form a chain. The
valency of the metal is the same in the peroxide as in the basic oxide ; e.g.
H-0 Na-O K-O-O ^ ^O ^ ^O
H-6 Na-6 K-0-6 ^^"^6 °^ ^^<^o
Hydrogen peroxide. Sodium peroxide. Potassium tetroxide. Barium peroxide.
The superoxides have also been called peroxidates or peroxites and regarded as
salts of hydrogen peroxide, for they are supposed to be constituted like this compound.
These oxides were Schonbein's antozonides — a term now obsolete.
Polyoxides, dioxides, or pseudo-peroxides. — Those peroxides in which the
oxygen atom or atoms, over and above those required to form the basic oxide, are
doubly linked to the metal so that the valency of the metal in the dioxide is greater
than the valency of the metal in the basic oxide ; e.g.
0<^ Pb<^ Mn<^
Ozone. Lead dioxide. Manganese dioxide.
These oxides possess feeble basic or feeble acidic properties — ^possibly both.
They are supposed to be constituted on the double water type with a quadrivalent
element talang the place of the four hydrogen atoms in two molecules of water.
These oxides were Schonbein's ozonides — a term now applied to quite difierent
compounds.
The peroxides which yield hydrogen peroxide when treated with water or a
dilute acid are probably constituted like hydrogen peroxide. Thus, sodium peroxide
with hydrochloric acid gives hydrogen peroxide ; and potassium tetroxide, which
gives oxygen and hydrogen peroxide, is probably constituted on the same plan.
The polyoxides or dioxides are not usually attacked by dilute acids. Both types
with concentrated sulphuric acid evolve oxygen. The mechanism of the reaction
is probably different in the two cases. With the super- or per-oxides, hydrogen
peroxide is probably formed as an intermediate product : Ba02+H2S04=BaS04
+H2O2; followed by 2H202=2H20+02. With manganese dioxide: 2Mn02
H-2H2S04=2MnS04+2H20-f O2. Similarly with hydrochloric acid, both give
chlorine, but with the peroxides hydrogen peroxide is first formed, and this reacts
with the excess of acid forming chlorine : 2HC1+H202=2H20+Cl2 ; with the di-
or poly-oxides, on the other hand, an intermediate perchJoride can often be detected
• — with manganese dioxide, probably MnCls ; and with lead dioxide, PbC^ is
formed.
The differences in the behaviour of the true and false peroxides — typified by
Ba02 and Pb02— has prompted many hypotheses. B. C. Brodie and C. F. Schonbein
have assumed that the normal oxygen molecule contains a negatively and a positively
charged oxygen atom, that substances undergoing oxidation have a preference for
oxygen carrying one kind of charge while the oppositely charged oxygen is consumed
in a secondary reaction. The positively charged oxygen was called antozone, and
the corresponding oxides— PbOg, KMn04, etc. — were called ozonides ; the negatively
charged oxygen atom formed antozonides — e.g. Na202, Ba02, H2O2, etc. For
instance, H20+0=hydrogen peroxide; and Mn04-0=manganese dioxide. The
OZONE AND HYDEOGEN PEROXIDE 959
union of antozone 0+ with ozone 0~ gives ordinary oxygen, and such a reaction
was supposed to occur when hydrogen peroxide (an antozonide) reacts with, say,
lead peroxide (an ozonide). No direct experimental evidence can be quoted
demonstrating the existence of Schonbein's antozone.
S. Tanatar^ has suggested that the differences between the true and false
peroxides are due to differences in the thermal values of the reactions which occur
when the oxides are treated with acids. The formation of hydrogen peroxide
from water requires 23 Cals., and if the thermal value of the reaction between the
metal of the peroxide and the radicle of the acid is less than this value, no hydrogen
peroxide can be produced. The strong acids — e.g. hydrochloric acid — give hydrogen
peroxide with barium peroxide, while the weaker acids — e.g. phenol — give oxygen
but no hydrogen peroxide. Hence, the distinction between a true or false peroxide
is arbitrarily determined by the strength of the acid used in making the test.
S. Tanatar further states that appreciable amounts of hydrogen peroxide are formed
when nickel dioxide is treated with sulphuric acid, because (i) the solution liberates
iodine from potassium iodide ; and (ii) decolorizes a solution of potassium per-
manganate ; but C. Tubandt and W. Riedel ^ could not get the confirmatory tests
with chromic and titanic acids, and it is hence inferred that the liberation of iodine
is due to the formation of traces of a persulplfuric acid ; and that the bleaching of
the permanganate is not real, but rather a masking of the pink colour of the per-
manganate by the green colour of the nickel solution. The formation of hydrogen
peroxide occurs not only when hydrocyanic acid acts on nickel peroxide, but also
when nickel hydroxide is used ; the peroxide is presumably formed by the auto-
oxidation of the complex nickel cyanides which are formed. G. Pellini and
D. Meneghini ^ have shown that there are possibly two nickel peroxides, one of which
gives hydrogen peroxide when treated with acids, and the other does not. According
to S. Tanatar, the former may really be a compound of hydrogen peroxide and nickel
monoxide, and not a dioxide at all.
Attempts have been made to show that two of the best-known dioxides, Pb02
and Mn02, are differently constituted because lead dioxide when decomposed by
sulphurous acid, H2SO3, furnishes lead sulphate, PbS04, while manganese dioxide
furnishes manganous dithionate, MnS206. It is more probable that the action in
both cases is similar, manganese dioxide forming the normal sulphite, Mn(S03)2 ;
and lead dioxide, the basic sulphate, PbO.SOs. Both salts then undergo internal
rearrangement, the former producing a dithionate, and the latter a normal sulphate.
There is a distinction between the peracids analogous to that occurring between
the peroxides. The true peracids are either formed by the action of hydrogen
peroxide on ordinary acids or their derivatives, or else they furnish hydrogen
peroxide when hydrolyzed with dilute sulphuric acid — with concentrated sulphuric
acid they behave like the peroxides and give ozonized oxygen. The prefix fer is
applied to many acids — ^perchloric acid, permanganic acid, etc. — to denote that
they contain relatively more oxygen than the acid indicated when the prefix is
deleted. These acids do not give hydrogen peroxide by hydrolysis, moreover
they are not formed by the action of hydrogen peroxide. In many cases, it is not
clear whether the so-called peracid is an additive compound of the acid with
hydrogen peroxide of crystallization, analogous to water of crystallization, or
whether it is a real peracid in which the acid radicle has united with the hydrogen
peroxide.
The elements which form peracids belong to the 3rd, 4:th, 5th, and 6th groups of
Mendeleeff's periodic system, although 0. Carrasco (1911) claims to have prepared
a perzincic acid. If the claim be established, zinc in the second group will have to
be included in the list of elements forming true peracids. T. S. Price ^ represents
the elements forming peracids in black type as illustrated in Table IV.
960 INORGANIC AND THEORETICAL CHEMISTRY
Table IV. — Elements in the Periodic System pobming Peracids.
Group ni.
Oronp IV.
Group V.
Group VI.
B
C
N
O
Al
Si
P
S
Sc
Ti
V
Cr
Ga
Ge
As
Se
Y
Zr
Cb
Mo
In
Sn
Sb
Te
(La)
Ce
—
—
Yb
. ,
Ta
W ~
Tl
Pb
Bi
—
• —
Th
■ — '
u
T. S. Price also shows that, omitting the first two rows, the elements which form
peracids are mainly confined to the members of the even series, and that the stability
of the peracids increases with increasing atomic weight of the element in agreement
with the rule that with the elements of the even series the higher the atomic weight
the greater their basicity — for instance, in the 6th group, peruranic acid is the
most stable peracid of the family, and can be prepared at ordinary temperatures ;
while permolybdic acid can be obtained only at temperatures approaching —10°.
Pertungstic acid is more stable than permolybdic acid.
The peracids can be regarded as derivatives of hydrogen peroxide in which one
of the hydrogen atoms is replaced by an acid radicle. Otherwise expressed, the
peracids can be regarded as acids in which one or more hydroxyl groups are
replaced by the monad radicle 0.0. H.
Ordmary acids.
Metaboric acid . . . HO.BO
Acetic acid . . . CH3CO.OH
Peracids.
Perboric acid .
Peracetic acid.
HOO.BO
CH3.CO.OOH
The diabasic acids form two series of peracids according as one or both the hydroxyl
groups are replaced by the HOO- radicle.
Carbonic acid.
^^^OOH
Monopercarbonic acid.
OOH
'"^^OOH
Dipercarbonic acid.
Similarly with the acids of higher basicity. The peracids form persalts. There
are theoretically two acid persalts and one normal persalt of the monoperacids,
and one each of the diperacids. For example,
^ .OONa
^Q .OOH
co<«r
p^ .OONa
^^^OONa
.OONa
^^^OOH
Salts of the percarbonic acids.
There is also a series of acid derivatives of hydrogen peroxide in which the acid
characteristics belong rather to the acid radicle itself than to the presence of the
HOO-group. They are generally prepared by electrolysis, and they can be
regarded as derivatives of hydrogen peroxide in which both the hydrogen atoms
are replaced by acid radicles. They are rather acid peroxides than true peracids,
although they are commonly called peracids. For example
HO
HO
Hydrogen i)eroxide.
HO.SO2.O
HO.SO2.6
Persulphuric acid.
HO.CO.O
H0.C0.6
Percarbonic acid.
The salts of these acid peroxides do not give the characteristic reactions of hydrogen
peroxide. For instance, the persulphates do not give the blue coloration by the
chromic acid and ether test ; potassium permanganate in acid solution is not
OZONE AND HYDROGEN PEROXIDE 961
decolorized ; iodine is not separated from potassium iodide ; etc. These salts,
however, may be hydrolyzed into hydroperoxides in aqueous solutions and they
may then show the characteristic reactions of hydrogen peroxide. In this series
. of persalts, the free acids corresponding with the persulphates have alone been
isolated. It will be observed that further complications are theoretically possible,
for these peroxide acids may form true peracids by the replacement of one or both
the hydroxyl radicles by HOO-radicles. Thus,
HO.CO.O HOO.CO.O HOO.CO.O
H0.C0.6 HO.CO.O HOO.CO.O
Peracids of the add peroxides.
References.
1 J. Sperber, Schweiz. Work. Chem. Pharm., 51. 409, 1913 ; Schiveiz. Apoth. Ztg.. 52. 2, 2459,
1913 ; 53. 717, 1915 ; W. Spring, ZeiL anorg. Chem., 8. 424, 1895.
2 A. von Baever and V. Villiger, Ber., 33. 2479, 1900.
3 R. Wolffenstein, German Paf., D.R.P. 196369, 1906 ; J. Tafel, Ber., 27. 816, 2297, 1894.
* E. Schcine, Liehig's Ann., 192. 257.
5 D. I. Mendeleeff, Journ. Russian Phys. Chem. Soc, 13. 561, 1881 ; M. Traube, Ber., 19.
Ill], 1115, 1117, 1886 ; F. Richar?, ib., 21. 1675, 1888.
« S. Tanatar, Ber., 33. 205, 1900 ; 36. 1893, 1903 ; 42. 1516, 1909.
7 C. Tiibandt and W. Riodel, Ber., 44. 2565, 1911 ; Zeit. anorg. Chem., 72. 219, 1911.
8 0. Pellini and D. Meneghini, Zeit. anorg. Chem., 60. 178, 1908.
^ T. S. Price, Per-acids and their Salts, London, 1912; 0. F. von Girsewald, Anorganische
Peroxide und Persalze, Braunschweig, 1914.
VOL. I. 3 Q
CHAPTEE XV
ELECTROLYSIS AND THE IONIC HYPOTHESIS
§ 1. The Products of Electrolysis
The electricity which decomposes, and that which is evolved by the decomposition of
a certain quantity of matter are (qualitatively and quantitatively) the same.^ — M. Faraday.
One or both of the products of electrolysis may be an insoluble solid, a soluble
liquid, a gas, etc. When an insoluble solid is formed it may stick to the electrode,
or fall to the bottom of the electrolytic cell ; if a gas, not too soluble in the electrolyte,
be formed, it can be collected in a suitable receiver. Substances are not always
visible when in solution. The soluble matter can often be isolated more or less
completely by surrounding the proper electrode with a porous pot whicli retards the
diffusion -and mixing of the products separated at the two electrodes. This is done,
for example, in the industrial preparation of chlorine.
The electrolysis of a solution of copper sulphate furnishes the products : copper,
sulphuric acid, and oxygen. This is more than was present in the copper sulphate
used at the start. It is therefore assumed, as a trial hypothesis, that Cu and SO4
ions are produced at the electrodes during the passage of the current ; that the
Cu-cation carries a positive charge of electricity, and the S04-anion a negative
charge. Consequently, the Cu-ion will be found at the
negative electrode, and the S04-ion at the positive
electrode. The ions are de-electrified at the electrodes —
the Cu-ion at the cathode, and the S04-ion at the anode.
The de-electrified copper ions are deposited as metallic
copper about the cathode ; and the de-electrified S04-ion,
at the anode, reacts at once with the solvent (water),
Fi f 1 • f P^^^^^^iiig sulphuric acid and oxygen : 2SO4+2H2O
Silver Nitrate!^*^ ^ =2H2S04+02. When an aqueous solution of potas-
sium nitrate is electrolyzed, potassium hydroxide and
gaseous hydrogen are formed at the cathode ; . and nitric acid and oxygen at the
anode. It is assumed that the potassium nitrate is first decomposed into two
electrified K+ and NO3- ions at the electrodes ; and that the K+-ion, when de-
electrified, reacts with water at the cathode, producing potassium hydroxide and
hydrogen ; and the NOs" ion, when de-electrified at the anode, reacts with
water, giving nitric acid and oxygen: 4N03~-}-2H20=4HN03-|-02. Again, if
a solution of copper sulphate be electrolyzed with copper electrodes, metallic copper
is deposited at the cathode, and the sulphuric acid produced at the anode
attacks and dissolves the copper cathode forming copper sulphate. This explains
how the total concentration of a solution of copper sulphate does not alter if it
be electrolyzed in a cell with a copper anode. Similar remarks apply to the
electrolysis of solutions of silver nitrate with a silver anode, Fig. 1 ; of ferrous
ammonium sulphate with an iron anode ; of nickel ammonium sulphate with
a nickel anode ; etc.
ElectropIating.^ — If a plato of silver bo used as the anode during the electrolysis of silver
nitrate, motallic silver will be dissolved by the nitric acid as fast as the acid is formed.
962
ELECTROLYSIS AND THE IONIC HYPOTHESIS 963
Thus, the concentration of the silver nitrate in the solution will remain unchanged and
metallic silver will be transported from anode to cathode. This is the principle of the
method of electroplating. In the case of silver-plating a firmer and more uniform deposit
of silver is obtained by using a solution of silver cyanide in potassium cyanide as the electro-
lyte in place of a solution of silver nitrate. The article to be plated, say a brass spoon, is
attached to a wire and dipped in the solution of silver salt, and this is made the cathode.
A bar or sheet of silver is made the anode. A rather weak electric current is sent through
the electrolyte. The electrolyte is decomposed, and silver (cation) is deposited on the
article to be plated (cathode) ; the anion collecting at the anode dissolves the silver anode,
and thus keeps the strength of the electrolyte unchanged. What is dissolved at the anode
is deposited at the cathode. Salts of other metals— nickel, iron, gold, platinum, etc. — can
be used as electrolytes in a similar maimer, and accordingly articles can be nickel-plated,
gold-plated, etc. The plated articles may be afterwards burnished.
Units. — We first inquire if there is any relation between the quantity of electricity
passing through an electrolytic cell and the amount of decomposition. In order to
fix a standard of measurement, let the quantity of electricity required to deposit
0-001118 gram of silver be called a coulomb. This is the so-called unit quantity of
electricity. Hence 108 grams of silver, that is, a chemical equivalent of silver, will
be deposited by 96,540 coulombs of electricity. This amount of electricity is often
called a farad.
The so-called " hydraulic analogy " of an electric current might here be cited. The
quantity of water flowing through a pipe can be expressed in gallons or cubic feet per
second ; in a similar way, quantity of electricity may be expressed in terms of coulombs
per second. An electric current carrying one coulomb per second is called an ampere.
This is the so-called unit current of electricity. A coulomb by the same analogy would
correspond with, say, a gallon or cubic foot of water ; and an ampere with a gallon or a
cubic foot of water per second. For example, if 20 coulombs of electricity pass through a
current in 20 seconds, the current is 60-^20=3 amperes, or 3 coulombs per second. The
total quantity of water delivered by a pipe is determined by the " head " or pressure of
water, so that in order to pass a certain number of gallons per second through a given pipe,
a certain pressure must be applied to overcome the frictional resistance of the pipe. In
the same way, a certain electromotive force — electrical pressure — is required on account
of the resistance offered by the wire to the flow of electricity. Just as water pressure is
measured in pounds per square inch, or in feet " difference of level " or " head," so the
unit of electrical pressure, the volt, is the difference of potential needed to produce a current
of one ampere in a conductor whose resistance is equivalent to that of a uniform column
of 14'45 grams of mercury, 106-3 cm. long. The resistance of such a column is called
an ohm. Hence a volt is the electric pressure required to produce a current of one
ampere in a conductor of one ohm resistance. The terms voltage, electrical pressure,
and electromotive force are generally applied synonymously to an electric current, or, if the
current be not directly under consideration, the term difference of potential is used. It is,
of course, needless to dwell on the fact that the analogy used above in comparing an electric
current with a moving fluid is merely a convenience. It is probable that electricity is not a
fluid, and the analogy must not be carried much further.
§ 2. Faraday's Laws of Definite Electrolytic Action
Nature presents us with a single quantity of electricity. For each chemical bond which
is ruptured within an electrolyte, a certain quantity of electricity traverses the electrolyte,
which is there in all cases. — G. J. Stoney (1881).
M. Faraday (1834) i found that the amount of chemical work done by an electric
current is directly proportional to the quantity of electricity which passes through
the electrolyte. If one farad leads to the separation of 108 grams of silver, two farads
will lead to the separation of 216 grams of silver, and so on. Similar results are
obtained with other electrolytes. Hence, said Faraday, " the chemical decomposing
action of a current is constant for a given quantity of electricity ; " or, *' the quantity
of chemical decomposition is exactly proportionate to the quantity of electricity
which has passed throu,s;h the electrolyte ; " if w denotes the number of grams of
an element set free by the decomposition of a compound by the passage of a quantity
964
INORGANIC AND THEORETICAL CHEMISTRY
of electricity C, then iv is proportional to C ; and consequently, *' the products of
decomposition . . . afford a very excellent and valuable measure of the electricity
concerned in their evolution." The increase in the weight of, say, the negative
electrode during the electrolysis of silver nitrate or copper sulphate owing
to the deposition of metallic silver or copper respectively, is a measure of the
quantity of electricity which has passed through the system. A cell specially
designed for such measurements is called a silver voltameter or a copper voltameter
respectively.
Provided there are no disturbing secondary " actions, the amount of electro-
decomposition is not affected by the strength (or intensity) of the current, the time
the current is passing, the concentration of the solution, the nature of the dissolved
substance, nor by the temperatiire. The same quantity of electricity will always
liberate the same quantity of the elements stated. The accuracy of the law is said
to have been established for " currents so small that a century would be required
for the separation of a milligram of hydrogen," and in large electrochemical works,
the law is continually being verified by the passage of millions of coulombs. In
every case, the law describes the phenomena exactly. The quantity of an element
liberated by the passage of one farad of electricity is called the electrochemical
equivalent of the element.
Again, let a current be simultaneously passed through six cells containing
respectively dilute sulphuric acid,
aqueous solutions of silver nitrate,
cuprous chloride and hydrochloric
acid, cupric sulphate, gold
chloride, and stannic chloride.
The experiment is conducted by
arranging the electrolytic cells as
illustrated in the plan, Fig. 2.
After about half an hour's elec-
trolysis the amounts of the different elements collected at the cathode can be
weighed or measured. The results will be very nearly :
Fig. 2. — Experimant illustrating Faraday's Laws.
Dilute H2SO4
AgNOa CuCl
CUSO4
AUCI3 SnCl4
Cathode. Anode.
Hydrogen. Oxygen.
Silver. Copper.
Copper,
Gold. Tin.
Amount found
. 00266 0-2126
2-9370 1-6000
0-8440
1-7476 0-7554 gram
IfH = l .
. 1 8
108 03-5
31-8
65-7 20-8
Atomic weight
. 1-01 16
107-9 63-6
63-6
107-2 110
Valency
. 1 2
1 1
2
3 4
Accordingly, chemically equivalent quantities of the different elements (that is,
atomic weight-;- valency) are liberated by the passage of the same quantity of
electricity. Consequently, the electrochemical equivalent of an element is numeri-
cally the same as the chemical equivalent. The equivalent weights of bodies are
those quantities of them which are decomposed by equal quantities of electricity.
Hence, it is inferred that electricity determines the combining forces because it
determines the combining weights.
At first sight, this result appears to contradict the principle of excluded perpetual
motion, because, if the current from a Zn|H2S04|Pt battery be sent through an
indefinite number of electrolytic cells containing dilute sulphuric acid, the same
amount of hydrogen would be liberated in each, and sufHcient hydrogen could be
collected to furnish, on combustion, enough heat to evaporate the solution of zinc
sulphate in the battery to dryness, to transform the zinc sulphate to metallic zinc
and suli)huric acid, and so reconstruct the battery ; and have some hydrogen
remaining in excess. The experiment fails. The current will not traverse an
indefinitely large number of cells. W. H. Wollaston 2 showed in 1801 that in dealing
with electrical energy we are concerned with two different factors, and that
" quantity of electricity " is only one of these factors. Faraday's law describes the
ELECTROLYSIS AND THE IONIC HYPOTHESIS 965
influence of *' quantity of electricity " upon electrolysis. It says nothing about
the electrical pressure —the electromotive force, described, say, in volts — required to
drive a given quantity of electricity through the system. Hence, Faraday's work
may be summarized : The same quantity of electricity passing through one or more
electrolytes connected up in series, will liberate in each cell chemically equivalent
amounts of the products of electrolysis, provided the electromotive force permits
the necessary current to be maintained. It might here be added that, for reasons
which will be discussed later, a certain specific electrical pressure or voltage —
called the decomposition voltage — is required to electrolyze a given solution ;
thus, hydrochloric acid requires about IJ volts, and fused sodium chloride about
4 volts.
Nomenclature.- — ^Let each positive charge of electricity be represented by a small dot,
and each negative charge by a small dash at the upper right-hand corner of the chemical
symbol for an element, then, a silver ion will be written Ag*; a zinc ion by Zn" ; a nitrate
ion, NOg' ; and a sulphate ion by SO4''. In the electrolysis of aqueous solutions of salts,
etc., the separation of an ion at one electrode is always attended by the separation of a
chemically equivalent ion or ions at the other electrode. For instance, with zinc chloride,
for every Zn" which is de-electrified at the cathode, two CI' ions will be de-electrified at
the anode. In order to designate positive ions J. Walker (1901) ' appends ion to the stem
with a prefix mono-, di-, tri-, ... to indicate the number of charges carried by the ion.
Thus H* is called hydrion ; Na*, sodion ; Fe**, diferrion ; Fe**% triferrion ; etc. For
negative ions, the termination of -ate becomes -anion ; -ite, becomes -osion ; and -ido,
becomes -idion. Thus OH' is hydroxidion ; CI', chloridion ; CIO3', chloranion ; OCl',
hypochlorosion ; SO 3", sulphosion ; SO/', sulphanion ; etc. To this, A. Smith (1901) adds
the term ionoyen for " bodies which are capable of undergoing ionization," reserving the
term electrolyte for the solution as a whole than for the substance dissolved.
We have just seen that the electrochemical and chemical equivalents are
numerically the same, and therefore the electrochemical equivalent of an element
is obtained by dividing the atomic weight by the valency. The same quantity of
electricity — positive or negative — must therefore be carried by each univalent
atom, and accompany it in all its movements in the electrolytic fluid. This quantity
has been called the unit charge of the ion. At first sight this deduction appears to
be rather startling, for it seems to imply either that the electric charges are divisible
or that the so-called bivalent atoms are composed of two sub-atoms, the tervalent
atoms, of three sub-atoms ; etc. The former hypothesis is generally accepted.
Accordingly, a univalent atom is supposed to carry one charge of electricity (96,540
coulombs) ; a bivalent atom two charges, and an w-valent atom, n charges. Accord-
ing to this view, valency represents the number of charges of electricity which are
associated with the respective ions, and chemically equivalent quantities of matter
have the same capacity for electricity. That is, '' the chemical equivalent is the
electrical unit of matter," or, as M. Faraday expressed it :
The equivalent weights of bodies are simply those quantities which contain equal
quantities of electricity, or have naturally equal electric powers ; it being electricity
which determines the equivalent number because it determines the combining force. Or,
if we adopt the atomic theory or phraseology, then the atoms of bodies which are equivalent
to each other in their ordinary chemical action have equal quantities of electricity naturally
associated with them.
So close is the relation between the chemical and electrochemical equivalents
that R. Luther (1905) proposed to define the combining weight of a univalent
element as the quantity corresponding with 10,(X)0 electromagnetic units, which,
in turn, is very nearly 100,000 coulombs. Hence this means little more than
multiplying the numbers at present in use by 3"46 per cent.
Quantity of electrolytic work done by a current. — ^M. Faraday has shown that
the amount of chemical decomposition in a given time depends upon the amount of
current employed ; and that the one magnitude can be computed when the other
is known. To find the relation between the chemical equivalent of an element and
quantity of electricity. Let e denote the chemical equivalent of a substance, then the
966 INORGANIC AND THEORETICAL CHEMISTRY
weight w of an element liberated by constant quantity of electricity is proportional
to e, but, w is also proportional to the quantity of electricity C ; and therefore
w=k€C, where A; is a constant sometimes called Faraday's constant. If suitable
units be chosen, io—€C. Consequently, the electrochemical equivalent of an
element is the amount in grams liberated by one coulomb. Careful measurements
have shown that 0001 118 grm. of silver will be deposited by one coulomb. Since
the chemical equivalent of silver is very nearly 107*88, hydrogen unity, 0*001118 grm.
is 107*88 times greater than the amount of hydrogen which will be separated by a
coulomb; accordingly, 000001036 grm. or 1*036x10-5 grm. of hydrogen will he
separated by the -passage of one coulomb of electricity . The electrochemical equivalent
of univalent copper is '63*5 X 1*036 X 10-^; of bivalent copper, J of 63*5 X 1*036 X 10-5 ;
of ferrous iron, J of 56x1*036x10-5; of ferric iron, | of 56x1*036x10-5; and
generally, if e denotes the chemical equivalent of an element — that is, atomic weight
-^ valency — the electrochemical equivalent is 1*036 X 10" 5e. Again, if a coulomb of
electricity liberates 1*036x10-5 grms. of hydrogen per second, a amperes will
liberate 1*036 X 10- 5a€ grms. per second, and generally, the number of gratns of an
eletnent ivltose chemical equivalent is liberated by the passage of a amperes of electricity
flowing for t seconds is \'OZQxlO~^eai. If € denotes the electrochemical equivalent
of an element, such that €=1*036 Xl0-5e, the number of grams of an element
whose electrochemical equivalent is € liberated by the passage of a amperes flowing
for t seconds is eat. It also follows that 1*036 X 10" 5^^^^ grms. of an w-valent
element of atomic weight A, will be deposited per coulomb. By Avogadro's rule,
a gram-molecule of a gas — of molecular weight M and iV-atoms per molecule —
occupies 22*4x103 c.c. Hence (22*4 X 1*036 XlO-2)/iVw c.c. of the gas are given off
per coulomb. For a univalent gas with two atoms per molecule, iV"=l, n=2, and
therefore 0*116 c.c. are obtained per coulomb at n.p.t., or 8*6193 coulombs are
required per c.c. of gas. This corresponds with 0*4176 litre per ampere hour, or
2*3943 amperes are needed per litre of gas ; or with 14*750 c. ft. are liberated per
100 amp. hours, or 67*798 amps, are needed per cubic foot of gas.
Examples. — (1) An electric current is passed simultaneously through the following
solutions : Hydrochloric acid, ferrous sulphate, ferric sulphate, and silver potassium
cyanide. If 5*2 litres of hydrogen at n.p.t. were evolved from the hydrochloric acid, how
much metal would be deposited in the case of the iron and silver salts ? Here 5 "2 litres of
hydrogen weigh 0'4664 grm. The chemical equivalent- — -that is, one gram — of hydrogen is
equivalent to ^ of 56 = 28 grms. of ferrous iron; to 18*67 grms. of ferric iron; and to
108 grms. of sUver. Hence, 0*4664 x 28 = 13 grms. of iron will be deposited from the ferrous
sulphate ; 8*7 grms. from the ferric sulphate ; and 50'1 grms. from the silver solution.
(2) A current of 0*04 amp. was passed through a solution of copper sulphate for 1^ hrs.
Hence, how much copper was deposited when the electrochemical equivalent of copper is
0000329 grm. t There are 5400 seconds in 1^ hrs., hence 0000329 X 0*04 x 5400 grms. of
cojjper were deposited.
(3) A current of 2^ amps, was obtained from a voltaic cell for 2\ hrs. How much
zinc was dissolved, given the electrochemical equivalent of zinc is 0*000337 '/ Ansr. 6*8
grams.
(4) What are the electrochemical equivalents c of hydrogen, of copper (cupric), and of
zinc ? Hydrogen, 1*036 X 10"^ x 1 =0*00001036 ; copper, ^ of 63*5 X 1*036 X 10-5=0*000329;
zinc, ^ of 65 X 1*036 x 10-5=0*000337.
(5) In C. Hopfncr's method (1900),^ a solution of ferric and sodium chlorides is used aa
electrolyte and the copper passes into solution as cuprous chloride. Compare the quantity
of electrical energy required to precipitate copper from Hopfner's solution, and from a
solution containing copper sulphate. Here bivalent copper requires 2 x 96,540 coulombs
for precipitating 63*57 grms., while univalent copper requires but 96,540 coulombs for
precipitating the same weight of copper. Hence the precipitation of a given weight of copper
from cuprous chloride requires but half the electrical energy as that required when cupric
sulphate is used.
H. von Helmholtz (1881) 5 has emphasized the fact that the evidence indicates
that electricity associates with the atoms of matter in multiples of one fundamental
quantity, but never in fractions of it ; these fractions may not be impossible, but they
have not yet been found. Hence the evidence for the atomic nature of electricity
ELECTROLYSIS AND THE IONIC HYPOTHESIS
967
is much the same as for the atomic nature of matter. The charge on a monad
atom is therefore a natural unit of electricity. To illustrate the prodigious electrical
capacity of the molecules, H. von Helmholtz estimates that if the opposite electrici-
ties were extracted from a milligram of water, and given to two spheres a mile apart,
these two spheres would attract each other with a force of ten tons.
References.
1 M. Faraday, Phil Trans., 123. 23, 1833 ; 124. 77, 1834 ; Experimental Researches in Elec-
tricity, London, 1. 107, 215, 230, 821, 1849.
2 W. H. Wollaston, Phil. Trans., 90. 427, 1801 ; R. Luther, Zeit. Elektrochem.y 11. 273, 1905.
3 J. Walker, Chem. News, 84. 162, 1901 ; A. Smith, ib., 84. 279, 1901.
* 0. Hopfner, German Pat., D.R.P. 704640, 1900.
6 H. von Helmholtz, Journ. Chem. Soc, 39. 277, 1881.
§ 3. The Velocity o! Electrolytic Conduction
The conduction of electricity through electrolytes is utterly indistinguishable from
metalHc conduction except for the action at the electrodes which is not part of true con-
duction at all. — J. T. Sprague (1892)
An electric current travels through an electrolytic solution as quickly as if the
same current were sent through a copper wire of the same resistance, and the
products of electrolysis appear simultaneously at
both electrodes, however far apart the electrodes
be placed. N. M. Hopkins (1905) i passed a
current through a tube 1500 cm. long, and
through another tube 10 cm. long, and measured
the time required for the current to pass by
means of a chronograph sensitive to nearly
10,000 cm. per second. The tubes were filled
with dilute sulphuric acid and fitted with elec-
trodes— the anode of copper and the cathode of
platinum. As soon as the current passed , bubbles
of hydrogen appeared at the cathode simul-
taneously with the blue colour of copper sulphate
at the anode. The electrolyte 1500 cm. long con-
ducted as quickly as the electrolyte 10 cm. long.
The experiment can be illustrated by the apparatus sketched in Fig. 3, which almost
explains itself. The long spiral tube contains the electrolyte as in Hopkins'
experiment. As soon as the circuit is closed electrolysis begins. As A. E. Dolbear
puts it :
If the two terminals of an electric circuit were on opposite sides of the Atlantic ocean,
and a current were sent through the circuit, hydrogen would appear on one side and oxygen
on the other . . . and in amounts defined by Faraday's laws of electrolysis.
Returning to Fig. 3, the known rates of diffusion of molecules in solutions are
altogether too slow to allow the SO4 which attacked the copper, to have come from
the same H2SO4 molecule as the hydrogen liberated at the cathode. Further,
it is supposed that the electrical energy used in electrolysis is entirely expended in
overcoming the resistance of the electrolyte, and no measurable quantity of work is
needed for tearing apart the components of the decomposing molecule. Hence, it
follows that (i) the molecules of an electrolyte in solution must be in a condition to
conduct the electric current immediately the necessary electrical stress is applied
to overcome the resistance of the liquid ; and that (ii) there must be either an
Fig.
3. — Velocity of Electrolytic
Conduction.
968 INORGANIC AND THEORETICAL CHEMISTRY
exchauge of partners among all the molecules of the liquid which take part in
conducting the current, or else anions and cations exist in the liquid in the free
state.
Refebences.
* N. M. Hopkins, ExperimeiUul Electrochemistry, London, 74, 1905 ; A. E. Dolbear, Matter
Ether, and Motion, London, 1891).
§ 4. The Effect of the Solvent
Tlioso bodies only are electrolytes which are composed of a conductor and a non-
conductor.— W. A. Miller.
Does the salt alone, or the water alone conduct the current ; or is the conduction
of the current shared between the solvent and solute ; or does neither the salt nor
the water alone conduct the current, but is the current carried by a hydrate which
conducts and is decomposed by the current as a whole ? The more care taken in
the purification of water, the less does it conduct electricity, and consequently, it
is assumed that 23ure water is a non-conductor in spite of the fact that perfectly
non-conducting water has not yet been made. Pure dry liquid hydrogen chloride,
like water, appears to be a non-conductor. A mixture of water and hydrogen
chloride is an electrolyte. Hence, it is inferred that the electrolytic conductivity
of a solution is a joint property of solvent and solute, and not a property of either
constituent alone. Solutions of dry hydrogen chloride in some solvents — e.g. dry
benzene or chloroform — conduct electricity so feebly, if at all, that they are said to
be non-conducting ; and solutions of some substances in water conduct no better
than water itself — e.g. solutions of sugar or alcohol in water. Hence, also, it follows :
the electrol3rtic conductivity of a solution depends upon some specific relation
between the solvent and the solute. The same conclusion can be deduced in the
following manner : If the solute— say copper sulphate — alone conducts the current
in a cell with platinum electrodes, then copper alone will be deposited at the cathode,
and the SO4 at the anode decomposes the water forming sulphuric acid and liberating
oxygen ; i.e. all the free acid appears at the anode. If the water alone conducts the
current oxygen alone appears at the anode, and at the cathode, hydrogen decom-
poses the copper sulphate forming sulphuric acid and depositing copper ; i.e. all
tlie free acid appears at the cathode. If water conducts l/:cth and the solute conducts
(1 — a;)/a:th of the current, then the ratio of the free acid formed at the anode, to the
free acid formed at the cathode, will be as {x—V) : 1. So far as observation shows,
some free acid is always formed at both electrodes, and hence the conduction of the
current is probably shared by the solvent and solute. The facts observed also fit
the assumption that a hydrate or hydroxylic compound exists in solution, and this
conducts and is decomposed as a whole.
In a general way, aqueous solutions of acids, bases, and salts conduct electricity,
and these substances are often called electrolytes, not because the salt conducts the
current, but because their aqueous solutions conduct the current electrolytically.
Some fused salts conduct electrolytically, e.g. with fused silver chloride and silver
electrodes, silver is dissolved at the anode and deposited on the cathode, so that the
total amount of silver chloride is maintained constant ; with carbon electrodes,
silver is deposited at the cathode, and chlorine evolved at the anode.
It is usually stated that an acid or alkali is added to water in order that the latter
may be decomposed into its constituent elements by the electric current. The function
of the acid (or, mutatis mutandis, of the alkali) has been a subject of some speculation.
(i) It has been said that the mere presence of the acid simply makes the water a con-
ductor and that the water alone is decom])osed by the current : 2H5,0=2A2+ 4-02~ ;
it has also been said that the acid alone is decomposed by the current, and that
the water is attacked bv the products of the electrolysis and chemically decom-
posed. Symbolically, 2H2S04=2H2+2S04 (electrolysis) followed by 2SO4+2H2O
I
ELECTROLYSIS AND THE IONIC HYPOTHESIS 969
=2H2S04-|-02 (chemical). As a matter of fact, it is very doubtful if pure anhydrous
acid or pure water is a conductor. Hence, it is not more logical to say that the
acid makes the water a conductor than that the water makes the acid a conductor.
Each constituent loses its individuality when mixed together. At first sight, it
seems as if during the electrolysis of acidulated water, the mixture must le
regarded as a unit which (1) conducts the current from one electrode to the
other ; and which (2) suffers decomposition by electrical influences at the
surfaces of the electrodes. Several working hypotheses can now be devised — e.g.
with dilute sulphuric acid, it is plausible to assume that the electrolyte contains
the complex H2S04.??H20. During electrolysis, neither the water nor the acid is
decomposed, but rather the complex: 2H2S04.^H20=27iH2+H-?i02~+2H2S04.
The action of the current is to deprive the complex of both hydrogen and oxygen in
the proportions 2H2 : O2.
In a general way it may be said that (i) electrolytic conduction is accompanied
by visible decomposition, or (ii) polarization phenomena {q.v.) may appear, (iii) An
assembly of metals at a constant temperature can give no current, but if an electro-
lyte be introduced into the series a current can be obtained, (iv) According to the
electromagnetic theory of light, if a conductor be transparent it will probably
conduct electrolytically, e.y. fused salts, hot glass, etc.
§ 5. The Ionic Hypothesis
Let us learn to dream, then perhaps we shall find the truth.- — A. Kekule.
In framing hypotheses we must see that they agree with facts ; in other respects, they
may be as inconceivable (not self- contradictory) as any fairy tale.- — M. M. P. MuiR.
The main facts so far established by the preceding discussion of the phenomena
attending electrolysis may now be summarized :
(1) Electrolytes in solution conduct electricity, and the process of electrical
conduction is attended by a splitting of the molecules of the solute into
anions and cations ; the anions appear at the anode, and the cations at the
cathode. The separation of a certain number of anions at the anode is
simultaneously attended by the separation of a chemically or electrically
equivalent number of cations at the cathode. During electrolysis, the anions
and cations appear to be discharged electrically, because electrically neutral
molecules appear as secondary products of the electrolysis.
(2) The anion which separates at the anode is not necessarily derived from
the same molecule as the cation which appears at the cathode.
(3) Solvent and solute together make a conducting medium, since as a rule
neither solvent nor solute alone shows any marked capacity for conducting
electricity.
(4) No measurable time is needed to put an aqueous solution in a condition
to conduct the current. Immediately the necessary difference of potential
appears at the electrodes the process of electrolysis begins.
(5) Osmotic pressure and related phenomena show that electrolytes in
dilute solution have what seems to be a molecular weight, which suggests
that the ordinary chemical molecule of the electrolyte dissolved in certain
solvents is dissociated into two parts.
It is generally agreed that during electrolytic conduction there is a convection
of electricity by the atoms of matter, but there have been differences of opinion as
to the mode of transit of the atoms through the liquid :
(1) The molecular chain hypothesis of C. J. T. von Grotthus was generally
accepted in the first half of the nineteenth century. In his Theorie de la
decomposition des liquides par Velectricile gahanique, he (1805) ^ assumed that
the molecules of salt in solution are distributed throughout the solvent in an
irregular manner without any signs of orientation, as represented diagrammatically
at A, Fig. 4. The molecules, in the presence of a pair of oppositely charged electrodes
970 INORGANIC AND THEORETICAL CHEMISTRY
range themselves in " chains," like little magnets. The positively charged ions —
cations — are directed towards the negatively charged cathode, and the negatively
charged ions — anions — to the positively charged anode, B, Fig. 4. If the charges
on the electrodes are great enough, the molecules in immediate contact with the
electrodes will decompose, C, Fig. 4, and the charge on one of the ions will be
neutralized by the charge on the electrodes. And the other ion will unite with the
neighbouring molecule and liberate an ion with a similar charge. The free ion
attacks the next molecule, and so the process is continued throughout the " chain."
To fix the idea, consider the end of the molecular chain at the cathode. There,
a negatively charged ion is se.t free when a positively charged ion is neutralized at
the cathode. This " negative " ion associates with the adjacent molecule of the
chain ; this molecule decomposes, forming a new molecule, liberating, at the same
time, a negatively charged ion which associates with the next molecule of the
chain, D, Fig. 4. This successive decomposition and recombination goes on
throughout the chain of molecules from electrode to electrode. The new molecules
80 formed turn about, and are again ranged in a " chain " resembling B, as shown
at E, Fig. 4. A cycle, of changes of this nature is supposed to be going on all the
time the current is passing through the electrolyte.
C. J. T. von Grotthus' mechanical interpretation of the phenomenon is very
_ . ingenious, and it satisfactorily
1^ I ex2)lained the facts known in his
^^c* 5^ 8 i^?.A day, but later knowledge has shown
« ! that the hypothesis is not tenable
11 in its original form. If the elec-
•o«o«o«o«o*o«o*c B tricity be conducted by Grotthus'
_ ' chain, no current can flow until the
^l electromotive force driving the
o«o«o«o«o«o#o» oC energy is equivalent to the energy
— 4- represented by the heat of forma-
II tion of the molecules undergoing
o.o.o.o»c»o»o.o. D decomposition in the solution.
_ J_ When a sufficient electromotive
II force is applied at the electrodes,
•o«o«o*o«o«o«o«o E the decompositions and recom-
Fio. 4.-Diagrammatic Representation of C. J T. Positions of the molecules might
von Grotthus' Chain Hypothesis. proceed m the way described by
Grotthus. No such critical electro-
motive force, however, has been found to be necessary for the passage of a
current through electrolytic solutions. In H. von Helmholtz's air-free cell,2
polarization is produced by an infinitesimal element, but no permanent leakage
of electricity goes on through the cell until the applied current attains a
certain voltage. The smallest electromotive force hitherto tried causes a current
to flow when it is applied to copper electrodes immersed in a solution of copper
sulphate. The total energy consumption is then nothing but that due to the
" resistance " of the cell. Again, solutions of electrolytes, like metallic wires,
conduct electricity in such a way that the rate at which "electricity ])asses through
the system is proportional to the electromotive force. This is true whatever be the
magnitude of the force, and consequently, if a certain amount of electrical energy
be expended in breaking up the molecules, this proportionality cannot obtain.
Hence, very little electrical energy can be expended in breaking up the dissolved
molecules into their respective ions, and it has therefore been urged that " the ions
cannot be held together by a force of finite value." Consequently, in the homo-
geneous electrolyte without polarization no hypothesis which involves the tearing
of the molecules asunder against the chemical binding forces can be admitted ;
there is no chemical cling of the atoms, but only a frictional rub. Otherwise,
those electrolytes whose atoms or radicles are held together by the weakest
ELECTROLYSIS AND THE IONIC HYPOTHESIS 971
attractions would most readily decompose electrolytically. This is by no means the
case. For instance, mercuric chloride is much less stable than sodium chloride,
and yet the latter is much more readily decomposed by an electric current.
G. E. Fitzgerald 3 has pointed out that the difficulty with Grotthus' hypothesis
can be overcome if it be assumed that when the molecules are polarized, they draw
one another apart at a rate proportional to the polarization. This at once makes the
relation between electric force and decomposition a linear one, and so satisfies Ohm's
law in the case of small currents. It also so far agrees with Clausius's hypothesis
that it explains electrolysis and double decomposition as properties of the same
kind. The molecules in a liquid will occasionally be arranged by accident in a proper
polarized condition in a closed circuit for drawing one another apart ; and if th,e
circuit includes molecules of different kinds, there will result double decomposition.
He added :
The supposition that it is a particular arrangement that is required before exchanges
take place, and that with this arrangement exchanges take place of their own accord, seems
to explain electrolysis and double decomposition without supposing free atoms to exist
within the liquid.
(2) The electrostatic strain hypothesis of H. von Helmholtz.— Here it is assumed*
that each kind of matter has a specific attraction for electricity — some kinds for
positive, other kinds for negative ; that accordingly, work must be done to separate
one atom from its electrical charge, or to remove electricity from an atom of high
specific attraction and give it to another lower in the scale. Further, the chemical
affinity is mainly due to the electrical attraction of oppositely charged atoms, and
that when such atoms combine into a compound molecule, they do not discharge
into each other, but retain their charge. During electrolysis, work is done, not in
tearing the atoms asunder, but in tearing their electrical charges from them.
(3) The ionization hypothesis o! R. Clausius.— As a trial hypothesis it may be
assumed that the mere presence of the solvent leads to the fission of the molecules
of the electrolyte into sub-molecules, each of which is charged with a definite amount
of positive or negative electricity equivalent to 96,540 coulombs per chemical
equivalent. The solution does not itself appear to be electrically charged, and
hence it is assumed that equal quantities of positive and negative electricity are
developed by the rupture of the molecules of the electrolyte during the process of
solution. Solutions of electrolytes are supposed to normally contain a definite
I)roportion of the sub-molecules charged with electricity. By a modification of
M. Faraday's definitions the '' sub-molecules " are called ions, and consequently :
ions are atoms or groups of atoms which carry a positive or negative charge of
electricity, and they are formed by the dissociation of the electrolyte in the solu-
tion. Each molecule, on dissociation, furnishes two kinds of ions with equal
and opposite charges of electricity. Consonant with M. Faraday's work, it is
further assumed that each monad ion carries a definite charge of electricity (96,540
coulombs) ; each dyad ion, two such charges ; a triad ion, three such charges ;
etc. ; but never a fraction of such a charge. To avoid confusing the phenomenon
of dissociation, in which the products are not charged electrically, with the dissocia-
tion of a molecule into electrically charged ions, the term ionization is reserved
for the latter phenomenon. The ionization of hydrochloric acid is represented in
symbols: HCl^H'-f CI' ; and of sodium chloride : NaCl^Na--f CI'.
A. W. Williamson's theory of the continuous interchange of the atoms of the
molecules of a compound was suggested in 1850, and it was followed in 1857 by
R. Clausius' suggestion that the molecules of the solute are ionized when dissolved
in the solvent, but R. Clausius appears to have assumed that only an infinitesimally
small fraction of the total number of dissolved ynoleculcs are so ionized. As the ions
are discharged at the electrodes during electrolysis, more molecules are ionized.
The un-ionized molecules keep the electrolyte constantly supplied with a definite
number of ions. The ions conduct the current ; the " undissociated " molecules
972 INORGANIC AND THEORETICAL CHEMISTRY
are inactive. Further, at any given temperature, there is a constant relation
between the number of un-ionized molecules, and the number of ions. S. Arrhenius
(1884), more bold or less cautious than R. Clausius, asserted that a considerable
fraction of the dissolved molecules are ionized, and that the number of ions increases
more and more as the solution becomes 7nore and more dilute. W. Ostwald, J. H.
van't HofE, W. Nernst, and a large number of other workers have followed the logical
consequences of Arrhenius' hypothesis in a great many directions ; the results, on
the whole, have been satisfactory, and the theory has thus stimulated the study
of the properties of solutions in a remarkable manner. Some hold that the great
cloud of subsidiary hypotheses which is needed to make the ionic theory presentable,
serves also to obscure progress towards a more satisfactory view of the nature of
solution. It is also maintained that the " principle of exhausting hypotheses " has
not been followed, and that the favoured child — the ionic hypothesis— has grown
into a tyrannical master ; for instance, G. F. Fitzgerald (1896) has said that " the
supposed advantage of the free ion theory is not only illusory but misleading." If
this be a correct diagnosis of the ionic hypothesis we have some consolation in
H. Davy's words : " The destruction of an error hardly ever takes place without
the discovery of truth."
At first sight the ionic hypothesis appears so incredible and so opposed to the
instinct, common sense, or prejudices of the chemist that it has been assailed by
much wholesome criticism — particularly by H. E. Armstrong. For instance, it is
asked :
1. In view of the great chemical activity of metallic sodiwn in contact with water,
is it profitable to postulate the existence of the elernent sodium m contact with water
without chemical action ? This objection is said to " rest on a misunderstanding,"
because electrically charged ions of sodium in an aqueous solution of sodium chloride
are very different from neutral atoms of metallic sodium. The ions of sodium
carry large charges of electricity. It is urged that " chemists know practically
nothing about the properties of atoms carrying large charges of electrical energy,"
and also that " the chemical activity of an atom of sodium charged with its 96,540
coulombs of electricity is much less than a neutral atom of sodium." In other
words, the presence of the electrical charge on the sodium ion keeps the ordinary
chemical activities of the atom in abeyance. This means that whenever a chemical
diflBiculty arises in the application of the ionic hypothesis the assumption is made
that " neutral atoms or atomic groups and ions are diiTerent substances," because
the properties of a substance are determined as much by the energy it contains as
by the kind of matter. In this way, the ions have been invested with such imaginary
properties as may be needed to keep the ionic hypothesis consistent with facts.
2. Compounds like mercuric chloride, very prone to thermal dissociation, are not
readily ionized ; while compounds like calciufn chloride ivhich resist thermal dis-
sociation are readily ionized. Would not the ionic hypothesis predict the converse
plienomena ? Mercuric chloride is very volatile and readily dissociates into its
elements by heat ; calcium chloride, on the contrary, does not readily volatilize or
dissociate except at very high temperatures, yet it is said that the latter is readily
ionized in solution while the former remains all but unchanged. Here again it is
said that totally difierent phenomena are confused, and that the forces which
produce ionization are quite different from those which produce thermal dissociation.
3. Bodies carrying electrical charges of opposite sign arc attracted and cling to one
another ; iftJierefore a mobile solution contains ^^ free and independent " ions carrying
erwrmous electrical charges of opposite sign, how can the charged ions remain 7nore
than momentarily free ? It is assumed that a certain proportion of the molecules of
the solute are continually breaking down into free (charged) ions, and a certain
proportion of the ions are continually recombining to form ordinary molecules, the
result is, that the ratio between the number of free ions and paired ions (molecules)
remains unchanged. This statement, of course, does not answer the perplexing
question. Attempts have been made to refer the difficulty to the specific insulating
ELECTROLYSIS AND THE IONIC HYPOTHESIS 973
properties — the so-called dielectric constant — of the solvent. The action of the
solvent has been compared with the function of the glass in a charged Leyden jar.
This agrees with the non-conducting qualities of pure water, but experiments have
shown that the relation between the insulating properties of a solvent and its
ionizing properties is not an adequate and sufficient explanation of the observed
facts. The two phenomena do not always vary concomitantly. A satisfactory'
answer to the question, therefore, has not yet been found.
4. If an ionized salt, say, sodium chloride, is present in solution as a mixture of
Na' and CV ions, it might be thought possible to separate the two components by diffusion
or by some other mechanical process. When the molecules of certain gases — hydrogen,
chlorine, etc. — exist free in a liquid, they will escape ; but when, say, sodium chloride
is ionized : NaCl^Na'-j-Cr, it is said that the chlorine ions do not escape because
of their electrical charge. S. Arrhenius also says that the great electrostatic attrac-
tion of the oppositely charged ions prevents any marked diffusion. W. Nernst,
however, has shown that the concentration currents produced when, say, a solution
of sodium chloride is carefully covered with a layer of water, leads to the conclusion
that the greater mobility of the chlorine ions charges the upper layer negatively,
and the lower layer positively, so that a current of electricity can be obtained by
placing the two layers in electrical contact. R. C. Tolman (1911) whirled aqueous
solutions of iodides — sodium iodide, potassium iodide, hydrogen iodide, etc. —
in tubes in a powerful centrifugal machine, and found that the two ends of the
tubes acquired charges of opposite sign. The extreme ends of the tubes acquired
a negative charge presumably because of the accumulation there of the heavier
positively charged iodine ions ; and the opposite ends of the tubes acquired a
positive charge presumably owing to the slight excess of positively charged sodium
ions at that end. There is the possibility that the electrification of the tube was
due to the friction against air.
5. Baits ivhichform solid compounds with two or more different amounts of water
of crystallization have different solubilities in their different forms . Hence it is ashed :
Is it not more reasonable to assume that the molecules of the solute exist in solution as
definite hydrates ? The ionic hypothesis answers : Only a definite fractional part
of the salt is ionized, and this part is proportionally less, the more concentrated the
solution. As a rule, in a saturated solution, only a small proportion of the solute is
ionized. A similar observation applies to the existence of liquid crystals. This
does not preclude the possibility that the un-ionized molecules and the ions are
themselves hydrated.
6. When a co7npound is formed from its elements ivith the loss of energy, the coin-
pound cannot be resolved into its ele/tnents unless energy be supplied. It is therefore
pertinent to inquire : What is the source of the energy tvhich leads to the fission of the
molecule into ions carrying equal hat opposite charges of electricity ? Here, again, it
is necessary to reiterate that the ionic hypothesis refers not to the separation of a
compound into its original constituents, but into charged ions ; and it is interesting
to observe that molecules of sodium chloride, etc., which appear to be very stable
when dry, react with great facility when in solution. A little heat is supposed to
be evolved during the ionization of many (not all) electrolytes, and the process of
ionization is then presumably accompanied by an exothermal reaction which more
than compensates for the energy needed for the fission of the molecule into oppositely
charged ions.
There are also hypotheses which suppose ionization occurs by collision.
G. T. Beilby (1905) considers that the ionization is essentially a mechanical operation,
the result of the kinetic activity of the solute molecules, for in a dilute aqueous solu-
tion of, say, hydrogen chloride, each molecule of the solute is surrounded by and
" at the mercy of " some millions of water molecules, all in a state of intense activity,
and the rude mechanical jostling to which the molecule of hydrogen chloride is
subjected will naturally tend to break it up into sim])ler portions mechanically more
stable. J. Kendall ^ made a similar suggestion, but B. de Szyskowsky and J. Perrin
974 INORGANIC AND THEORETICAL CHEMISTRY
regard this hypothesis as untenable — the latter says la prohahilite de rupture d^une
molecule ne depend pas deji chocs qu'elle subit.
W. Nernst and J. J, Thomson have found that the ionizing power of a solvent is
related with its specific inductive capacity or dielectric constant as indicated in the
chapter on " Water." W. Nernst has pointed out that water has a higher specific
inductive capacity than other liquids, and that liquids like methyl alcohol, formic
acid, and others, which, as solvents, give solutions having electrolytic conductivity
also have high specific inductive capacities. From this it is argued that the dis-
sociating power of a solvent, or its power of producing ions, is greater the greater
its specific inductive capacity. C. B. Thwing's numbers show that the dielectric
constants, K, of the hydrocarbons and non-asssociated liquids approximate to
K=2'6D, where D is the specific gravity. Other liquids, particularly those which
contain hydroxyl groups, have higher values than correspond with this rule,
and H. Crompton has shown that if i be the association factor, C. B. Thwing's
data can be represented by K=2'6Di^, or i=^{KI2'6D), where the values of i so
obtained run quite parallel with the values obtained by I. Traube, but are a little
higher. R. Abegg also showed that the temperature coefficient of the dielectric
constant is very small for non-associated liquids— there is only a slight change between
15° and —80° for toluene and ether, but with other liquids there is a larger change
as the temperature is lowered, probably due to increasing density and increasing
association — with ethyl and amyl alcohol, and acetone, the change is quite marked.
H. Crompton therefore argues that it is almost impossible to doubt that association
plays an all-important part in determining the specific inductive capacity of a liquid,
and that if there is any connection between specific inductive capacity and the power
of ionization, it may be looked for rather in the fact that electrolytes are solutions
of approximately non-associated salts in an associated solvent than in there being
any peculiar ionizing power attaching to the solvent. According to P. Dutoit and
E. A. Aston, and P. Walden, the association of the solvent may in turn be referred
to the unsaturated character, of the molecules of that menstruum. H. Crompton
added : The supposed decrease in electrolytic dissociation with rising temperature
is also accounted for, and means nothing more than the decrease in the association
of the solvent. What has been termed the ionic fluidity of a given solution no
doubt increases continually with rising temperature, and, therefore, up to a certain
point, the conductivity increases. But the association of the solvent is also con-
tinually decreasing, and in the end this effect will become the more powerful, the
conductivity, therefore, never rising above a certain maximum value. This
explains the maximum in the molecular conductivity with rising temperature, first
noticed by S. Arrhenius.
G. Ciamician (1890) attributed ionization to the attraction hetiveen the solvent and
the positive and negative "parts of the salt. The idea is supported by the fact that those
solvents which ionize readily are the very ones which unite readily with the salts as
water of crystallization, alcohol of crystallization, etc. Again, not a great many
salts whose ions are univalent separate with water of crystallization while those
with ions of higher valency nearly always do. This is taken to mean that possibly
univalent ions are less hydrated than those of higher valency. This corresponds
with experiments on the speeds at which ions move through the solvent — univalent
ions generally travel faster than the bivalent ions, and the latter in turn go
faster than tervalent ions. However, this evidence is inconclusive, since it does
not follow that because such compounds separate from a solution therefore these
compounds exist in solution. J. D. van der Waals (1891) expressed a similar idea
and considered that the heat of hydration of the ions furnished the energy needed
for the ionization of the salt. To this, D. Konovalof! (1893) adds, " only those
solvents which react chemically with the solute furnish solutions which conduct
electricity," and R. Abegg (1899) expressed it, " the degree of ionization depends
upon the capacity of the ions to unite with the molecules of the solvent."
In the early days of the ionic hypothesis it was assumed that the solvent acts
ELECTROLYSIS AND THE IONIC HYPOTHESIS 975
like a passive medium preventing the re-combination of the ions ; and it came as
an antithesis to the so-called hydrate theor}^ in which the solvent was regarded as
intensely active, and formed complex molecular systems with the solute. The one
hypothesis assumes that simplification (ionization) precedes chemical combination,
the other assumes that complication (association) precedes chemical action. The
accumulation of new facts rendered it necessary for the former to borrow from the
latter, so that " the hydration of ions in aqueous solutions is admitted by almost
all advocates of the ionic theory." It was A. Kekule who, in his Ueher die Consti-
tution unci die Metamorphosen der chemischen Verhindung (1858), emphasized that
molecular coalescence is a preliminary condition for many reactions ; the complex
may then undergo a reversible change ; molecular re-arrangement ; or it may
dissociate into more stable components. In the reaction between ammonia and
hydrogen chloride water plays an essential part, here the molecular coupling is
probably dependent on unsaturated valencies of water and ammonia ; it has been
suggested that the water first reacts with the ammonia :
H3N= +=OH2 -> H3N=OH2 -> NH4OH
and that the acid and base then react either through their ions or molecularly :
^^H>0 = + =C1H -> ^4j[^>0<^^ -> H4N.CI+H.OH
With water and calcium oxide CaO = +=OH2->CaO=OH2->Ca(OH)2. The
same view was emphasized by J. H. van't Hoff in 1878, and by H. E. Armstrong in
1891. The addition compound brings together elements, previously separate, into
one common sphere of activity within which it is possible for them to interact.
J. Kendall and J. E. Booge now consider ionization to be preceded by a combination
between solvent and solute. Any unsaturated solvent possesses the power of
forming complexes with itself by association and also with any unsaturated solute.
It is assumed that the attractive forces holding together the radicles of the solute
are now so diminished that ionization can take place. The union of the solute
with molecules of the solvent thus promotes ionization, or the dissociation of the
complex into radicles of opposite charge. They add :
The mechanism of the electronic hypothesis of J. J. Thomson, G. N. Lewis, W. A. Noyes,
etc., is based on the assumption that an unsaturated molecule contains electrons not rigidly
fixed but free to move under the influence of an electric field- — e.g. the oxygen atom in
a water molecule has at least two electrical doublets H+ — > ~0~ <— +H, which can move
into such a position or orbit as to exert a maximum attraction on the positive part of the
doublet, thus producing an unequal distribution of the electrical charges. When two such
molecules come into close proximity, this inequality will be greatly accentuated, and the
mutual attraction may be sufficient to enable complex molecules to be formed by association.
The associated molecule possesses a larger electrostatic moment, and the constraints on
the electrons are accordingly weakened. The higher the degree of association, the larger
the electrostatic moment, the weaker the constraints holding the charges, and the higher
the dielectric constant. Such associated molecules may exert considerable attractive
forces on the molecules of a solute especially when the latter also furnish strong fields of
electric force. As a result, the complex may be much less stable than the simple component
molecules, and dissociation occurs.
In conformity with the hypothesis that the formation of complexes between
solvent and solute is the immediate cause of ionization, J. Kendall, J. E. Booge, and
J. E. Andrews found that conducting solutions invariably afford evidence of the
formation of complexes — e.g. the formation of stable hydrates is characteristic of
strong acids, and the less the strength of the acid, the less the tendency to form
stable hydrates. There is no indication of hydrate formation with weak organic
acids unless the acid has also the character of a phenol base. Similar remarks
apply to the bases. In general, therefore, there is a uniform increase in the ionization
with the acids and bases most readily hydrated.
The proposition has not yet been established with salts, although A. Werner
976 INORGANIC AND THEORETICAL CHEMISTRY
states that the stability of an aquo-salt decreases as the strength of the acid or base
increases ; but R. Abegg and G. Bodlander hold that salts of weak acids and bases
are most extensively hydrated. With the mercuric salts also, all the highly ionized
salts yield hydrates, and all slightly ionized salts are not hydrated. If ions are
formed at all they art* complexes formed by association with the molecules of the
solvent. The ionic theory primarily assumes that the apparent numhcr of solute
" molecules " is increased by ionization, and in a general sense it may be said that
it makes very little difference to the applications of the ionic theory whether it be
assumed that each ion is isolated as a distinct individual, or whether each ion forms
a complex with the molecules of the solvent. The number of ions is the same in
both cases.
The explanation of the phenomena, particularly when solutions other than water
are considered, is beset with many difficulties on account of the complex relations
between the solvent and solute. So much is this the case that M. le Blanc (1907)
considered " it to be very fortunate for the advance of electrochemistry that such
complications are generally, though not always, absent in the case of the aqueous
solutions. It is due to this fact that it has been possible to deduce simple laws from
the study of aqueous solutions." On the other hand, 0. N. Witt (1901) has said :
Water is a very complicated substance, and the process of solution in that liquid must be
attended by very great exceptions from the simple rules which exist for other solvents, not
so complicated. It has therefore been a great mistake to study aqueous solutions and then
other solutions. The chapters in treatises on physical chemistry entitled " Theory of
Solution " should be rightly entitled " Theory of Aqueous Solution."
Several attempts have been made to work out a consistent explanation of the
fundamental facts without a theory of charged ions, but with hypotheses based upon
the formation of imaginary molecular complexes by a reaction between polymerized
solvent and the molecules of the solute.
These controversial matters emphasize the fact that an explanation of a
phenomenon may contain part of the truth, and yet not " the whole truth, and
nothing but the truth." In that case, we try the hypothesis by the test indicated
in the first chapter, and ask : Is the hypothesis useful ? The answer is that the
ionic hypothesis has done good work, and it promises to do more. An hypothesis
is not always to be discarded as a first approximation, because troublesome exceptions
crop up from time to time. Newton's theory of gravitation, for instance, appeared
to be afflicted with such blemishes — particularly in its early days ; so was the theory
of opposing reactions once considered to be unreasonable folly ; and the present-
day theory of light seems highly absurd when it is remembered that it is based upon
the existence of an aether pervading all space, an aether which is of the highest
elasticity, and denser than steel. In spite of important difficulties, we shall try
how the ionic hypothesis fits in with a few important phenomena.
References.
1 C. J. T. von Grotthus, Gehkn's Journ., 5. 816, 1808 ; Ann. CMm, Phys., (1), 58. 64, 1806 ;
(1), 63. 20, 1808 ; Ostwald's KUssiker, 152, 1906.
2 H. von Helmholtz, Journ. Chem. Soc, 39. 277, 1881 ; Proc. Roy. Sor. Edin., 12. 596, 1884 ;
Pogg. Ann., 150. 483, 1873 ; Sitzber. Akad. Berlin, 587, 1873.
3 G. F. Fitzgerald, B.A. Rep., 326, 1900 ; H. E. Armstrong, Journ. Chem. Soc, 67. 1122, 1895.
* H. von Helmholtz, IJeher die ErJmltung der Kraft, Berlin, 1847 ; Wied. Ann., 11. 737, 1880 ;
Journ. Chem. Soc, 39. 277, 1881 ; O. J. Lodge, B. A. Rep., 723, 1885.
«» J. H. Poynting, Phil. Mag., (5), 42. 289, 189() ; J. D. van der Waals, Zeit. phys. Chem., 8. 215,
1891 ; G. Ciamician, ih., 6. 401, 1890 ; 69. 100, 1909 ; A. Hantzsch, ih., 61. 307, 1907 ; A. Sachanoff,
ib., 80. 20, 1912 ; 1>. Konovaloff, Wied. Ann., 49. 733, 1891 ; A. Werner, Zeit. anorg. Chem., 3. 294,
1893; R. Abegg, ih., 39. 330, 1904; J. Walker, Journ. Chem. Soc, 85. 1082, 1904; J. Walker,
D. Mcintosh, and P]. Archibald, ih., 85. 1098, 1904 ; E. Fitzgerald and A. Lapworth,?7>., 93. 2ir)3,
2200, 1908 ; 107. 857, 1915 ; W. R. Rousfield, ih., 105. 1809, 1914 ; W. R. Bousfiekl and T. M. Lowry,
Phil. Tra7i.«., 204. 281, 1914 ; Tramf. Faraday Soc, 1. 197, 1905 ; 3. 125, 1907 ; M. M. (Jarver, Journ.
Phys. Chem., 14. 651, 1910; J. M. Nelson and K. G. Falk, Journ. Amer. Chem. Soc, 27. 285,
ELECTROLYSIS AND THE IONIC HYPOTHESIS 977
17:j;i, I'Jlo ; J. Kendall and J. E. Booge, ib., 39. 2323, 1917 ; J. Kendall, ib., 36. 1073, 1914 ;
J. Kendall, J. E. Booge, and J. C. Andrews, ib., 39. 2303, 1917 ; E. W. Washburn, ib., 31. 322,
1909 ; G. Carrara, Electrochemie der nichtwdssrigen Losungen, Stuttgart, 1908 ; S. Arrhenius,
Tlieories of Chemistry, London, 83, 1907 ; Theories of Solution, New Haven, 184, 1912 ; A. Kekul6,
Liebig's Ann., 106. 129, 1858 ; J. H. van't Hoff, Ansichten iiber organische Chemie, Braunschweig,
1. 224, 225, 1878 ; R. Clausius, Pogg. Ann., 101. 347, 1875 ; S. Arrhenius, Eecherches sur la
conductibilite galvanique des electrolytes, Stockholm, 1884 ; B. A. Rep., 357, 1886 ; P. Kohlrausch,
Wied. Ann., 26. 161, 1885 ; B. A. Rep., 334, 1886 ; R. C. Tolman, Journ. Amer. Chem. Soc, 33.
121, 1911; W. Ostwald, Lehrbuch der allgemeinen Chemie, Leipzig, 1886-7 ; Elektrochemie
Geschichte und Lehre, Leipzig, 1896 ; W. Nernst, Theoretische Chemie, Stuttgart, 1907 ; London,
1916 ; Zeit. phys. Chem., 2. 613, 1888 ; 3. 372, 1889 ; M. le Blanc, Lehrbuch der Elektrochemie,
Leipzig, 1904 ; New York, 1907 ; R. Lupke, Grundzuge der Elektrochemie, Berlin, 1907 ; London,
1903 ; H. Jahn, Grundriss der Elektrochemie, Wien, 1905 ; R. Abegg, Theorie der electrolytischen
Dissoziation, Stuttgart, 1903 ; J. J. van Laar, Lehrbuch der theoretischen Elektrochemie, Leipzig,
1907 ; S. Arrhenius, Lehrbuch der Elektrochemie, Leipzig, 1901 ; London, 1902 ; F. Foerster,
Elecktrochemie wdsseriger Losungen, T^eipzig, 1905 ; R. A. Lehfeldt, Electrochemistry, London,
1904 ; Zeit. phys. Chem., 1. 75, 1887 ; 2. 270, 840, 1888 ; 3. 170, 241, 369, 1889 ; J. H. van't HofE
and T. L. Reicher, ib., 2. 777, 1888 ; 3. 198, 1889 ; M. Lob and W. Nernst, ib., 2. 962, 1888 ;
A. W. Williamson, Phil. Mag., (3), 37. 350, 1850 ; Proc. Roy. Inst., 1. 90, 185 ; B. A. Rep., 65,
1850 ; G. F. Fitzgerald, Journ. Chem. Soc, 69. 885, 1896 ; H. Davy, Collected Works, London, 8.
346, 1840 ; H. E. Armstrong, Nature, 55. 78, 1896 ; 49. 100, 1893 ; B. A. Rep., 962, 1885 ; Proc.
Roi/. Soc, 40. 287, 1886 ; 70. 90, 1902 ; 74. 86, 1904 ; Journ. Chem. Soc, 49. 112, 1886 ; 67.
1122, 1895 ; 83. 1088, 1903 ; Proc Chem. Soc, 1. 39, 1885 ; 8. 22, 1892 ; Encyc Brit., 26. 740,
1902; C. B. Thwing, Zeit. phys. Chem., 14. 286, 1894 ; W. Nernst, ib., 14. 622, 1894 ; P. Walden,
ib., 46. 103, 1903 ; H. Crompton, Journ. Chem. Soc, 71. 925, 1897 ; I. Traube, Ber., 30. 265,
1897 ; R. Abcgg, Wied. Ann., 60. 54, 1897 ; J. J. Thomson, Phil. Mag., (5), 36. 320, 1894 ;
P. Dutoit and E. A. Aston, CompL-Rend., 125. 240, 1897.
§ 6. The Electrolytic Conductivity of Solutions
In electrolytic conduction, the electricity does not slip through between the molecules,
it goes wdth them. The constituents of each molecule are free from one another, and while
one set of atoms conveys positive electricity, the other set conveys negative electricity in
the opposite direction, and so it is by a procession of free atoms that the current is trans-
mitted. The atoms act as carriers. The free locomotion of charged atoms is essential for
electrolysis.— O. J. Lodge (1892).
The process of electrolysis, according to the ionic hypothesis, is supposed to
proceed somewhat as follows : When a salt — say, sodium chloride — is dissolved in
water, (i) some of the dissolved molecules are, by hypothesis, ionized, and the
ions immediately begin to recombine to form molecules. The speeds of the two
reactions are supposed to behave analogously with those of opposing reactions,
and a state of equilibrium is reached when the number of molecules reformed by
the combination of the ions is equal to the number of molecules ionized in the same
time. Still further, (ii) when the two poles of a battery — say platinum electrodes —
are dipped in a solution of sodium chloride, all the chlorine ions, carrying a negative
charge, are attracted to the anode or positively charged electrode, and the positively
charged sodium ions are attracted to the cathode or negatively charged electrode.
If a constant difference of potential be maintained between the electrodes of the
battery, as each ion comes in contact with the electrode with a charge of opposite
sign to its own, the charge is torn from the ion which thereby reverts to an ordinary
atom of chlorine or sodium. The battery reproduces the same difference of potential
as before by generating more electricity ; this is again discharged at the electrodes ;
and so, by an alternate process of charge and discharge, electrolysis continues,
(iii) The chlorine atoms, being unable to attack the water or the electrode, unite
in pairs to form molecules of chlorine gas. As soon as the liquid in the vicinity of
the anode is saturated with chlorine, this gas bubbles to the surface of the liquid.
Similarly, the sodium ions are relieved from their charges at the cathode, and the
resulting sodium atoms immediately attack the water, forming hydrogen gas and
sodium hydroxide: 2Na-f2H20=2NaOH+H2 ; the hydrogen bubbles off as a
gas. (iv) The equilibrium between the un-ionized molecules and the ions is disturbed
by the annihilation, so to speak, or the removal of ions at the electrodes. The
VOL. I. 3 R
978 INORGANIC AND THEORETICAL CHEMISTRY
difEerence of potential at the electrodes is maintained by the battery, and the supi)ly
of ions is kept up by the steady ionization of the salt as fast as the ions are de-
electrified at the electrodes, until practically the whole of the salt in the solution
has been electrolyzed. The charges drag the atoms to the electrodes, and only at
the electrodes can the charges be torn from the atoms. Thus A. Smith (1890) said
that " the ions do not transport the electricity of the battery, but their own."
The charged ions are already present in the solution before connection is made with
the battery.
Usually, the electrical conductivity of a solution is measured indirectly. The resistance
which the solution offers to the passage of a current is directly measured. It is more
convenient to take the reciprocal of the resistance and call it the conductivity of the
solution. The specific resistance is first determined, that is, the resistance in ohms which
is equivalent to the resistance of a cubical mass of the solution whose length of side is
1 cm., Fig. 5, The reciprocal of this quantity in reciprocal ohms, is the specific conductivity.
Hence the specific conductivity, *c, represents the current in amperes which is produced in
a cube of one centimetre side when a potential difference of one volt is applied to the opposite
faces of the cube. From this, the so-called equivalent conductivity— symbolized A — is
calculated. The equivalent conductivity of a substance represents the conducting power
of one gram-equivalent of the substance dissolved in the solvent, and placed in a cell whose
opposite walls, one centimetre apart, form the electrodes. Otherwise expressed, the
equivalent conductivity represents the conducting power of a layer of the solution 1 cm.
thick, and containing one gram-equivalent of the substance in solution ; or, the quantity of
electricity which, under a potential difference of one volt, passes per second between electrodes of
indefinite extent, and one centimetre apart, between which is placed that quantity of solution which
contains one equivalent weight of the ionizing substance. If the conductivity be referred to a
gram-molecule, and not a gram-equivalent, it is termed the mole-
cular conductivity of the solution- — symbolized /x.
C. J. Reed^ has pointed out the need for emphasizing the fact
that electrical conductivity is a property of matter like transparency,
'icm/ diathermacy, magnetic permeability, and heat conductivity. The
electrical conductivity of different substances is measured or com-
Fig. 5. pared by employing a unit in which there is a definite section across
which the flux is measured, and a definite length in the direction of
the flux. Units of this kind are defined by certain relations of form, in which the ratio of
length to cross-section is a fixed ratio. He applies the terms conductance and resistance
to bodies of definite form, length, and cross-section, while the terms conductivity and
resistivity refer to properties of matter independent of shape, or form, or quantity. Con-
ductivity is therefore specific conductance ; and resistivity is specific resistance. The two
ideas, however, are commonly merged in the one term conductivity or resistivity as the
case might be.
When an electric current is passed through a cell composed of two metallic
electrodes and an electrolyte, the total opposition to the passage of the current is made
up of the resistance of the cell and the back electromotive force at the electrodes,
which latter is the sum of the over- voltages at the two electrodes. The resistance
of the cell is made up of the resistance of the electrolyte and the transfer resistance
from electrode to electrolyte. The transfer resistance has been particularly studied
by E. Newbery, etc.2 The conditions which favour high transfer resistances are
(i) low current density ; (ii) low temperature ; (iii) polished electrode surfaces ;
and (iv) high over-voltages.
References.
1 C. J. Reed, Trans. Atner. Electrochem. Soc, 5. 103, 1904.
2 G. Gore, Proc. Roy. Soc, 38. 209, 1885 ; C. F. Burgess, Tratis. Amer, Electrochem. Soc, 7
61, 1905 ; W. S. Franklin and L. A. Freudenberger, ib., 7. 33, 1905 ; 8. 227, 1905; H. J. S. Sand
and T. P. Black, Zeit. phys. CJiern., 70. 496, 1910 ; E. Newbery, Trans. Faraday Soc, 15. 126, 1919
§ 7. The Number of Ions in a Solution
The greater the number of ions in a liquid the better is the liquid likely to conduct
electricity .—0. J. Lodge (1892).
If water be progressively added to an aqueous solution of hydrogen chloride,
containing, say, one gram- equivalent (36'5 grams) per litre, at 18°, the equivalent
ELECTROLYSIS AND THE IONIC HYPOTHESIS
979
conductivity of the solution gradually increases as illustrated by the following
numbers, when v denotes the number of litres of solution containing one gram-
molecule of the solute, and A the corresponding equivalent conductivity :
v= 2
A-305-4
8
328-5
16
331-5
32
342-3
128
349-1
512
349-3
1024
349-3 units
150
::: ::: :::±: !:::_:
±-±----
lOOzi::::
= :::::::::::::::::::"::: =
=:::::=::::::::=::
J
n ::.:::
±_: :;_:;:!_::
200 400 600
1000 1200 1400 mo 1800 2000
These numbers show that the electrical conductivity of the solution increases
until a certain limit is reached. Subsequent additions of water have no further
influence on the equivalent conductivity of the solution. This is further emphasized
by the curve, Fig. 6, which
represents the rapid rise in
the equivalent conductivity
of sodium chloride solutions
with decreasing concentra-
tion ; the conductivity reaches
a maximum very quickly,
when the further additions
of water have no further ^m. 6
influence on the result. Hence
the equivalent conductivity
of an electrolytic solution increases with dilution, reaching a maximum value
approximately corresponding with infinite dilution. This fact is called F. Kohl-
rausch's first law. When the dilution has reached the limit beyond which no
further increase in the equivalent conductivity can be observed, it is supposed that
the salt is all ionized, and no more ions can be supplied by the solute, however
much more solvent be added. All the ions which can be obtained from the solute
take part in conducting the electric current at infinite dilution.
The determination of the electrical conductivity of a solution.— Let i? denote the resistance
of the solution expressed in ohms. The cell containing the solution^ — the conductivity cell — is
Effect of Concentration on the Equivalent
Conductivity of Aqueous Solutions of Sodium Chloride.
Key,K
Induction Coil.N
Bridge Wire,a,b
Telephone, T
Fig. 7. — The Determination of the Electrical Conductivity of Solutions.
arranged as illustrated in the sketch and plan. Fig. 7, with a known resistance R (ohms) in the
box of resistance coils ; a " metre " wire of known resistance ; a telephone, T ; a battery, B ;
an electrolytic cell of resistance, S ; a key, K, for starting the current from the battery ; and
an induction coil, 'N , which furnishes a rapidly alternating high potential current. It may
be asked : How can there be anode and cathode when working with an alternating current ?
If the impulse in one direction of the alternating current cannot bridge the gap between
980 INORGANIC AND THEORETICAL CHEMISTRY
the electrodes, and the impulse in the other direction can, the general effect is tliat of an
intermittent current in one direction. If the sliding contact of the bridge, C, be moved
until the telephone is " silenced "• — that is, makes least noise^ — ^the resistances on both
sides of the system are in equilibrium. The readings are most accurate when the resistance
i? is so adjusted that the contact C is near the middle of the bridge. From the principle of
Wheatstone's bridge — discussed in text-books on electricity — it follows that S : E=a : ?>,
where a is the resistance of the bridge wire on the same side as the resistance R, and h the
resistance of the wire on the other side. If the wire be 1000 mm., that is, one metre long,
h = 1000 — o. Consequently,
1 b
Observed conductivity = == - rec. ohms . . • (1)
S Ra
The spreific conductivity* — -the reciprocal of the specific resistance- — nmst be proportional to
the observed conductivity, oi*, the specific conductivity is k times the observed conductivity,
where k is the constant of proportion. Consequently,
« .r. . • . ^ 1000-a
Specific conductivity = - • rec. ohms . . • (2)
R a
The observed conductivity depends upon the capacity of the cell employed, and this,
in turn, depends upon the surface area and the distance ajyart of the electrodes. If these
two magnitudes are known, it is possible to compute the specific conductivity of the solution
from the observed conductivity. It is simpler, however, to determine the constant of the
cell employed, by using a solution of known conductivity. All the factors, except k, are then
measured, and k is calculated from (2). With the same cell, the conductivity of the given
solution can be determined and the specific conductivity computed by (2).
In illustration, the specific conductivity of a j^A^-solution of potassium chloride at
18° is 0'0112. With a resistance of 10 ohms, the telephone was silenced when a = 535 mm.
Hence, from (2), 0*01 12 = (A;/10t( 1000 -535)/535 ; or fc = 0-1289. Hence, for this particular cell,
the specific conductivity =0'1289( 1000— a)/i?a rec. ohms ; with a ^^r^-solution of potassium
chloride for which a = 490, when i? = 12, the specific conductivity =0-1289(1000~490)/12
X 490=0*011199 rec. ohms. If v denotes the number of cubic centimetres containing
one gram -equivalent of the compound under investigation, the quotient obtained by dividing
the specific conductivity by v will represent the specific conductivity of the j'jy^-potassium
chloride at the dilution v, hence, since a ^^^-solution of potassium chloride has one-tenth
of a molecular weight of the salt expressed in grams per litre, one gram-molecule will bs
present in 10 litres, or v = 10,000 c.c, and, since k denotes the specific conductivity, the
equivalent conductivity k/v will be 0-011199/v, or 111*99.
During the measurements, the conductivity cell must be kept at a constant temperature
since the conductivity varies with changes of temperature. Hence, the need for the water-
bath, stirrer, and thermostat (temperature regulator) shown in Fig. 6. If the conductivity
of the solution be large, a form of cell is employed with electrodes further apart than is
the case with the cell shown in the diagram for solvitions of feeble conductivity. Details
are discussed in any laboratory handbook. ^ It is becoming increasingly clear from the work
of E. Newbery, S. F. Acree, etc., that this method of measuring conductivity can be made
much more sensitive and accurate by directing attention on the nature of the electrodes,
and the frequency of the alternating current.
It is now assumed that the number of ions which take part in conducting the
electric current at any particular concentration of the solution is proportional to
the equivalent conductivity, A, of the solution. If the number of ions in a given
solution be doubled, the conductivity will be doubled. Consequently, if a repre-
sents the fraction of a gram-equivalent which is dissociated into ions when the
solution occupies v litres, we have, at dilution v, the conductivity X^—ka, where k
is the constant of proportion. At infinite dilution, the whole gram-equivalent is
supposed to be ionized, and consequently, a=l, and therefore the conductivity,
X^ , at infinite dilution, is A^^^ =k. Substitute this value of k in the preceding equation,
and we get
'-
which, by hypothesis, means that the fractional number of molecules ionized
in a solution is numerically equal to the equivalent conductivity of the solution
divided by the equivalent conductivity of the solution at infinite dilution ; or,
_^ ,. . . Number of molecules ionized Xv
Degree of ionization, a = = .
Total number of molecules A^
ELECTROLYSIS AND THE IONIC HYPOTHESIS
981
This formula enables the electrical conductivity of a salt to be expressed in
terms of the degree of ionization of the salt in solution. Thus, the equivalent
conductivity of a solution of hydrochloric acid is 305'4, and the same acid at infinite
dilution has the equivalent conductivity 349 '3. Hence, the degree of ionization
is 305'4— 349*3=0874 per gram-equivalent, or 87'4 per cent, ionization. This
means that 12*6 per cent, of the molecules of the solution are present as un-ionized
neutral molecules, HCl ; and 87 '4 per cent, of the molecules are present in the ionic
formH-+Cr. Or,
HCl H--fCl'
12*6 per cent. ^87 '4 per cent.
The percentage ionization must not be confused with the absolute concentration
of the ions. The former may be the greater in dilute solutions, and the latter
greater in concentrated solutions. The ionic hypothesis thus assumes that an
aqueous solution of hydrochloric acid contains three distinct kinds of " solute
molecules," electrically charged molecules (hydrogen and chlorine ions), and neutral
hydrogen chloride molecules.
Table I shows the degree of ionization of normal solutions (unless otherwise
stated) of a few typical acids, bases, and salts selected merely for illustrative
purposes.
Table I.^— Degree of Ionization of Some Typical Acids, Bases, and Salts.
Acids.
Bases.
Salts.
g-d
o-d
o-d
Acid.
0.2
Base.
ti.§
Salt.
•ii,H
II
m
11
Nitric acid (62%) .
0-096
Potassium hydroxide
0-77
Potassium chloride
0-74
Nitric acid (dil.)
0-820
Sodium hydroxide .
0-73
Ammonium chlo-
Sulphuric acid (dil.).
0-510
Lithium hydroxide.
003
ride .
0-75
Carbonic acid {^^N).
0-002
Ammonium hydroxide
0-01
Potassium nitrate
0-64
Hydrosiilphuric acid
Calcium hydroxide
Zinc sulphate
0-24
(tV^) . . .
0-001
(bW)
0-90
Copper sulphate .
0-22
Perchloric acid {IN)
0-880
Barium hydroxide
Silver nitrate
0-58
Acetic acid {^^N)
0-013
(itW)
0-92
Barium chloride .
0-57
Trichloracetic acid
■
Methylamine
0-12
Potassium sul-
i^^N) .
0-850
Ethylamine .
0-13
phate
0-24
The effect of increasing the concentration of a solution is to increase the internal
friction. This retards the movements of the ions and thus diminishes the conduc-
tivity more rapidly than would occur if the results were not affected by this disturbing
factor. As the concentration decreases, the friction diminishes ; and, with the more
dilute solutions, the effects of internal friction can be neglected. The two factors —
internal friction and conductivity — do not change with dilution in the same way,
and, in consequence, the conductivity may increase with increasing dilution ;
reach a maximum ; and then decrease with increasing dilution. This is the case,
for instance, with sulphuric acid, where the maximum conductivity occurs when
30 per cent, of acid is present. The application of the ionic theory to concentrated
solutions is beset with many difficulties, and consequently the theory has been
mainly developed from results obtained with dilute solutions.
Strong and weak acids and bases. — The terms '' strong " and " weak " are
sometimes applied to the acids and bases, and these terms refer to the conductivity
or to the degree of ionization in aqueous solution of moderate dilution. A strong
acid or base has a high conductivity and accordingly a high degree of ionization at
moderate dilutions ; while the converse is the case with a weak acid or base. There
982 INORGANIC AND THEORETICAL CHEMISTRY
is no real line of demarcation between the two. Acids like carbonic and hydro-
sulplfuric acids, and bases like ammonia, are weak. Their degree of ionization is
less than one per cent. If the degree of ionization exceeds 70 per cent, the acid is
undoubtedly strong. Electrolytes like solutions of sodium chloride are good
conductors, and some solutions with a conductivity midway between good con-
ductors and non-conductors are sometimes called semi- or half-electrolytes. Most
salts are highly ionized, even at moderate dilutions, but there are many exceptions,
e.g. mercuric chloride, the cadmium halides, and mercuric cyanide are but slightly
ionized in moderately dilute solutions.
W. Ostwald - noted that, as a rule, the ionization of analogous salts is the greater
the further apart the anions and cations are in the electrochemical series. R. Abegg
and G. Bodlander found that the tendency to form complex ions increases as the
tendency of the salt to ionize decreases. The electrical conductivity of aqueous
solutions increases with a rise of temperature» and F. Kohlrausch ^ represented his
observations at 0° by the formula /i^=^o{l+«(^~^o)+M^~^o)"K where ^o repre-
sents the initial temperature, in the present case 18°. The constants a and h are
characteristic of each solution of a given electrolyte. Thus, for nitric acid, a=0"0163,
6=— 0-000016 ; potassium chloride, a=-0-0197, 6=+0-000047 ; etc. F. Kohlrausch
suggested that on a falling temperature, the conductivities of all aqueous solutions
approach zero at about the same temperature, because of the decreasing fluidity
or increasing viscosity of the solvent. J. Kunz found this temperature approxi-
mates to —40°. Owing to the opposite effects of viscosity and temperature on
conductivity, negative coefficients might be anticipated, and S. Arrhenius found
this to be the case with phosphoric acid. There is also a maximum in the curve
with many salts — and the more concentrated the solution the lower the temperature
at which a maximum occurs. There does not appear to be any anomaly in the
conductivity as the temperature passes 4°, or even with undercooled solutions in
passing through the freezing point.
The effect of pressure has been studied by G. Tammann, J. Fanjung, A. Bogo-
jawlensky, etc.* If k represents the ionization constant ; p, the pressure ; R, the
gas constant ; T, the absolute temperature ; and dv, the decrease in volume which
occurs during the ionization of a gram-molecule of the salt, then, M. Planck's formula
at constant temperature, {d log k)ldp=dvlRT, describes the results of observation.
G. Wiedmann & noticed a relation between the conductivity and the viscosity
of a solution, and 0. Grotrian noted that there is a parallelism between the two
variables, but not strict proportionality. According to H. von Euler, the viscosity
7) oi a. solution of a binary electrolyte is 'r]=r]i^^^~'^\T]2y)s)^''^, where a represents
the degree of ionization ; t^j, r)2, and 7^3 respectively denote the viscosity coefficients
of the un-ionized salt, the cation, and the anion ; and C denotes the total con-
centration. If the electrolytic mobilities of the ions be plotted as abscissae, and the
viscosity coefficients as ordinates, the values for most of the salts lie on a straight
line. M. G. Levi did not find that the stiffening of a solution with gelatine made an
appreciable difference on the conductivity, but in very concentrated gelatine
solutions, E. Wiedemann, C. Liideking, and B. von Tietzen-Hennig found that the
conductivity is diminished ; similar observations were made by C. Stephan,
S. Arrhenius, and P. Massoulier with respect to solutions containing sugar or
glycerol. W. von Beetz, B. von Tietzen-Hennig, and M. Oker-Blum studied the
influence of suspensions of sand, gypsum, and blood corpuscles on the conductivity
of salts.
J. Bosi,6 E. H. Hall, and J. Nabl investigated the effect of the agitation of the
solution on the conductivity, but no marked difference was observed. F. Neesen,
H. Bagard, and G. Melani studied the effect of magnetism on the conductivity but
without establishing any definite conclusion. C. H. Wind. E. van Everdingen,
H. Bagard, F. Florio, and F. Chiavassa have studied the Hall effect with electrolytes.
F. Kohlrausch, H. M. Goodurn, A. Wiener, and A. Miolate, etc., 7 noticed that some
electrolytes when diluted, mixed together, etc., show a transient change in the
ELECTROLYSIS AND THE IONIC HYPOTHESIS 983
conductivity ; this is attributed to chemical action. Light may also set up chemical
reactions. According to J. Gibson, light always acts in such a way as to increase
the conductivity, but whether the effect is reversible is not clear. The effect of
X-rays has also been examined.^
References.
* W. Ostwald and R. Luther, Hand- und Hillfshuch zur Ausfilhrung physiko-chemischer
Messiingen, Leipzig, 1902; London, 1894; A. Findla,y, Practical Physical Cheynistr y,Ijondon.,
1906.
2 W. Ostwald, Lehrbuch der allgemeinen Chemie, Leipzig, 2. i, 791, 1903 ; R. Abegg and
G. Bodlander, Zeit. anorg. Chem., 20. 471, 1899.
3 F. Kohlrausch and 0. Grotrian, Pogg. Ann., 154. 1, 215, 1875 ; 0. Grotrian, ib., 151. 378,
1874 ; F. Kohlrausch, ib., 159. 233, 1876 ; W. van Beetz, ib., ill. 1, 1862 ; P. Sack, Wied. Ann.,
43. 212, 1891 ; R. J. Holland, ib., 50. 349, 1893 ; E. Doen and B. Volhner, ib., 60. 468, 1897 :
D. Dennhardt, ib., 67. 325, 1899 ; F. Kohlrausch, ib., 6. 28, 1879 ; C. Heim, ib., 27. 643, 1886 ;
C. Deguisne, ib., 52. 604, 1894 ; Temperatur Koeffizienten des Leitvermogens sehr verdilnnier
imssriger Losungen, Strassburg, 1893 ; E. Krannhals, Zeit. phys. Chem., 5. 250, 1890 ; H. Jahn,
ib., 16. 72, 1895 ; M. Rudolphi, ib., 17. 277, 1895 ; J. Kunz, ib., 42. 591, 1903 ; H. von Euler,
ib., 21. 257, 1896 ; C. Schaller, ib., 25. 497, 1898 ; H. von Steinwehr, ib., 38. 185, 1901 ; E. Cohen,
ib., 31. 164, 1899 ; F. Kohlrausch, ib., 44. 197, 1903 ; W. Bottger, ib., 45. 521, 1903 ; M. Maltby,
ib., 18. 155, 1895 ; A. A. Noyes and W. D. Coolidge, ib., 46. 323, 1903 ; G. Foster, Phys. Rev., (1),
8. 258, 1899 ; L. Kahlenberg, Journ. Phys. Chem., 5. 339, 1901 ; H. C. Jones and J. M. Douglas,
Amer. Chem. Journ., 26. 428, 1900 ; F. L. Kortright, ib., 18. 365, 1896 ; H. C. Jones and
E. Mackay, ib., 19. 83, 1891 ; H. C. Jones, The Electrical Conductivity, Dissociation, and Temperature
Coefficients of Conductivity, Washington, 1912 ; R. T. Lyle and R. Hosking, Phil. Mag., (6), 3.
487, 1902 ; G. E. Hulett and L. E. Allen, Journ. Amer. Chem. Soc, 24. 667, 1902 ; A. Hantzsch and
W. B. Davidson, Ber., 31. 1612, 1898 ; A. Hantzsch, ib., 32. 3066, 1899 ; J. Guinchard, ib., 32.
1732, 1899 ; R. Abegg, ib., 33. 393, 1900 ; F. Kohlrausch, Sitzber. Akad. Berlin, 1026, 1901 ;
572, 1902 ; P. Rivals, Compt. Rend., 125. 574, 1897 ; J. Kunz, ib., 135. 788, 1902 ; A. Hagenbacli,
A7in. Physik, (4), 5. 276, 1901 ; J. C. H. Kramers, Arch. Neerl., (2), 1. 455, 1898 ; S. Lussana,
Nuovo Cimento, (3), 36. 41, 1894 ; C. Deguison, ib., (4), 1. 59, 1895 ; T. Gnesetto, Atti 1st. Veneto,
(2), 59. 987, 1900 ; M. Pacher, ib., (2), 58. 785, 1899 ; H. M. Dawson and P. WilUaras, Zeit. Electro-
chem., 6. 141, 1899 ; W. R. Bousfield and T. M. Lowry, Proc. Roy. Soc, 70. 42, 1902 ; F. Kohl-
rausch, ib., 71. 338, 1903.
" G. Tammann, Wied. Ann., 69. 767, 1899 ; M. Planck, ib., 32. 494, 1887 ; J. Fink, ib., 26. 481,
1885 ; F. Braun, ib., 30. 250, 1887 ; Zeit. phys. Chem., 1. 259, 1887 ; G. Tammann, ib., 17. 725,
1895 ; G. Tammann and A. Bogojawlensky, ib., 27. 457, 1898 ; J. Fanjung, ib., 14. 673, 1894;
B. Piesch, Sitzber. Akad. Wien, 103. 784, 1894 ; S. Lussana, Nuovo Cimento, (4), 2. 263, 1895 ;
G. Foussereau, Compt. Rend., 104. 1161, 1887; C. Barus, Amer. Journ. Science, (3), 40. 219,
1890 ; E. Cohen and W. Schut, Piezochemie kondensierten Systeme, Leipzig, 1919.
« a Wiedmann, Pogg. Ann., 99. 177, 1856 ; 0. Grotrian, ib., 157. 130, 1876 ; 160. 238, 1877 ;
Wied. Ann., 8. 529, 1879; C. Stephan, ib., 17. 673, 1882; F. Kohlrausch, ib., 6. 196, 1879;
E. Wiedemann, ib., 20. 537, 1883 ; B. von Tietzen-Hennig, ib., 35. 467, 1888 ; C. Ludeking, ib.,
37. 172, 1889 ; W. von Beetz, ib., 26. 20, 1884 ; H. von Euler, Zeit. phys. Chem., 27. 536, 1898 ;
8. Arrhenius, ib., 9. 487, 1892 ; P. Massoulier, Com^^t. Rend., 130. 773, 1900 ; M. G. Levi, Gazz.
Chim. Ital, 30. ii, 64, 1900; W. Oker-Blum, Archiv. gesamt. Physiol, 79. Ill, 510, 1900; 81.
167, 1900.
« J. Bosi, Nuovo Cimento, (4), 5. 249, 1897 ; F. Florio, ib., (4), 4. 106, 1896 ; G. Melani, ib., (4),
6. 191, 1897; F. Chiavassa, ib., (4), 6. 296, 1897: H. Bagard, ib., (4), 7. 187, 1898; Compt.
Rend., 122. 77, 1896 ; 129. 152, 1899; E. van Everdingen, Versl. Akad. Amsterdam, 7. 46, 1898;
C. H. Wind, Verh. Akad. Amsterdam, (1), 5. 499, 1896 ; J. NabI, Anz. Akad. Wien, 356, 1899 ;
E. H. Hall, Phys. Rev., (1), 38. 246, 1898 ; F. Neesen, Wied. Ann., 23. 482, 1884.
' F. Kohlrausch, Zeit. phys. Chem., 12. 773, 1893 ; 33. 259, 1900 ; J. Gibson, ib., 23. 349,
1897 ; H. M. Goodurn, ib., 21. 1, 1896 ; A. Weiner and A. Miolate, ib., 21. 225, 1896 ; W. R.
Whitney, ib., 20. 40, 1896 ; A. Hantzsch, Ber., 35. 210, 1902 ; G. Foussereau, Compt. Rend., 104.
116, 1889 ; C. F. Lindsay, Amer. Chem. Journ., 25. 62, 1901 ; H. C. Jones and B. P. CaldweU,
ib., 25. 349, 1901.
8 J. A. Cunningham, Proc. Cambridge Phil. Soc, 11. 431, 1902 ; K. Regner, Phys. Zeit., 4.
862, 1903 ; L. Graetz, Ann. Physik, (4), 1. 530, 1900 ; J. C. Beatty and M. S. de Smolan, Phil.
Mag., (5), 43. 418, 1897.
§ 8. The Migration of Ions
It is impossible to get one kind of ion liberated at one electrode Avithout having a
precisely equivalent c^uantity of an oppositely charged ion appearing at the other electrode ;
984 INORGANIC AND THEORETICAL CHEMISTRY
it is impossible to have a procession of positive atoms through a liquid without a corre-
sponding procession of negative ones. In other words, an electric current in a liquid
necessarily consists of a flow of positive electricity in one direction combined with a flow
of negative electricity in the opposite direction.- — O. J. Lodge (1802).
Many early investigators — e.g. M. Faraday (1834), i J. F. Daniell and W. A.
Miller (i844), etc. — noticed that changes in concentration are produced about the
electrode during the electrolysis of a solution, and that although the quantities of
anions and cations liberated at the electrodes during electrolysis are always strictly
equivalent, nevertheless, the rates at which the concentrations of the electrolyte
changes about anode and cathode are not the same. W. Hittorf studied the effect
of strength of current, concentration of solution, and temperature on the phenomenon
which was attributed to differences in th^ speeds at which the anions and cations
drifted in the solution.
The changes of concentration about the electrodes are illustrated by an experiment
due to A. A. Noyes and A. A. Blanchard. A U-tube contains a solution of gelatine
colored with cupric chloride and covered with a layer of sodium chloride. The
object of the gelatine is to prevent any movement of the liquid. When electrolyzcd,
the blue colour rises into the sodium chloride at the cathode side, and descends
below the level of the gelatine on the anode side. With potassium dichromate
instead of cupric chloride, the yellow colour rises at the anode side and descends at
the cathode side. With copper dichromate, the yellow colour rises in the anode
compartment and the blue in the cathode. The ionic . theory interprets these
experiments by assuming that in the case of cupric chloride, blue copper ions travel
towards the cathode and colourless chlorine ions towards the anode ; in the second
oooooooo
Fig. 8. Fig. 9.
J oooooooo| I coooo I ooooooooo|oo
experiment, that yellowish Cr207"-ions travel towards the anode and colourless
potassium ions towards the cathode ; and in the third experiment, blue copper
ions travel towards the cathode and yellow dichromate ions towards the anode.
The fact observed is that the electrolysis of the coloured solutions occurs at the boundary
surfaces betioeen the gelatine and the supernatant solution. M. Faraday (1833) has
described experiments illustrating the phenomenon : " the surfaces of separation
of liquids in contact act as electrodes to each other, and separation may there
occur just as at a plate."
By a modification of these experiments, it is possible to measure the rates at
which the concentration of the solution changes about the electrodes, or, in the
language of the ionic theory, the rates at which the anions of copper, etc., drift
towards the electrodes. Let the cations be represented by • and the anions by o,
and suppose the two sets of anions to be arranged in rows as shown diagrammatically
in Fig. 8, with an equal number on each side of a porous diaphragm in the
electrolytic cell. There are eight molecules in each compartment, and the con-
centration of ion about each electrode can be represented by eight. If the cations
move twice as fast as the anions, then, after a certain interval, the conditions may
likewise be represented by Fig. 9, where each ion with no partner is supposed
to have been discharged at the electrodes. Six cations and six anions have there-
fore been set free. The concentration in the cathode compartment has decreased
from 8 to 6, and in the anode compartment from 8 to 4, with a loss of 2 and 4
respectively. Hence the loss in concentration about the cathode is to that about
the anode, as the velocity of the anion is to that of the cation ; that is, the losses
in concentration about the electrodes are inversely as the speed of the correspond-
ingly named ions. The relative velocities of the ions through the solution under
ELECTROLYSIS AND THE IONIC HYPOTHESIS
985
a potential-gradient of 1 volt per cm. are best termed Hittorf's transport numbers
of the ions.
If a solution of silver nitrate of known concentration be electrolyzed between
silver electrodes in an apparatus similar to that illustrated in Fig. 7, the only change
in the solution is a transfer of silver from the anode to the cathode, and a change in
the concentration of the silver salt round the two electrodes ; for the apparatus is
constructed so as to reduce the mechanical convection of the dissolved salt to a
minimum. The change in the concentration of the solution, after a few hours'
electrolysis, can be measured by withdrawing about half the solution from the
apparatus, via the stopcock, and determining the amount of silver in the solution
by analysis. From the results, numbers can be obtained which are supposed to
represent the speeds of migration of the anions and the cations. The following
numbers, due to W. Hittorf (1853-59), serve to illustrate the principle.
W. Hittorf's experiment. — A solution of silver nitrate containing one part of
silver to 49 '44 parts of water was electrolyzed for nearly an hour in a cell with silver
electrodes. Silver dissolved from the anode and a similar quantity deposited on
the cathode. The concentration of the whole solution remained unchanged, but the
concentration of the solution about the cathode decreased while that about the anode
increased. In the cathode compartment, W. Hittorf found
Silver before electrolysis
Silver after electrolysis
Loss
0-7162 gram
0-5862 „
0-1300 ,.
The solution about the cathode thus lost 0'1300 gram of silver, and the solution
about the anode must have increased by this amount owing to the action of an
equivalent quantity of nitric acid on the silver electrode.
At the same time, by the simultaneous interposition of a silver voltameter in
the circuit, it was found that sufficient electricity had passed through the electrolyte
to deposit 0'2470 gram of silver at the cathode. If no silver ions
have passed from the anode chamber, the quantity of silver in the
anode chamber would have increased by 0*2470 gram owing to
the transport of NOg'-ions from the cathode chamber. The
observed increase was only 01300 gram of silver ; hence 0"2470
less 0'1300 gram ; in all, 0-1170 gram of Ag'-ions were transported
from the anode cKamber to the cathode chamber while the cathode
chamber simultaneously lost 0*1300 gram of silver due to the
deposition of 0*2470 gram of silver on the cathode. Hence since
the relative speeds of the ions are proportional to the fall of the con-
centration about the oppositely named electrodes, or
Loss in cathode chamber 0-1300
Loss in anode chamber 0-1170
Speed of anion, NO 3'
Speed of cation, Ag-
or the rate of transport of the anions is to the rate of transport
of the cations as 130 : 117 ; or the N03'-ions migrate 1*1 times
as fast as the Ag*-ions.
The mobility of the elementary ions was found by G. Bredig to
vary periodically with the atomic weight ; and with complex ions,
W. Ostwald and G. Bredig 2 noted that the mobility decreases as the combining
weight increases, and with isomeric ions the mobility is approximately the same.
Fig. 10. —Deter-
mination of the
Speed of Ionic
Migration.
References.
1 J. F. Daniell and W. A. Miller, Phil. Trans., 134. 4, 1844 ; J. F. Daniell, ih., 129. 103,
1839 ; M. Faraday, ib., 123. 23, 1833 ; 124. 77, 1834 ; G. Wiedemann, Fogg. Arm., 99. 177, 228,
1856 ; H. Hankel, ib., 69. 263, 184() ; O. (Irotrian, ih., 157. 130, 237^ 1876 ; 160. 238, 1877 ; Wied.
986 INORGANIC AND THEORETICAL CHEMISTRY
Ann., 8. 520, 1879 ; E. Wiedemann, i6., 20. 537, 1883 ; G. Stephan, ib., 17. 673, 1882 ; A. A.
Noyes and A. A. Blanchard, Journ. Amer. CMm. Soc, 22. 726, 1900 ; R. Lenz, Mhn. Acad. St.
Petersburg, (7), 26. 51, 1878 ; F. Kohlrausch and L. Holborn, Leitvermdgen der Electrolyte, Leipzig,
1898 ; H. Scudder, The Electrical Conductivity and Ionization Constants of Organic Compounds,
New York, 1914 ; H. C. Jones, The Electrical Conductivity, Dissociation and Temperature Co-
efficients of a Number of Salts and Organic Acids, Washington, 1912 ; 0. P. Tower, The Conductivity
of Liquids, Easton, Pa., 1905 ; W. Hittorf, Pogg. Ann., 89. 177, 1853 ; 98. 1, 1856 ; 103. 1, 466,
1858 ; 106. 337, 513, 1859 ; Wied. Ann., 4. 405, 1878 ; Ostwald's Klassiker, 21, 1903 ; 23, 1904.
« G. Bredig, Zeit. phys. Chem., 13. 191, 1894 ; W. Ostwald, ib., 2. 840, 1888.
§ 9. The Speeds of Moving Ions — Kohlrausch's Laws
If the anions in a solution remained stationary, the whole current would be
carried by the cations ; but if both ions be moved, the current will be shared between
them, and the share of each will be proj)ortional to the speed at which it moves. i
Let v represent the speed of migration of the cations, and v' that of the anions.
The total current carried from one electrode to the other will be proportional to the
rate at which the ions separate — that is, to the joint velocity of the two ions v'-\-v\
Accordingly the current carried across the electrolyte is shared in such a way that
the
v' iy
Anions share = — - — ; Cations share =
Hence in W. Hittorf's experiment, the relative speeds of the Ag*- and NO3'-
ions were as 0117 : 0*130, =v : v' , the transport numbers are respectivelv
0-130/(0-130-fO-117)=0-52 and 0-117/(0'130+0-117)=0-48. Hence, if n denotes
the transport number of one ion, \—n will represent the transport number
of the other. It might be observed that the change in the concentration of the
solution about one anode is sufficient to compute the relative speeds of migration
of the anions and cations under the given conditions. In W. Hittorf's experiment
just cited, the solution before the passage of the current contained the equivalent
of 0*7162 grm. of silver, and the anode compartment lost 0*1170 grm. of silver.
The total quantity of electrical energy transported through the cell was equivalent
to the 0*2470 grm. of silver deposited in the voltameter. The fraction transported
by the silver ions is therefore 0*1170/0*2470=0*473. If n' denotes the transport
number of the anion, and n' that of the cation, n -|-w'=l- Hence, the transport
number of the NOs'-ions is obtained by subtracting 0*473 from unity ; the result
0*527 agrees with that obtained from the ratio 0*1300/0*2470=:0*527. Obviously
the ratio 0*473 : 0*527 is the same as 117 : 130.
Examples.— (1) F. Vogel (1903)2 found that 100 c.c. of a solution of barium nitrate
contained the equivalent of 0'4419 grm. of barium, and it weighed 100-6660 grms. This
solution was electrolyzed in a suitable apparatus while a silver voltameter deposited
0'5618 grm. of silver, which is equivalent to 0*3577 grm. of barium, 187-2880 grms. of the
anode solution contained the equivalent of 0-6728 grm. of barium ; the concentration of the
solution between the two electrodes did not change. Compute the transport numbers of
anion and cation. Since 187-2880 grms. of the electrolyte before electrolysis contained
the equivalent of 0-8223 grm. barium, and after electrolysis 0-6728 grm. Hence, the loss
in the anode chamber was 0*1495 grm. of barium. The transport number of the cation,
Ba-, is therefore 0-1495/0-3577=0-418 ; and of the anion, NO/, 1 -0-418 =0-.582.
(2) F. Warschauer (1903) found that a solution of sodium metaphosphate NaPOg
contained the equivalent of 0-1775 grm. of P2O5 per 25 c.c. or 25-1998 grms. Before the
electrolysis the anode solution contained 91-6632 grms. of liquid ; hence, it contained the
equivalent of 0-6384 grm. PgOg ; 0-2792 grm. NagO ; and 90-7456 grms. of water. After
electrolysis 0-7370 grm. PgOg ; 0-3223 grm. NaaO ; and 90-6039 grms. of water. Hence,
90-6039 grms. during the electrolysis gains 0-7370 — 0-6374=0-996 grm. P2O5, and this is
equivalent to 0-1 108 grm. PO3. While the electrolysis was in progress, a solution containing
6-0035 grms. of silver nitrate per half litre deposited 0-2625 grm. of silver during the
electrolysis. This amount of silver is equivalent to 0-1922 grm. of PO3. Hence the trans-
port number of thePOg'-ion is 0-l]08/0-1922=-0-58, and that of the Na--ion is 1—0-58 = 0-42.
ELECTROLYSIS AND THE IONIC HYPOTHESIS 987
The behaviour of electrolytes with a bivalent radicle united with two univalent
radicles is a little more complex. For example, solutions of sulphuric acid may
furnish two H'-ions, and one S04"-ion, or one H'-ion and one HSO^-ion. This is given
as an explanation of the fact why the transport numbers for such solutions show
appreciable differences when determined in dilute and in concentrated solutions.
Thus, the anion of barium chloride in aqueous solution varies from 0*611 to 0"555
by progressive dilution as shown in the following scheme :
Normality of solution
. 0-50
0-20
0-10
0-05
002
0-01
Transport numbers
. 0-61
0-59
0-57
0-56
0-56
0-56
KCl NaCl
KNO3 NaNOg
KF NaF
129-1-108-1
125-5- 104-5
110-5-89-4
210
21-0
21-1
The concentration of the solution determines whether the ionization proceeds
HaSO^^H'+HSO/ or H2S04^2H-l-SO/'. With decreasing concentration the
ions become less complex, the solution yields transport numbers approaching a
constant value. The variation is due to changes in the relative proportions of the
anions and cations with increasing dilution : BaCl2^BaCl"+Cr^2Cr+Ba'\
All the molecules of the solute at infinite dilution are ex hypothesi ionized and take
an active part in conducting the current, while the number of molecules which take
an active part in conducting the current at any particular dilution is proportional
to the molecular conductivity, fjL, and therefore, if x represents the fraction of a gram-
molecule which is ionized when the solution is diluted to v litres, fi=kx, where k
is the constant of proportion. At infinite dilution, the whole molecule is ionized,
and consequently 07=1 ; yL^=h ; smd fi—fji^x. On comparing the molecular con-
ductivities of a number of different salts at infinite dilution, F. Kohlrausch (1876)
noticed a curious fact : The difference in the molecular conductivities of potassium
and sodium chlorides is equal to the difference between potassium and sodium
nitrates ; and this in turn to the difference between sodium and potassium fluorides ;
etc. In illustration :
/^oo • . •
Difference .
The difference in each pair of salts with a common anion thus depends on the
difference in the speeds of migration of the cations — K" and Na'^ — and this is
constant. Similar relations hold good for other salts containing a common cation.
It is thence inferred that the molecular conductivity at infinite dilution is the sum
of two factors, one dependent upon the nature of the anion, the other on the cation
— this is Kohlrausch's second law. At infinite dilution, the molecular conductivity
is represented by Kohlrausch's equation: ix^=v'-\-v\ Consequently, the speed
of migration of any particular ion in a particular solvent is constant at infinite
dilution, is dependent on its own chemical nature, and is independent of the nature
of the other ion or ions which may be present. Otherwise expressed, the speeds
of migration of the different ions in a solution are independent of one another.
In illustration, the speeds of migration of the chlorine ion in O'OOOliV-solutions
of lithium, sodium, and potassium chlorides were respectively 65*7, 65*7, and 65*6.
W. Hittorf showed how to determine the numerical value of the ratio v/v', and
F. Kohlrausch's equation furnishes the numerical value of v'-\-v\ Hence it is
possible to compute the absolute velocities of the anion and cation. The results
agree closely with those obtained by direct measurement.
ExAMPLE.^ — ^The molecular conductivity of a solution of silver nitrate at infinite dilution
is 115, at 18°. This means that 115 coulombs of electricity are carried 1 cm. per second ;
but each gram ion of silver nitrate carries 96,540 coulombs (Faraday's law), hence 115
coulombs are carried 115-^96540=0*00121 cm. per second. This represents the velocity
at which the two ions draw apart. The carriage of the electric charge is shared by both
ions ; the migration constants of the two ions are Ag- 54, and NO3' 62. The ratio
ar/v' = 54/62 =0-87. Hence v-=0-87v'. From Kohlrausch's equation 0-00121 =v-+v' or
l-87v'=0-00121, or v'=0-00065, and v- -=0-00050 cm. per second.
988 INORGANIC AND THEORETICAL CHEMISTRY
F. Kolilransch's equation was extended by W. Ostwald to represent other than
infinite dihitions by introducing the term />t=a/x^. Hence, when ionization is not
complete, Kohlrausch's equation becomes fjL=a(v'-\-v'). As the solution is diluted
more and more the term a becomes more and more nearly equal to unity, and
finally when ionization is complete, a=l and Kohlrausch's equation ^^=v'-^v'
appears. The expressions assume a similar form if the equivalent conductivities
be in question.
ExAMPLK. — The transport numbers for the ions in a solution of sodium metaphosphate
are 0*573 for PO3', and 0-427 for Na*. The maximum equivalent conductivity of the
solution is 126*2 luiits. What are the relative speeds of migration of the two ions ? If n be
the transport number of the cation, 1 —n will be the transport number of the anion, and
nl{l—n)=v'/v' ; and X=x{v'-{-v') ; by substitution therefore va = (l— n)A, and v'x==nX.
When the equivalent or molecular conductivity is a maximum, ionization is complete, and
a = l. Accordingly, n=0*427 ; 1— n = 0'573; A = 126*2. Hence, the relative speeds of
migration of the cation Na* is 53*9, and of the anion PO'3, 72*3 ; or the sodium ion travels
53*9-T-96540 =0*00056 cm. per second when the electromotive force is one volt, etc.
By measuring the rate of rise of the blue colour in A. A. Noyes and A. A. Blan-
chard's experiment, the velocity of copper ions can be determined under standard
conditions, and in that way, with other solutions, a series of numbers have been
obtained which represent the velocities of migration of the respective ions. In
0. J. Lodge's experiment (1886) a current was passed through a j — \,-shaped tube
containing a stiff gelatine solution of sodium chloride coloured red with phenol-
phthalein and a trace of sodium hydroxide. The ends of the tube dipped in beakers
contained dilute sulphuric acid. The time taken to decolorize a certain measured
distance was determined, and this was corrected by the time taken for the acid to
diffuse in the jelly when no current was passing. In this w^ay it was found that the
hydrogen ions travelled through the jelly at the rate of 0"0026 cm. per second when
a difference of potential of one volt was applied to the electrodes. H. B. Denison
and B. D. Steele, and G. M. Lewis used the method of moving boundaries for
measurements of the transport numbers.
The transport numbers are expressed in the same units as the equivalent and
molecular conductivities, and they are proportional to the velocities of the ions.
The speeds of the ions are decreased by increasing the viscosity of the solution,
say, by adding non-electrolytes like cane sugar, alcohols, ether, etc. Gelatine has
very little influence. The decrease in viscosity which occurs on raising the tempera-
ture is also of influence. The speeds of the ions are also augmented by using
currents of greater electromotive force. Hence, the rate of motion of any given
ion is determined (i) by the intensity of the electric pressure which directs or drives
the ions to the electrodes ; (ii) by the damping effect of the liquid on the moving
ion.
It must be remembered also that the raising of the temperature or the addition
of a third substance to the solution may materially modify or even mask the normal
relations between solvent and solute. The conductivity of a solution may virtually
vanish suddenly at the point of solidification of a cooling liquid ; so that by plotting
the conductivity of a molten salt cooling through a range of temperature, there will
be an abrupt change in the direction of the curve at the freezing teiiiperature, and
it has been proposed to apply this principle to determine transition temperatures —
solid or liquid, etc.
Many solids are poor conductors at ordinary temperatures but good conductors at
temperatures above their melting points. One explanation assumes that the conductivity
of fused salts is due to self-ionization, in other words, that a small portion of fused substance
is ionized. Glass and jDorcelain are poor conductors at ordinary temperatures, but they
conduct very fairly at more elevated temperatures, and this principle has been utilized in
the so-called Nernst lamp.
At 18°, with a difference of potential of one volt between the electrodes^ the
a})8olute velocities of some ions are :
ELECTKOLYfSlS AND THE IONIC HYPOTHESIS 989
Anions
Speeds
OH'
5-6
cr
212
r
2-19
NO3'
1-91
CIO'3
1-70 cm. por hour
Cations
Spc^eds
Cs-
. 2-32
Rb-
2-32
K-
205
Na-
1-26
Li-
1-11 cm. per hour
Under similar conditions, the charges carried by different ions may be equal ;
their speeds are different. The heaviest ions, in the alkali series of elements — that
is, the ions with the greatest " atomic " weights — here appear to move fastest.
This has been supposed to be due to the slower-moving ions dragging along with
them a number of molecules of the solvent. This assumption seems to be justified
by experiments with dilute solutions, but with concentrated solutions there is a
discrepancy between results obtained by W. Hittorf's method and the method of
moving boundaries, which, is explained by assuming that in dilute solutions the
change in the concentrations is negligibly small, whereas in a concentrated solution
the difference in -^he concentration of the ions at each electrode must be affected
by the water carried by the ions. W. Hittorf's method does not distinguish between
changes of concentration about the electrodes due to the transport of ions and
those due to the carriage of water by the ions.^ In Hittorf's method it is assumed
that the solvent is quite stationary during the electrolysis, and that no solvent is
transferred from the one electrode to the other. If the ions are hydrated, not only
the electrolyte but some of the solvent will be transferred from the one electrode to
the other, unless the respective ions happen to be equally hydrated. This argu-
ment has been tested by W. Nernst, G. Buchbock, etc., by using solutions of an
electrolyte containing a third substance — raffinose, mannite, sugar, etc. — not
affected by the electrolysis, as a standard of reference for the concentration of the
water. As a result, it was found that the numbers are not generally in agreement
with Hittorf's transport numbers although they agree with those obtained by the
method of moving boundaries ; it was also found that during the electrolysis of
hydrogen, sodium, potassium, and lithium chlorides, the ratio water : sugar
decreases at the cathode and increases at the anode under conditions where no
change occurs in the absence of the electrolyte. This fact can be explained either
by assuming that the non-electrolytic sugar is carried from cathode to anode ; or
that water is carried from anode to cathode during the passage of the current. . The
former hypothesis assumes that the ions (or electrolyte) form a complex with the
sugar, and for this there is no satisfactory evidence ; the ionic hypothesis assumes
that the ions are hydrated, and carry the water molecules as well as their electric
charges. The molecular equivalents of water transferred from anode to cathode
per farad of electricity when the electrolyte has normal concentration :
HCI KCl NaCI LiCI .
0-24 0-60 0-76 1-5
These numbers taken in conjunction with the transport numbers show that the
degree of hydration varies with the different ions and with the concentration of the
solution. The relative degrees of hydration for the different ions (chlorine assumed
to be unity) are :
Cl' ir K; Na- Li-
1 0-46 2-3 3-6 7-0
the row of numbers must represent minimum values since a negative value for
chlorine is impossible. Hence all the cations in the above-mentioned solutions
must be hydrated, and the hydration increases markedly in passing along the
series H, K, Na, Li. If the speeds of migration of the different ions be plotted on
squared paper against the atomic weights of the elements, a periodic curve is
obtained. H. Kemy has calculated hydration numbers from the changes in volume
which occurred in the anodic and cathodic solutions during electrolysis.
Evidence indicating the union of ions with the solvent has been sought in the
990 INORGANIC AND THEORETICAL CHEMISTRY
changes which occur with boiling points a,nd freezing points of concentrated solu-
tions. These phenomena depend on the ratio of the number of molecules (or ions)
of solute and of solvent. If union occurs, the number of molecules (or ions) is
not changed, but only their size. The combination in dilute solutions can
remove but a relatively small amount of solvent from the field, and the consequent
effect is inappreciable. Even in concentrated solutions the deviations from the
theoretical values do not show whether it is the free ions that are so combined, or
whether it is not the result of something else. Measurements of the boiling and
freezing points of solutions of hydrated salts in other solvents have also been made
with the idea of finding if the water of crystallization remains attached to the salt.
In virtually all cases the evidence points one way : the ions do unite with molecules
of the solvent, and thus move more slowly than if they were not harnessed with the
molecules of the solvent.
The data show that the rates at which ions move through water are surprisingly
small — one centimetre per hour for theK'-ion under a pressure of one volt. And it
requires a pressure equivalent to about 300,000 tons to drive a gram ion of hydrogen,
H', through water at this speed.
Example.- — ^By definition W dynes of energy are required to drive a coulomb of
electricity one centimetre against a potential difference of one volt. A force of 10' dynes
is equivalent to 10' 18 kilograms. Hence, the force required for a gram-ion charged with
96,540 coulombs will be 96,540 X 10*18 = 983,000 kilograms ; and measurements of the absolute
velocity of the hydrogen ion show that this force is required to drive these gram -ions with
a velocity of 0*00325 cm. per second. Hence to drive the gram-ion with a velocity of one
centimetre per second will require a force of nearly 983,000 -s-0'00325 = 302,000,000 kilograms.
One ton is nearly equivalent to 1016*05 kilograms, and therefore 302,000,000 kilograms is
nearly equivalent to 297,000 tons.
W. Hittorf 's results can be deduced from the assumption that both the solvent and
the solute conduct the current, for if, say, silver nitrate conducts the whole current,
free acid appears only at the anode ; if the water conducts the whole current, free
acid appears only at the cathode ; and if both salt and solute conduct the current,
free acid appears at both electrodes, and the current can be portioned between the
solvent and solute so as to make the observed facts fit the hypothesis. Hence,
W. Hittorf's migration data do not prove that ions travel at unequal rates, for the
observed facts can be explained by at least two plausible hypotheses.
In these experiments the fact observed is the changing molecular concentration
of the solution about the anode and cathode during electrolysis ; the hypothesis
is that during the passage of the current the anions and cations move in the same
electrolyte with different velocities, and are yet the anions and cations given off at
the respective electrodes at the same time.
References.
1 F. Kohlrausch, Wied, Ann., 6. 1, 1879 ; 26. 213, 1885 ; 0. J. Lodge, B. A. Rep., 723,
1885 ; 389, 1886 ; B. D. Steele, Journ. Chem. Soc, 79. 414, 1901 ; B. 1). Steele and H. B. Denison,
ib.y 81. 456, 1902 ; W. Hittorf, Pogg. Ann., 89. 177, 1853 ; Zeit. phys. Chem., 39. 613, 1902.
2 F. Vogel, Zeit. anorg, Chem., 35. 385, 1903 ; F. Warschauer, ih., 36. 137, 1903.
3 E. W. Washburn, Journ. Amer. Chem. Soc, 31. 322, 1909 ; G. N. Lewis, ib., 32. 862, 1910 ;
J. L. R. Morgan and C. W. Kanolt, ib., 28. 572, 1906 ; H. C. Jones and H. P. Basse tt, Atner.
Chem. Journ., 32. 409, 1906 ; G. Buchbock, Zeit. phys. Chem., 55. 563, 1906 ; H. B. Denison and
B. D. Steele, ib., 57. 110, 1906 ; H. Remy, ib., 89. 529, 1915 ; W. Nernst, Gott. Nachr., 86, 1900.
§ 10. " Abnormal " Osmotic Pressures and Ionization
It is natural to assume that substances are ionized which give in aqueous solution osmotic
pressures which are too great.' — ^S. Arrhenius.
We are now in a position to resume our study of the abnormal osmotic pressures
furnished by solutions of electrolytes. Just as the abnormally high vapour density
ELECTROLYSIS AND THE IONIC HYPOTHESIS
991
exhibited by hydrogen fluoride was traced to polymerization : 2HF^H2F2, and the
abnormally low vapour density of iodine above 700° was traced to the dissociation :
I2=2I, so S. Arrhenius argued that salts in solution which give an abnormally
high osmotic pressure are similarly dissociated. Suppose that one molecule of an
electrolyte furnished m ions, and further let a denote the fraction ionized when a
gram-molecule of the electrolyte is dissolved in water. The solution will then con-
tain (1— a) non-ionized molecules, and ma ions. The total number of individual
molecules in the solution — that is, electrically charged molecules (ions) and neutral
molecules — will be (1— a)-|-ma. As in our previous study of solutions, let n denote
the total number of individual molecules formed by the ionization of a substance in
a given solution. Then n^=l-{-ma—a. The numerical value of n, as we have seen,
can be determined from conductivity data, and from osmotic pressure and related
phenomena — freezing and boiling point determinations. If the value of n so deter-
mined be divided into the value of n calculated on the assumption that no ionization
occurs, the vahie of a can be computed ; and if a be known, the value of n can be
computed.
Examples.- — (1) The solution of hydrochloric acid just studied gives a=0"874 and m = 2.
Hence w = l +(?/» — l)a becomes «.== 1*874. Hence every 100 molecules of HCl furnish the
equivalent of 187 "4 individual molecules. If the electrolyte has been non-ionized, n would
have been unity ; and if completely ionized, n would have been 2.
(2) The boihng point of a solution of 3*400 grms. of barium chloride, BaClg, in 100
grms. of water is 100°-208°, what is the degree of ionization of the solute ? From the above
expression, there will be 1+ma — a = l + 3a — a = l+2a "molecules" in the solution, viz.
Ba-, 2Cr, and hence one molecule of barium chloride furnishes the equivalent of 208/85
= 2-447 "molecules." Consequently, 2*447 = 1 -f 2a, or a=0-723; or 72*3 per cent, of the
salt is ionized.
(3) A solution of 11*07 grms. of barium nitrate in 100 grms. of water raised the boiling
point 0*466°. What proportion of the salt is ionized ? Ansr. 55*8 per cent.
(4) A solution of 0*3668 grm. of sodium chloride in 100 grms. of water freezes at —0*221°.
What proportion of the salt is ionized ? Ansr. 89*2 per cent.
A comparison of the values of n calculated from osmotic pressure, freezing point,
and electrical conductivity data is indicated in Table II. The numbers in the last
three columns show that the values determined by independent processes are
Table II.— Molecular Weights of Some Electrolytes in Solution.
Molecular
concentration.
Values of n.
Salts.
Osmotic
pressure.
Freezing
point.
Conductivity.
Calcium nitrate, Ca(N03)2 .
Magnesium sulphate, MgS04
Strontium chloride, SrCU .
Potassium chloride, KCl"
Lithium chloride, LiCl
Magnesium chloride, MgCL .
0*18
0*38
0*18
0*14
0*13
0*19
2*48
1*25
2*69
1*81
1*92
2*79
2*47°
1*20°
2*52
1*86°
1*94°
2*68°
2*46
1-35
2*51
1*86
1*84
2*48
strikingly concordant ; and it is therefore inferred that the abnormal osmotic
pressures indicated in Table VI, Cap. X, arise from the more or less complete
ionization of the electrolytes in aqueous solution.
Modes of ionization.- — ^The ionization of some of the bivalent electrolytes- — ^HgCOa ;
H2SO4 ; BaClo ; CdCla ; K0SO4 ; CUSO4 ; etc.- — ^in moderately dilute solutions appears to
furnish complex ions. " Thus, cadmium chloride, CdCl.,, not only furnishes Cd** + 2Cr, but
also Cd** + CdCl4" ; sulphuric acid, H2SO4, not only gives 2H* + S04", but also H*+HS04' ;
carbonic acid, HXO,, gives 2H- + C03" ; and H*4-HC03'; copper sulphate, CUSO4, not only
gives Cu-' + SOi", but also Cu.SO^** and Cu(S04)/'; etc. If, however, the solutions be still
992 INORGANIC AND THEORETICAL CHEMISTRY
furtlier diluted, the complex ions break down into simpler ones. Hence the ionization of
concentrated polybasic acids like H..SO4 proceeds in stages: lirst HaSOi^^H'+HSO/ ;
foUowed by H + HS04'-2H+S04". "
§ 11. Equilibrium between Ionized and Non-ionized Solute
The evidence is so un-ambiguous and convincing that ions and some molecules combine
with niore or less of the solvent that it seems that it can now be accepted as a fact of science. —
H.C.Jones (1913).
Reference has previously been made to the assumption that the molecules of an
electrolyte, when dissolved in water, are ionized ; that the ions, at the same time,
recombine to form neutral molecules ; and that equilibrium will ensue when the
speeds of the two opposing reactions — ionization and de-ionization— are e([ual.
Consider the ionization of ammonium hydroxide, NH4OH, represented by NH4OH
v=^NH4"-f-0H'. Here the process of ionization bears some analogy with the dissocia-
tion of iodine by heat : I2=I+I. Let [NH4OH] denote the concentration of the
ammonium hydroxide ; [NH"4], that of the concentration of the ammonium ion ;
and [OH'l, that of the hydroxide ion. Then, applying the principle of opposing
reactions, the condition for equilibrium is: (NH4"]x[OH']=iC[NIl40HJ. If this
theory applies to ions, it follows that the numerical value of the equilibrium constant,
K, now called the ionization constant, remains unchanged whatever be the concen-
tration of the solution. This relation is sometimes called W. Ostwald's dilution
law,i or Ostwald's mass law, and, for univalent electrolytes, it is also symbolized :
:, — - =A ; or, " =A ; or . \ =^
1— a Cu (\—a)v
where a denotes the degree of ionization ; (7, the total concentration ; Cj, the con-
centration of each of the ions ; C„, the concentration of the non-ionized part ; and
V the number of litres of solvent containing a gram-molecule of the salt.
Example.— In a solution containing 0-125 gram-molecules of NH4OH per litre, the
equivalent conductivity shows that 0'0135 gram -molecules are ionized, and hence, 0"0135
X0'125=0*0017 represents the molecular concentration of the ammonium hydroxide which
is ionized. This number thus represents the concentration of the NH4'-ions. But every
NH4"-ion is accompanied by one OH'-ion, and accordingly, O'OOIT also represents the concen-
tration of both the NH4*- and the OH'-ions. Hence, from Ostwald's dilution law, 0'0017
xO-0017-T-(0-125 -0-0017) =0-000023. This last number represents the value of the ioniza-
tion constant for a seminormal solution of ammonium hydroxide.
It follows from this rule that the greater the dilution v the greater will be the
percentage amount of ionization (although, of course, the actual concentration
of the ions must decrease). Since a^X^jX^ the value of K can be readily computed ;
and this has been done by W. Ostwald for a large number of organic acids, and by
G. Bredig for organic bases, with results so very satisfactory that S. Arrhenius could
say "in no other field has the law of mass action been applied with such good
results " — the converse of this statement aj)plies with strong electrolytes. Table
III represents values for the ionization constant for solutions of ammonium
hydroxide of different strength.
The constancy of the value K means that although the last-named solution of
ammonium hydroxide is nearly 300 times more dilute than that named first, and the
degree of ionization of the last is nearly 16 times as great as the first, the expression
represented by K, deduced on the supposition than the process of ionization follows
the rule for opposing reactions, is constant within the limits of experimental error.
It will be observed that two important assumptions have been tacitly made in
deriving the formula : (1) that the concentration of the ions can be obtained from the
conductivity ratios ; and (2) that the ions and un-ionized molecules of an electrolyte
I
ELECTROLYSIS AND THE IONIC HYPOTHESIS
993
follow the laws of ideal solutions. In spite of the very successful results obtained
with the feebler acids and bases, for some unknown reason, the application of the
Table III. — ^Effect of Dilution on the Ionization of Aqueous Ammonia.
Ammoiiiiira
hydroxide.
(G ram-molecules
per litre.)
Proportion
ionized.
Molecular con-
centration of NH4-
and of OH' ion.s.
(Gram '* ions "
per litre.)
Molecular con-
centration of non-
ionized NH4OH.
(Gram-molecules
per litre.)
K
10000
0-1250
00159
0-0039
0-0047
0-0135
0-0376
0-0754
0-0047
0-0017
0-0006
0-0003
1-0000-0 0047
0-1250-0-0017
0-0159-0-0006
0 0039-0-0003
0-000023
0-000023
0-000023
0-000023
dilution law to the stronger electrolytes has been a signal failure, and the so-called
anomaly is regarded as the bete noir of S. Arrhenius' ionic hypothesis. The numerical
value of K, instead of remaining constant, increases rapidly with increasing concen-
tration. In the first place, the equivalent conductivity of a univalent substance is
X=aF{v-{-v'), where a is the fraction of the substance ionized, and F the quantity
of electricity associated with each equivalent ion ; and the equivalent conductivity
at zero concentration or infinite dilution is A^=JP(Vqo-|-?/oo). Hence,
A v'-{-v'
which shows that A/A^ is equal to a only when the mobilities of the ions are
constant for the concentrations under consideration, for only then does v-{-v^
==v'^±v'^. In view of the possible electrical effects resulting from the large
electric charges on the ions, this assumption is by no means certain.
Again, S. Arrhenius and W. Ostwald 2 inclined to the view that the cause of the
discrepancy is due to the failure of the mass law with strong electrolytes ; H. Jahn
believed that imperfections in the method of measuring Kl^ao render the ratio an
imperfect measure of a ; and C. A. Kraus suggested introducing a correction term
for the viscosity of the medium, Kvol^a:>V—^> when t^q and 7] respectively denote
the viscosities of solvent and solution. E. W. Washburn proposed A</>5g/A^^»^=a,
where (/> represent the fluidities of the media ; for solutions less than normal,
m=0'9i, but F. G. Donnan and W. E. Garner showed that the value of m decreases
with increasing concentration for 2N- to 12iV-solutions of lithium chloride.
Another set of workers — S. Arrhenius, C. A. Kraus and W. C. Bray, J. Kendall, and
W. R. Bousfield, etc.3 — have corrected the results for the effect of traces of impurity
in the solvent on the conductivity measurements. A. A. Noyes and W. C. Bray,
J. Walker, etc.,* have suggested that the deviation is to be ascribed to the abnormal
behaviour of the un-ionized molecules of strong electrolytes slightly increasing the
concentration of the ions. The stimulating effect of the un-ionized molecules on
ionization is also assumed by B. de Szyszkowsky. P. Walden 5 suggested that the
solute increases the dielectric constant of the solvent and this augments its ionizing
power. H. Jahn and G. N. Lewis have suggested that there is an increase in the
ionic mobility with increasing ionic concentration. J. C. Ghosh assumes with
W. Sutherland that the electrolyte is completely ionized, but that part of the ions are
free to move as contemplated by the kinetic theory, and part are bound so as to
hold fixed positions in the solvent-like structural units in the space lattice of a crystal.
Another hypothesis assumes that a Grotthus chain conductivity is superposed on the
ionic conductivity, a view negatived by the work of C. A. Kraus and W. C. Bray ; ^
yet another hypothesis suggests that the solvent in the solution also conducts some
of the current.
VOL. I. 3 s
994 INORGANIC AND THEORETICAL CHEMISTRY
The question was raised by 0. J. Lodge in 1886, if pure water and pure hydrogen
chloride are non-conductors, and the conductivity of hydrochloric acid is due to the
ionization of the hydrogen chloride by water ; presumably also ' the effect is not
one-sided, and the water is likewise ionized by the acid.^ J. W. McBain has pointed
•out that if the movement of the cation and anion add up to unity, the solvent cannot
have taken a share in the conductivity, and this conclusion is not affected by the
presence of various kinds of complex ions. The work of G. Poma and B. Tanzi ^
makes it appear as if the water in the presence of sodium chloride is not so much
ionized as the solvent alone. W. Palmaer and K. Melander draw the opposite
conclusion with respect to aqueous solutions of calcium and lithium chlorides, so
also did H. S. Harned with respect to solutions of sodium bromide or the chlorides of
the alkalies and alkaline earths in hydrochloric acid. A. Sachanoff found that
with solvents of low dielectric constant, an increased ionization occurs when other
electrolytes are added, although with water there is a decrease.
The failure has been attributed to the gradual hydration of the ions during
dilution. If so, the mobility of the ions will be constant only when the hydration
is complete. The conductivity, therefore, will not depend upon ionization alone,
but will also depend on the degree of hydration of the ions. If the dissolved sub-
stance combines with the solvent, then, in the more concentrated solutions, part of
the liquid in which the substance is dissolved will no longer function as solvent
because it is in combination with the ions and the non-ionized molecules of the
solute. This is virtually the solvate theory of solution which so much attracted
H. C. Jones.9 However, this does not affect the principles which are based upon
the dilution law, for it is merely necessary to introduce an additional factor to
provide for the fact that the solution is much more concentrated than it would be if
all the liquid in which the substance is dissolved is doing the work of a solvent.
H. C. Jones, for instance (1913), has given very good evidence for assuming that
what we call a normal solution of aluminium chloride and of several other substances
in water is about twice normal because about five-eighths of the water is actually
combined with the dissolved substance.
Numbers of formulae ^^ have also been proposed by M. Rudolphi, L. Storch,
J. H. van't Hoff, F. H. McDougall, C. A. Kraus, and W. C. Bray, etc., in place of the
one based on the law of mass action. In most of these, empirical changes are made in
the indices or terms are added to the simple dilution law. The results are satis-
factory only when empirical correspondence with fact is desired. The present
position of the theory of ionization discussed at the Faraday Society n (19i9), made
it very clear that in spite of twenty years' labour there is something fundamentally
wrong in the application of the mass law to ionization, excepting in the case of weak
electrolytes and of very dilute solutions of the stronger electrolytes. The general
conclusien to be drawn from the work is that electrical conductivity is not an accurate
measure of ionic concentration, and for strong electrolytes in concentrated solutions,
A„/A is not equal to a. This is confirmed by a comparison, by H. N. Lewis and G. A.
Linhart, of A. A. Noyes and K. G. Falk's value of a from the ratio A/A^ and the value
computed from thermochemical data for solutions with O'Ol gram-molecule of the
salt per litre. The discrepancies are very pronounced :
KCl NaCl KIO3 NalOa KoSO* BaCIg CdSOi CuSO* La(N0g)2
a . . 0-925 0-925 0-872 0-872 0-687 0-716 0-338 0290 0-571
A/Aoo . 0-941 0-936 0-928 0-917 0-872 0883 0614 0629 0-802
Beferenoes.
1 W. Ostwald, Zeit. phi/s. Chem., 2. 36, 1888 ; 3. 170, 241, 369, 1889 ; G. Bredig, ib., 13. 289,
1894 ; S. Arrhenius, Textbook of Electrochemistry, London, 163, 1902.
2 S. Arrhenius, Zeit. phys. Chem., 36. 28, 190J ; 37. 490, 1901 ; W. Nemst, ib., 36. 696, 1901 ;
H. Jahn, ib., 27. 364, 1898; R. A. Lehfeldt, ElectrochemisPry, London, 1904; W. Ostwald,
QrUndriss der allgemeinen Chemie, Leipzig, 406, 1899.
» S. Arrhenius, Medd. Akad. Nobel-fnst., 2. 42, 1913 ; C. A. Kraus and W. C. Bray, Journ.
ELECTROLYSIS AND THE IONIC HYPOTHESIS 995
Amer. Chem. Soc, 35. 1413, 1915 ; J. KendaU, ib., 38. 1480, 2460, 1916 ; 39. 9, 1917 ; E. W.
Washburn, ib., 33. 1461, 1911 ; 40. 106, 1918 ; H. J. Weiland, ib., 40. 13J, 1918 ; W. R. Bousfield,
Jaurn. Chem. Soc, 103. 310, 1913 ; R. Bourdillon, ib., 103. 191, 1913 ; J. W. McBain and P. C.
Coleman, ib., 105. 1517, 1914 ; R. Wegscheider, Zeit. phys. Chem., 69. 621, 1909 ; G. N. Lewis,
ib., 61. 129, 1907 ; Froc. Amer. Acad., 43. 269, 1907 ; F. G. Donnan and W. E. Gamer, Journ.
Chem. Soc., 115. 1313, 1919.
♦ A. A. Noyes and W. C. Bray, Journ. Amer. Chem. Soc, 33. 1643, 1911 ; A. A. Noyes, W. C.
Bray, and F. S. Farrell, ib., 33. 1630, 1911 ; A. A. Noyes and I). A. Maclnnes, ib., 42. 239,
1920 ; W. C. Bray and W. J. Winninghoff, ib., 33. 1663, 191] ; W. C. Bray, ib., 33. 1643, 1911 ;
J. Walker, B. A. Hep., 81. 349, 1911 ; B. de Szyszkowaky, Medd. Vet. Nobel-Inst., 3. 2, 3, 4, 5, 1914 ;
Compt. Bend., 157. 767, 1913.
5 P. Walden, Journ. Amer. Chem. Soc, 35. 1649, 1913 ; G. N. Lewis, ib., 34. 1631, 1912 ;
A. Sachanoff, Zeit. phys. Chem., 87. 441, 1914 ; H. Jahn, ib., 33. 545, 1900 ; J. C. Ghosh, Journ.
Chem. Soc, 113. 449, 627, 707, 790, 1918 ; W. Sutherland, Phil. Mag., (6), 3. 161, 1902 ; (6),
11. 781, 1905 ; (6), 12. 1, 1906 ; (6), 14. 1, 1908 ; (6), 16. 497, 1908 ; S. R. Mihier, ib., (6), 35. 214,
1918.
« C. A. Kraus and W. C. Bray, Journ. Amer. Chem. Soc, 35. 1369, 1913.
' 0. J. Lodge, B. A. Rep., 391, 1886 ; A. A. Noyes, Zeit. phys. Chem., 9. 614, 1892 ; M. le
Blanc, ib., 8. 413, 1891 ; P. Walden, Journ. Amer. Chem. Soc, 35. 1659, 1913 ; J. W. McBain,
Zeit. Elektrochem., 11. 216, 1905.
8 H. S. Harned, Journ. Amer. Chem. Soc, 37. 2460, 1915 ; W. Palmaer and K. Melander,
Zeit. Electrochem., 21. 418, 1915 ; A. Sachanoff, Zeit. phys. Chem., 87. 441, 1914 ; G. Poma and
B. Tanzi, ib., 79. 55, 1912.
9 J. D. van der Waals, Zeit. phys. Chem., 8. 215, 1891 ; V. Gordon, ib., 18. 8, 1895 ; W. Roth,
ib., 24. 114, 1897 ; L. Braun, ib., 33. 721, 1900 ; W. Knopp, ib., 48. 97, 1904 ; G. Hiifner, ib., 57.
611, 1907 ; G. Senter, Trans. Faraday Soc, 3. 24, 1907 ; Discussion : ib., 3. 1, 1903 ; J. C. Philip,
Journ. Chem. Soc, 91. 711, 1907 ; E. Baur, Von den Hydraten in wdsseriger Losung, Stuttgart,
1903 ; H. C. Jones, Journ. Franklin Inst., 173. 217, 1912 ; 176. 479, 677, 1913 ; The Nature of
Solution, London, 306, 1917 ; J. S. Guy, E. J. Schaeffer, H. C. Jones, Amer. Chem. Journ., 49,
265, 1913.
10 M. Rudolphi, Zeit. phys. Chem., 17. 386, 1895 ; J. H. van't Hoff, ib., 18. 300, 1895 ;
L. Storch, ib., 19. 13, 1896 ; 26. 645, 1900 ; W. D. Bancroft, ib., 31. 188, 1899 ; F. Kohbausch,
ib., 18. 662, 1895 ; J. R. Partington, Journ. Chem. Soc, 97. 1158, 1910 ; F. H. McDougall, Jourrt.
Amer. Chem. Soc, 34. 855, 1912 ; C. A. Kraus and W. C. Bray, ib., 35. 1315, 1913 ; E. W. Wash-
bum, ib., 40. 106, 1918 ; G. A. Lewis and G. A. Linhart, ib., 41. 1961, 1919 ; A. A. Noyes and
K. G. Falk, ib., 34. 474, 1912.
11 Trans, Faraday Soc, 15. 1, 1919.
§ 12. The Solubility Law
In the case of an aqueous solution of sodium chloride, containing, say, 58*5
grams, that is, one gram-molecule per litre, we have: NaCl=Na"-j-Cr, where
68 per cent, of the salt is ionized. The condition of the equilibrium, according to
the dilution law, is
[Na-JLCr] 0-68x0'68 _ ^ ^ ^^
iNaOT^^' -0^32— =^^"'^=^*^
If Cl'-ions be added they must necessarily be accompanied by an equivalent number
of oppositely charged ions of, say, K', from, say, a solution of potassium chloride ;
or of Na'-ions from, say, sodium hydroxide ; NaOH^Na'+OH' ; etc. If either
Na'- or Cr-ions be added to the solution — say, hydrogen chloride — making the
concentration of the Cl'-ions 075 instead of 0*68 — then, in order to preserve the
constancy of the ratio 1*44, the concentration of the Na*-ions must be diminished.
This can only occur by the union of some of the Na'- and Cl'-ions to form NaCl
until the ratio K is again 1'44.
Solubility of mixtures with a common ion. — Sodium chloride is in equilibrium
with its aqueous solution, when, at a given temperature, the concentration of the
substance in solution has a certain definite and constant value — the solubility of
the substance at the given temperature. Since the sodium chloride in solution is
partly ionized, there are two equilibria to consider : first, the relation between
996 INORGANIC AND THEORETICAL CHEMISTRY
the non-ionized and the ionized salt in sohition, NaCl^Na*+Cr just indicated ;
and second, the relation between the non-ionized salt and the solid. If the solution
be saturated, we have :
NaClsoiid^NaCl8oiutionr=^Na- +Cr
If the concentration of the Na" or the CI' be augmented by the addition of
hydrochloric acid, some of the Na'- and Cl'-ions will recombine to form non-ionized
NaCl as indicated above. Consequently, some sodium chloride will be precipitated
or the solution will be supersaturated. Hence the solubility of a salt is usually
diminished in the presence of another compound with a common ion. If the
solution of the hydrochloric acid had been isohydric with the salt solution — i.e. if
the number of chlorine ions per cubic centimetre had been the same — no alteration
in the concentration of the ions would occur, and therefore no salt would be pre-
cipitated on mixing the solutions, provided no disturbing secondary action occurs.
The solubility of potassium nitrate is influenced in the same way by equivalent
solutions of potassium chloride and bromide. Hence, it follows that the two
latter salts are ionized to the same extent. A familiar example of this phenomenon
is the precipitation of sodium chloride from its saturated solution by the action of
hydrogen chloride or a concentrated aqueous solution of the acid.
The solubility of sodium chloride in the presence of increasing amounts of
hydrochloric acid, decreases more rapidly than corresponds with the simple theory.
The phenomenon is complicated by the chemical union of hydrochloric acid with
some of the water, as is evidenced by the development of heat during the reaction.
This increases the concentration of the sodium chloride solution by removing some
of the effective solvent. The precipitation of sodium chloride by the addition of
alcohol to a saturated solution of that salt is due to the union of the solvent with the
alcohol, so that less solvent is available for the sodium chloride. The phenomenon
is quite general ; for example, potassium chloride in saturated solution is precipitated
by hydrochloric acid, sodium chloride, or potassium sulphate ; copper sulphate by
copper nitrate, or sulphuric acid ; barium chloride by hydrochloric acid, or sodium
chloride ; calcium sulphate, by sulphuric acid, potassium sulphate, or calcium
nitrate. Again, nitric acid will precipitate barium nitrate from concentrated
aqueous solutions ; a nearly saturated solution of silver bromate will give a precipi-
tate of silver bromate, if either silver nitrate or sodium bromate be added to the
solution ; sodium chlorate added to a saturated solution of potassium chlorate
furnishes a precipitate of the last-named salt.
Solubility product. — It is sometimes convenient to discriminate between the
total or apparent solubility of a salt, and the amount of the non-ionized salt present
in the solution. The latter is sometimes called the real or molecular solubiUty
of the salt. According to W. Nernst, if W. Ostwald's law applies to solutions, in
a saturated solution the real solubility, like the apparent solubility, is regarded as
constant. Hence, in the dilution law for sodium chloride : [Na'][Cr]^X[NaCl],
indicated above, the concentration [NaCl] is invariable, K is constant, and
consequently also the product of the two is constant. Therefore, for saturated
solutions :
[Na-][Cr] = Constant
This relation means that in a saturated solution, the product of the " mole-
cular " concentrations of the ions is constant. This product is sometimes
called the solubility product because, from what has been already stated, the
product of the two ion concentrations determine the magnitude of the " real "
8olul)ility of the substance.
Rule for precipitation. — The solubility product of sodium chloride in solution
is not very great, and, in consequence, if concentrated hydrochloric acid be added to
a concentrated solution of sodium hydrogen sulphate, NaHS04, the solubility
ELECTROLYSIS AND THE IONIC HYPOTHESIS 997
product of sodium chloride may be exceeded, and that salt will be precipitated.
The condition of equilibrium of the mixed solutions is :
NaHS04 ^
Na- + (H- + SO4
+ +
HCIv=^
cr + (h-
I T^
NaCl H2SO4
Hence, if the product of the " molecular " concentration of any pair of ions
(with equal and opposite electrical charges) in a solution be greater than the solu-
bility product for the saturated solution formed by the union of these ions, that
substance will be precipitated ; and conversely, if a substance be present in excess,
it will be dissolved if the product of the " molecular " concentration of any pair of
ions (with equal and opposite electrical charges) in a solution be less than the solubility
product for the saturated solution formed by the union of these ions.
Apparent exceptions to the solubility law. — There are some apparent excep-
tions. The solubility of silver sulphate is increased, not diminished, by the addition
of potassium or ammonium sulphate. Lead chloride is not precipitated from its
saturated solution by lead nitrate, nor is barium chloride precipitated by barium
nitrate. In both these cases a precipitate is produced by the addition of alkali
chloride or hydrochloric acid. Calcium sulphate is not precipitated by sodium or
ammonium sulphate, but it does separate on the addition of potassium sulphate or
sulphuric acid ; potassium sulphate is not precipitated from its saturated solution
by sulphuric acid, but it does separate on the addition of potassium chloride or sodium
sulphate. The explanation is that in each case double or complex salts are formed,
which really diminish the number of lead or chloride ions in the case of lead chloride
by forming a lead chloronitrate, PbClNOs. Similarly, in the other cases BaClNOs,
CaNa.2(S04)2, Ca(NH4)2(S04)2, KHSO4, ^^c. In many cases the solubility of the
double or complex salt so formed is less than the solubility of either constituent, and
when such a solution is concentrated, the double compound separates in preference
to the single components, c.^. potassium sulphate has a solubility of 12*5; aluminium
sulphate, 8*5 ; and the double salt, KA1(S04)2.12H20, 9'5 ; potassium sulphate,
12*5 ; nickel sulphate, 6'7 ; and the double salt, K2Ni(S04)2.6H20, 6"3 ; potassium
carbonate, 110 ; sodium carbonate, 21 ; the double salt, KNaC03.6H20, 13. With
potassium and calcium sulphates the solubility of the latter (0'205) is less than the
solubility of either potassium sulphate (12*5) or the double salt CaK2(S04)2.H20 —
0"25. Calcium sulphate is accordingly precipitated in preference to the double
salt.i
The solubility law developed by W. Nernst (1889) and A. A. Noyes (1890) 2 is
based on assumptions which are not always valid. For example : (i) That the ioniza-
tion conforms with W. Ostwald's law of mass action. As a matter of fact the ratio
of the product of the concentrations of the positive and negative ions to the con-
centration of the un-ionized salt becomes greater as the concentration of the salt
increases, otherwise expressed, the value of K in, say, [M'][Cr]=^[MCl] increases
with increasing concentration, (ii) It is assumed that the concentration of the un-
ionized molecules of an electrolyte in solution and in equilibrium with its solid
phase is always constant. S. Arrhenius, however, has shown that this is not in
accord with fact, for the concentration of the un-ionized salt, i.e. the molecular solu-
bility, diminishes as the total concentration of the salt in solution is increased.
S. Arrhenius (1899) argued that the non-ionized part of a salt in a saturated solution
is not a constant because the non-ionized part of thallous chloride, TlCl, in aqueous
solution, as calculated from the electric conductivity, is 0'00179iV, while if 0'8iV-
potassium chloride is present, the total solubility of thallous chloride is 000170iV,
or less than that of the non-ionized salt in aqueous solution — of this, 0'00170, perhaj)S
40 per cent., is ionized, leaving the non-ionized part of less concentration than in
998 INORGANIC AND THEORETICAL CHEMISTRY
aqueous solute, (iii) It is assumed that the 'degree of ionization is independent of
the presence of other electrolytes. S. Arrhenius, W. D. Harkins, M. S. Sherrill,
have shown that this postulate is not in accord with fact. The solubility of calcium
sulphate is increased up to a certain point by the addition of sodium chloride, and
beyond that is diminished. A. E. Hill (1909) found that the solubility of silver
sulphate in nitric acid behaves similarly. In both cases an increased solubility
might have been anticipated owing to the formation of new compounds which would
be opposed by a small decrease in the amount of un-ionized salt in solution, for with
highly ionized salts like gypsum and silver sulphate in aqueous solution, the increased
ionization is negligibly small. The diminution in the total solubility is therefore
an effect of a diminution in the concentration of the ions of the solute, as distinct from
chemical interaction, called by S. Arrhenius (1899) the neutral salt effect. According
to this phenomenon, the presence of neutral salts increases the ionization of weak
acids present with them in solution, either by the water acquiring a greater ionizing
power, or the salt itself acting as a dissociating medium. The Neutralsalzwirkung
has been studied by many workers. G. Poma and A. Patroni found that the ioniza-
tion of cupric nitrate is augmented in aqueous solution by the presence of other
nitrates — excepting those of potassium and rubidium. A. Sachanoff and P. J.
Gontscharoff found the ionization of silver nitrate to be depressed by the addition
of other salts — vide § 13. In general, observations indicate that the addition of
sodium chloride to an acid solution results in an increase in hydrogen-ion concen-
tration, whereas adding it to an alkaline solution causes an increase in concentration
of the hydroxide ion.
The constancy of the solubility product is not maintained, for the work of
F. K. Cameron and of A. E. Hill and J. P. Simmons has shown that the product of the
concentration of the ions diminishes in value as the concentration of the electrolyte
increases — e.g. the solubility of gypsum increases with the addition of sodium
chloride up to a certain concentration beyond which it decreases ; similarly, with
silver sulphate and nitric acid ; thallous chloride in acetic acid ; with tetramethyl-
ammonium iodide in potassium hydroxide ; etc. A continuous increase in solu-
bility due to the formation of new compounds might have been anticipated ; this will
be opposed by a small decrease in the amount of an ionized salt as indicated by
S. Arrhenius' experiments, but not sufficient to counterbalance the increase due to
chemical action, since the quantity of un-ionized salt is negligibly small with such
salts as silver sulphate and calcium sulphate. Hence, in order to explain the diminu-
tion in the solubility product it seems necessary to assume that the concentrations
of the ions of the solute have been lessened. The solubility of mercuric chloride is
increased by the addition of hydrochloric acid owing to formation of a combination
of J hydrochloric acid and mercuric chloride, which has a greater solubility than
mercuric chloride alone. Again, a concentrated solution of cupric potassium sul-
phate, K2Cu(S04)2.6H20, can be separated into its components, potassium and
cupric sulphates, by diffusion. A solution saturated with respect to both salts is
possible in which the two salts are present in the ratio of their solubilities, but this
condition is disturbed by the tendency of the more soluble salt to crystallize out.
Either salt added in a soluble form will not drive the other out of its saturated
solution, but there is a definite state of equilibrium when the solution is not changed
by further additions. Equilibrium is the same whether one salt be added to a
saturated solution of the other, or directly to the double salt. These facts strike
at the foundations of the solubility product law, for the constancy of both relations
was assumed in deducing that law. The solubility product is not constant, hut dimin-
ishes in value in the presence of other electrolytes. With a decreasing value of [MR]
and an increasing value for the whole proportion, it follows that the ion product
[M-]x[R'] might remain approximately constant if perchance the proportion grows
larger to the same extent as the true solubility grows smaller. Hence, although
the premises from which the solubility product law has been deduced are not the
whole truth, yet it is conceivable that the law itself might be a sufficiently close
ELECTROLYSIS AND THE IONIC HYPOTHESIS 999
approximation to be of practical assistance in the study of reactions involving
precipitation, solution, etc. This alternative has been tested by J. Stieglitz (1908)
with respect to the solubility of the silver salts of a number of organic acids. He
finds that the solubility product must be treated as an " approximate empirical
principle" without a theoretical foundation. This conclusion is confirmed by
A. Findlay's experiments (1900) on the relative solubility of lead iodide and sulphate ;
G. Bodlander's work (1900) on the solubility of calcium carbonate in aqueous solu-
tions of carbon dioxide ; A. A. Noyes (1890-1903) on the solubility of lead iodide in
solutions of potassium iodide, of lead chloride in solutions of potassium chloride, of
calcium hydroxide in presence of ammonium chloride ; etc.
Solubility of mixtures with no common ion. — If potassium nitrate be added
to a saturated solution of silver bromate, a number of molecules of silver nitrate and
potassium bromate will be formed by double decomposition, and the solution will
be in equilibrium when these four salts have attained a definite concentration, and
each salt is itself ionized and is in equilibrium with the corresponding ions. The
condition of equilibrium is therefore complex. It may be symbolized :
KNOg + AgBrOg ^ KBrOg + AgNOg
I I I I
t + t +
""^ e? «^ =^ .
The net result is that the number of ionized and non-ionized molecules of silver
bromate in the solution is lessened, and the equilibrium :
AgBrOgsoiid ^ AgBrOggoiution
is disturbed. The original relation is restored by the passage of more silver bromate
into solution. Similarly, when nitric acid is added to a saturated solution of silver
acetate, some silver nitrate is formed, and the equilibrium :
Silver acetategon^j ^ Silver acetategoiutlon
can only regain its former value by the passage of more silver acetate into solution.
Consequently, the solubility of a salt may be increased in the presence of a
compound containing no common ion. There are a number of complications
in special cases owing to the dehydration of the solution by the added substance ;
the solute may form polymerized molecules in the presence of the solvent; etc.
The dehydrating action is illustrated in the familiar method of preparing hydrogen
chloride by dropping concentrated sulphuric acid into a concentrated hydrochloric
acid. The sulphuric acid abstracts water, and thus diminishes the effective amount
of the solvent, the hydrogen chloride is then evolved as a gas. The action is in
part attributed to the repression of the ions of hydrogen chloride at the same
time, and the consequent elimination of molecules from the water already saturated
with molecules of the same kind. Similarly, an excess of sulphuric acid precipitates
the di- or monohydrate of cadmium chloride almost completely ; the particular
hydrate formed depends on the amount of sulphuric acid added. Cadmium bromide,
mercuric chloride or bromide, and copper and stannous chlorides or bromides
behave similarly. The behaviour of many salts is thus not completely described
by the solubility product law.
References.
1 P. N. Evans, Chem. News, 86. 4, 1902 ; J. Gibson and R. B. Denison, Proc. Boy. Soc. Edin.,
30. 562, 1909.
2 W Nemst, Zeit. phys. Chem., 4. 372, 1889 ; A. A. Noyes, ib., 6. 241, 1890 ; 9. 613, 1892 ;
1000 INOKGANIC AND THEOEETICAL CHEMISTRY
16. 125, 1895 ; 26. 162, 1898 ; 42. 336, 1903 ; 46. G03, 1903 ; 8. Arrhenius, ibJ, 11. 391, 1893 ;
2. 284, 1888; 31. 218, 1899; J. G. MacGregor and E. H. Archibald, Phil. Mag., (5), 45. 151,
1898 ; J. G. MacGregor, Trans. Roy. Soc. Canada, 2. (i5, 1890 ; F. K. Cameron, Journ. Phys.
Chem., 5. 556, 1901 : A. Findlay, Zeit. phys. Chem., 34. 409, 1900 ; G. Bodlander, ih., 35. 23,
1900 ; A. E. Hill and J. P. Simmons, ib., 67. 594, 1909 ; Journ. Amer. Chem. Soc, 31. 821, 1909 ;
A. E. HiU, i6., 31.82, 1909; 32. 1186, 1910; 39.218, 1917; J. Stieglitz, ib., 30. 940, 1908;
W. D. Harkins, ib., 33. 1107, 1911 ; W. C. Bray, ib., 33. 1073, 1911 ; M. S. Sherrill, ib., 32. 749,
1910; S. Arrhenius, Zeit. phys. Chem., 4. 381, 1889 ; G. N. Lewis and P. Wheeler, ib., 56. 190,
1906 ; G. Poma and B. Tanzi, ib., 79. 55, 1912 ; G. Poma and A. Patroni, ib., 87. 190, 1914 ;
G. Poma, ib., 88. 671, 1914; B. von Szyszkowsky, ib., 78. 426, 1912; H. Goldschmidt, ib., 70.
627, 1910 ; H. C. S. Snethlage, ib., 85. 212, 1912 ; Zeit. Elektrochem., 18. 539, 1912 ; A. Sachanolf
and P. J. Gontscharoff, Journ. Ritssian Phys. Chem. Soc, 47. 1244, 1915 ; H. S. Harned, Journ.
Amer. Chem. Soc., 37. 24()0, 1915; H. A. Fales and J. M. Nelson, ib., 37. 2709, 1915; A. W.
Thomas and M. E. Baldwin, ib., 41. 1981, 1919; J. A. Witson, ib., 42. 715, 1920.
§ 13. Acids and Bases according to the Ionic Hypothesis
The ionization hypothesis enables us to place the whole subject of acidity upon a rational
basis. Without this theory, the subject would still be only so many empirically established,
disconnected, and meaningless facts.' — H. C. Jones (1913).
Acids. — It will be remembered that C. Gerhardt defined acids to be "salts of
hydrogen," the ionic hypothesis expresses a similar idea another way : " all acids,
when dissolved in water, furnish hydrogen ions." Although many substances not
usually called acids when completely ionized furnish hydrogen ions — e.g. potassium
hydrogen sulphate, KHSO4, etc. — ^yet their acidic properties are due to the presence
of H*-ions, and consequently it has been said that " there is only one acid, and that
is the H'-ion," and that " the two terms acidity and hydrogen ions are co-exten-
sive." 1 Hydrogen ions thus become the primordial acid of the older chemists.
The general and characteristic properties of acids are assumed to be the general and
characteristic properties of H*-ions, and thus the H'-ions are said to have a sour
taste, redden blue litmus, conduct electricity in solutions containing them, behave
as univalent radicles, etc. The basicity of an acid is fixed by the number of H-ions
furnished by the complete ionization of one molecule of the acid. Thus monobasic
hydrochloric acid, HCl, furnishes one H*-ion, HClF^^H'-j-Cr ; and dibasic sulphuric
acid furnishes two H'-ions, H2S04v^2H'-|-S04". Sulphuric acid also furnishes the.
ions H' and HSO4', so that it behaves also as a monobasic acid when it forms the so-
called " acid sulphates."
Why is the hydrogen ion acidic? — The action of a metal, say zinc, on an
acid is usually represented by the equation : Zn+2HCl=ZnCl2+H2 ; the ionic
hypothesis assumes that Zn+2H*+2Cr^Zn"+2Cr+H2 ; and that the difiterent
atoms have different affinities — electro-affinities — for the electric charges as assumed
by H. von Helmholtz in 1881. Since the Cr-ions are but little affected by the
change, the last equation reduces to Zn+2H"=Zn"'+H2. The action is thus
independent of the negative ion, for it involves little more than a transfer of
the positive electric charges from the two hydrogen ions to the zinc, and generally,
such typical chemical reactions can be represented as the transfer of electrical
charges from hydrogen to the metal. When the solution of zinc chloride is concen-
trated by evaporation the ionizing solvent is removed from the system and the
Zn*-- and 2Cr-ion3 recombine to form zinc chloride. The hydrogen ion is an
acid because it holds its charge less tenaciously than the metals hold their
charges. If it were otherwise, says H. C. Jones, if hydrogen held its charge as
firmly as the average metal, the acids would not be acids.
Why is hydrogen acidic in some compounds and not in others?— The
answer furnished by the ionic hypothesis is that hydrogen separates from the former
compounds as ions, but not from the latter. Hydrogen chloride in aqueous solution
is an acid because it furnishes H'-ions, but it is not an acid when dissolved in dry
benzene or dry chloroform because it does not then furnish H'-ions. Hence, says
ELECTROLYSIS AND THE IONIC HYPOTHESIS 1001
H. C. Jones, an acid is a compound which srields hydrogen ions when dissolved
in an ionizing solvent. Dry benzene and dry chloroform are not ionizing solvents.
A compound becomes an acid only when it is ionized into hydrogen ions, etc. Many
substances contain hydrogen, and they are not regarded as salts of hydrogen.
Methane, CH4 ; ammonia, NH3 ; alcohol, C2H5OH, etc. Again, H3PO2 only gives
one hydrogen ion per molecule, and the remaining two hydrogen atoms are not
ionizable, for they form an essential part of the cation H2PO2'. Silicic acid is very
slightly soluble in water, so that its aqueous solution has no effect on blue litmus.
Silicic acid is acid because it forms a salt, sodium silicate, NagSiOs, which dissolves
in water and ionizes: Na2Si03^2Na*+Si03", when electrolyzed. Neither pure
dry hydrogen chloride nor pure dry sulphuric acid is acidic ; neither compound is
ionized ; nor acts on metals ; nor decomposes carbonates ; nor colours blue litmus.
The chemical activity of different solutions containing equivalent amounts
of the different acids is referred to the concentration of the H -ions in the
solution. — The concentration of the H'-ions depends upon the degree of ionization
of the different acids. Hence, the relative strengths of the acids can presumably
be expressed in terms of the electrical conductivity of equivalent solutions. The
speed of a reaction dependent upon an acid is thus connected with the concentration
of the H*-ions. Reverting to the measurements given previously for hydrochloric,
sulphuric, and acetic acids, although the solutions contained equivalent quantities
of replaceable hydrogen per litre, acetic acid has but one two-hundredth the activity
of hydrochloric acid.
Hydrochloric acid. Sulphuric acid. Acetic acid.
Fraction ionized . . 0-78 0'51 0*004
Relative strength . . 100 70 0*5
In hydrochloric acid, a greater number of hydrogen ions are ready to react with the
metal than with acetic acid, and consequently the available hydrogen in hydrochloric
acid is more rapidly exhausted than with acetic acid where but few ions are in a
condition to react with the metal at any moment, and consequently the reaction
progresses slowly for a long time ; as fast as the available ions are exhausted, new
ions are formed by the ionization of the molecule of acetic acid. The total number
of hydrogen ions is the same in both cases, but the number in a condition to react
with the metal at any moment is very different in all three cases.
Bases. — Just as an acid has been defined to be a substance which can furnish
hydrogen ions when dissolved in water, so bases, according to the ionic hypothesis,
are substances which yield HO'-ions when dissolved in water. The basic properties
of bases are due to the OH'-ions, and in this sense it has been said that " there is
only one base, and that is the OH-ion," and that " the two terms hydroxyl ions and
base are co- extensive." The general and characteristic properties of the bases are
supposed to be the general and characteristic properties of the OH'-ions. Thus the
OH'-ions are said to have a soapy feel, turn red litmus blue, conduct electricity in a
solution containing them, etc. The acidity of a base is fixed by the number of
OH'-ions it furnishes on complete ionization of a molecule of the base. Thus, the
monoacid bases, like sodium hydroxide, ionize : NaOx=^Na*-f OH' ; and the diacid
bases, like barium hydroxide, ionize: Ba(OH)2=Ba"*-|-20H'. In a non-ionizing
solvent, the bases do not furnish hydroxyl ions, and they do not then behave like
bases. Hence, says H. C. Jones, a base is a compound which furnishes
hydroxyl ions when dissolved in an ionizing solvent. A compound becomes a
base only when it is ionized into hydroxjd ions.
The strength of a solution containing equivalent quantities of the different
bases is referred to the concentration of the OH'-ions in the solution.— The
strength of a base depends upon the degree of ionization, or on the concentration
of the OH'-ions. The strength of a base can thus be determined from the electrical
conductivity. In equivalent solutions, bases, like acids, differ very much in strength.
The alkalies and alkaline hydroxides are very strong bases, for they are ionized
to very nearly the same extent as hydrochloric acid in aqueous solution. Ammonia
1002 INORGANIC AND THEORETICAL CHEMISTRY
is a comparatively feeble base. The following numbers represent the relative
strengths of a few bases in ^^jiV^-solution on the assumption that the strength of the
base is proportional to the electrical conductivity :
LiOH NaOH KOH NH4OH
Relative strength ... 100 98 "98 2
Reactions between acids and salts. — When a highly dissociated acid is mixed
with a salt, the two react, forming another acid and salt. The change is reversible,
and the reacting system is then a further illustration of the principle of opposing
reactions. For instance, the action of dilute nitric acid on potassium hypochlorite,
HNOg+KOCl^KNOa+HOCl. If both products are highly ionized, there will
be no perceptible change in the system, but in the illustration just cited, hypo-
chlorous acid is but feebly ionized, and, since the H*-ions of the nitric acid, and the
OCl'-ions of the potassium hypochlorite, react to form feebly ionized hypochlorous
acid, the result of the reaction in dilute solutions is ionized potassium nitrate, and
feebly ionized hypochlorous acid :
HH-NOg'+K'+OCr ^ K'+NOg'+HOCl
Reactions between bases and salts. — Similar remarks apply, mutatis mutandis,
to the action of a salt on a base, and this explains how feebly ionized ammonium
hydroxide is formed in relatively large quantities when highly ionized solutions of
potassium hydroxide and ammonium nitrate are mixed together. The reaction
proceeds almost to the end :
NH^'+NOs'+K'+OH' ;^ K-+N03'+NH40H
When the base is insoluble, it will be precipitated, and the reaction will proceed
to an end quite apart from the degree of ionization of the reacting compounds. This
is the case, for example, with ferric, aluminium, zinc, and other hydroxides :
Fe-+3Cl'+3Na-+30H' ^ 3Na-+3Cr+re(OH)3
The determination of the basicity oJ acids. — W. Ostwald (1887) 2 and P.
Walden (1887) noticed empirically that if na denotes the valency of the anion, and
rig the valency of the cation, the difference D between the equivalent conductivity
of the corresponding electrolyte dilutions ^=1024 and v=^2 is D=Kna*i^c, where K
is a constant which is a function of the concentration of the solution, and for the
v=1024 and f =32 is nearly 10. If the sodium salts be employed, Wc=l ; and at
25°, the basicity of the acid is approximately ~D. For monobasic acids D is nearly
10 ; for dibasic acids, 20 ; for tribasic acids, 30 ; etc. ; otherwise expressed, the
equivalent conductivity of aqueous solutions of the sodium salts of n basic
acids increases by about nxlO units on diluting a solution from ^l^N to lo^i^.
This empirical relation is sometimes called Ostwald and Walden's rule.
The unit of conductivity is now rather smaller than that used in 1887. The
constant is now more nearly 10 "8 than 10, so that the empirical rule is more nearly
exact if the basicity of the acid be taken as one-eleventh of the difference, D, between
the equivalent conductivities of the sodium salt at dilutions t;=32 and v=1024.
The amended rule then reads :
Basicity of acid = jjD
Numerous tests of the rule have been made with organic acids. Table IV shows
some results with inorganic acids.
Examples.- — (1) G. von Knorrc (1900) showed that the equivalent conductivity of the
so-called sodium trimetaphosphato with a composition corresponding with Na3(P03)3.6H20
at dilutions v = 32, and v = 1024, is respectively 89*5 and 122-4. Hence show that these
numbers are in agreement with the inference " Nach der Ostwald- Walden'schen Regel
ELECTROLYSIS AND THE IONIC HYPOTHESIS
1003
ist deinnach die Trimetaphosphorsaure als dreibasische Saure zu betrachten." Here
jyAl024—A32)=^,-(122'4 — 89-5) = 3-3 nearly ; and the acid is accordingly tribasic.
(2) R. Kremann and W. Decolle (1907) found that the conductivity of sodium fluoride
increases on dilution from v = 32 to v = 1024 by 20*8 units. Explain how it might be
inferred that " hydrofluoric acid is dibasic."
Table IV.- — The Equivalent Conductivities of Some Sodium Salts.
Equivalent conductivities.
Sodium salt.
Differences.
J>
11
Basicity
of acid.
i?=32
t7=1024
Chloride, NaCI
113-6
126-3
12-7
11
1
Chlorate, NaClOg .
101-3
112-3
110
1-0
1
Perchlorate, NaC104
112-9
121-2
12-3
1-1
1
Nitrate, NaNOg
108-2
120-1
11-9
1-1
1
Sulphite, NaSOg .
94-5
114-6
20-1
1-8
2
Tungstate, N'a2W04
95-9
116-4
20-5
1-8
2
Cyanoplatinate, NagPtCyg
110-6
130-4
19-8
1-8
2
Orthophosphate, Na3P04
97-5
114-2
26-7
2-4
3
Orthoarsenate, Na3As04 .
101-2
127-6
26-6
2-4
3
Trimetaphosphate, Na3(P03) .
89-4
119-4
30-0
2-8
3
Fleitmann's metaphosphate,
Na4P40i2.4H20 .
85-6
126-2
40-6
4-1
4
At extreme dilutions, more particularly with, inorganic salts, secondary reactions —
e.g. hydrolysis — may interfere with the conductivity determinations ; in . other
cases, the dilution v==1024 does not appear great enough to ionize all the salt. There
are accordingly some difficulties, but, in a great many cases, the method furnishes
circumstantial evidence of the basicity of the acid in question.
References.
1 H. C. Jones, A New Era in Chemistry ^ New York, 1913 ; The Nature of Solution, London, 1917.
2 W. Ostwald, Zeit. phys. Chem., 1. 74, 1887 ; 2. 843, 1888 ; P. Walden, ib., 1. 52, 1887 ; 2.
49, 1888 ; R. Kremann and W. Decolle, Monatsh., 28. 917, 1907 ; G. von Knorre, Zeit. anorg.
Chem., 24. 369, 1900.
§ 14. The Strengths of Acids and of Bases
1 have no doubt that fixed salts choose one acid rather than another in order that they
may coalesce with it in more intimate union.- — John Mayow (1674).
I maintain that when several acids act upon one metallic base, the action of one acid
does not overpower that of the others so as to form an insulated combination, but each of
the acids has a share in the action proportionate to its capacity for saturation and to its
quantity.- — ^C. L. Berthollet (1803).
The strength of an acid or base refers to the extent to which the acid or base
exhibits acidic or basic properties respectively. The terms affinity, avidity, and
activity are sometimes employed synonymously with " strength," but there are
objections to each of these. The term strength, too, is often used where concentra-
tion is really meant. Concentration refers to the "quantity in unit volume"
expressed in, say, " grams per litre ; " or in any other convenient form — say the
number of gram-molecules per litre, etc.
The action of sulphuric acid on sodium chloride which results in the formation
of hydrochloric acid, seems to prove that sulphuric acid is stronger than hydrochloric
acid ; ] again, when hydrochloric acid is added to a solution of silver sulphate,
silver chloride is precipitated. The hydrochloric acid expels the sulphuric
acid from its combination with silver : Ag2S044-2HCl=2AgCl+H2S04, and it
1004 INORGANIC AND THEORETICAL CHEMISTRY
seems as if hydrochloric acid is stronger than sulphuric acid. These two con-
clusions are contradictory, and there must therefore be a fallacy in our reasoning.
We have wrongly assumed that the two acids were competing for sodium and for
silver under similar conditions. This is not the case. When hydrochloric and sul-
phuric acids compete for the sodium, the hydrochloric acid, being volatile, escapes
, from the system as fast as it is formed ; while the non- volatile sulphuric acid alone
remains behind. Again, when sulphuric and hydrochloric acids are competing for
silver, the hydrochloric acid carries the silver away from the sulphuric acid as an
insoluble precipitate of silver chloride. Still further, hydrosulphuric acid is notori-
ously a very feeble acid, and yet it can displace relatively strong acids from combina-
tions with the metals. Thus, it will precipitate lead sulphide from solutions of lead
chloride ; copper sulphide from solutions of copper sulphate, etc. Here, again, the
feeble acid does its work by removing the metal from the solution as an insoluble
sulphide.
To compare the relative strengths of the acids, and, mutatis mutandis, of the bases,
it is necessary that the comparison be made under conditions where the reacting
acids and the products of the reaction are in the same physical condition — say, all
in solution. Thus, if an equivalent of a solution of sodium hydroxide be mixed with
an equivalent of a solution of sulphuric and of hydrochloric acids, the two acids can
compete for the one base under the same conditions, and hence the stronger acid
will be able to unite with more sodium than the weaker acid. In 1803, C. L. BerthoUet
pointed out that :
When a neutral salt is dissolved and an acid added to the solution, the free acid enters
into competition with the combined acid, and they both act on the alkaU base in the ratio
of their respective concentrations as though no combination had existed. It cannot there-
fore be said that if all the conditions remain equal one acid displaces another from the base
with which it has been united, but the base is shared between the two acids in the ratio of
the concentrations and affinities of the respective acids.
It is found experimentally that the same result is obtained when equivalent quantities
of sodium hydroxide, sulphuric acid, and hydrochloric acid are mixed together as
when equivalent quantities of sodium sulphate and hydrochloric acid, or equivalent
quantities of sodium chloride and sulphuric acid are mixed, provided, of course, the
whole of the system has been allowed to stand long enough for equilibrium. This
fact is represented by the equation : 2HCl+Na2S04=F^2N*aCl+H2S04, which,
when translated into the language of ions, reads :
2HC1
^
2H-
+
+
2cr
+
Na2S04
^
SO4"
+
2Na-
I
11
€
g
»
o
Q
The measurement of the relative strengths of acids and bases. — The
proportions of a base shared between two acids, or of an acid between two bases,
cannot be determined by the ordinary methods of chemical analysis without
disturbing the equilibrium of the mixture. The distribution of an acid between
two bases, or of a base between two acids, must be determined by physical processes
which do not interfere with the solution.
In illustration, the heat of neutralization of sodium hydroxide by sulphuric acid is 31 "38
Cals. ; and by hydrochloric acid, 27 '48 Cals. If, therefore, on mixing hydrochloric acid
with sodium sul])hato, all the sulphuric acid wore displaced by the hydrochloric acid, the
thermal effect rc^sulting from the decompo.-iition of the sodium sulphate, and the formation
of the sodium chloride would be 27-48— 31-38 = —3-9 Cals. After making a small allowance
for secondary reactions between sodium sulphate and sulphuric acid, J. Thomsen found
that the thermal value of the reaction was —2-6 Cals. Hence it follows that — 2-6-i-— 3-9,
ELECTROLYSIS AND THE IONIC HYPOTHESIS
1005
or about two-thirds of the hydrochloric acid combines with about two-thirds of the base
to form sodium chloride ; and about one-third of the sulphuric acid combines with the
other third of the base to form sodium sulphate.
A similar result was obtained with a mixture of sodium chloride and sulphuric acid,
as with sodium sulphate and hydrochloric acid. Consequently, in the competition
of sulphuric and hydrochloric acids for sodium under comparable conditions, the
hydrochloric acid can hold twice as much of the base as the sulphuric acid, and conse-
quently, hydrochloric acid is nearly twice as strong as sulphuric acid. Similar
results have been obtained by measuring the specific gravity, index of refraction,
absorption of light, etc. Some results obtained by three different methods are
shown in Table V.
Table V.—
Relative Strengths ov Acids.
Acid.
Thomsen's thei^al
process.
Ostwald's specific
gravity process.
Molecular
conductivity.
Hydrochloric acid .
Nitric acid ....
Hydrobromic acid .
Sulphuric acid
Phosphoric acid
Acetic acid ....
100
100
89
49
• 1
98
100-00
95-00
66-7
1-23
100-0
99-6
100-0
65-1
7-3
0-4
The relative strengths of the different acids and bases have also been determined
by measuring the effects of the different acids on the speed of different reactions —
e.g. the hydrolysis of acetamide, cane sugar, methyl acetate, etc. The actual
numbers obtained by the different methods are not always quite the same, possibly
because of the different conditions under which the experiments are made. If
two acids are under the same physical conditions and differently influence the
speed of a given reaction, the acid which induces the greater velocity is assumed to
exert the greater chemical force. It has been found that in many cases the effects
produced by one acid of different concentrations is roughly proportional to the
electrical conductivity of its solution, which in turn is proportional to the concen-
tration of the hydrogen ions. A. A. Noyes and A. A. Blanchard illustrate this by
the following experiment :
Mix 40 c.c. of 0-5iV-KI, 40 c.c. of O-S^V-KBrOj, 40 c.c. of starch solution, and make
the whole up to two litres. Put 400 c.c. of this solution in each of four bottles and introduce
10 c.c. of each of the |A'^-acids- — hydrochloric, sulphuric, chloroacetic, and acetic acids- — into
each of the four cylinders, as quickly as possible. The bottles are immediately stoppered
and shaken. The solution containing the hydrochloric acid turns deep blue almost immedi-
ately ; the sulphuric acid in about half a minute ; the chloroacetic acid in three or four
minutes ; and the acetic acid in three or four hours.
Again, the addition of an excess of foreign acid to the reaction between potassium
chlorate, potassium iodide, and hydrochloric acid : KC103+6HC1+6KI=7KC1
4-3H2O+3I2, or to a mixture of bromic acid and hydriodic acid accelerates
the change. The effect with different acids varies with their strength (affinity) :
HBr, HCl, HNO3, H2SO4. This order is virtually the order as deduced from
measurements of the degree of ionization of the different acids, and accordingly
the catalytic effect is said to be due to the presence of hydrogen ions. Organic
and other feebly ionized acids exert very little catalytic action. It was therefore
postulated by W. Ostwald that the hydrolytic activity of acids is proportional to
the degree of the assumed ionization.
S. Arrhenius found that the hydrolytic activity of strong acids is augmented by
the addition of a neutral salt of the acid. Thus, the rate of hydrolysis of cane
1006 INORGANIC AND THEORETICAL CHEMISTRY
sugar by an aqueous solution of hydrogen chloride is considerably increased by the
addition of sodium chloride or calcium chloride, although these substances decrease
the degree of ionization of the acid. The effect was styled neutral salt action by
S. Arrhenius. If the action of the acid is due to hydrogen cations, the presence of
a neutral salt having the same anion should diminish the dissociation of the acid
and consequently its activity, but this is not the case. Actual experiment yields
results which are just the reverse of that which the theory of ions would indicate.
J. W. McBain and co-workers have shown that the alleged effect can be accounted
for on the assumption that the un-ionized salt is catalytically active without assuming
that the solvent acquires a greater ionizing power, or that the salt acts as an ionizing
medium.
Again, the molecular hydroly tic activity of strong acids is decreased by increasing
dilution, whereas, if the activity were proportional to the degree of ionization, the
reverse should obtain. Hence, G. Senter, H. C. S. Snethlage, S. F. Acree, etc.,i
conclude that an electrolyte in solution may enter into chemical reactions not only
by means of its ions, but also by means of its un-ionized molecules — otherwise
expressed, the catalytic activity of acids is the joint effect of the hydrogen ions and
of the un-ionized molecules. F. P. Worley has emphasized that instead of the rate
of hydrolysis of ethereal salts and cane sugar being proportional to the concentration
of the hydrogen ions, the two properties are altered in different directions by changes
of concentration. For example, if n denotes the molecular proportions H2O : HCl,
H, the molecular hydrolytic activity, and a, the degree of ionization of hydrochloric
acid.
n
,
30
40
50
60
80
100
200
H
.
385
323
290
269
243
229
201
a
.
. 0-726
0-778
0-813
0-837
0-859
0-871
0-901
Consequently, the catalytic activity of the acids is not proportional to the concen-
tration of the hydrogen ions ; and, adds F. P. Worley, " if hydrogen ions and un-
ionized molecules are both chemically active, ... in time, the extreme supporters of
the ionic hypothesis may admit that both ions and un-ionized molecules may be
concerned in electrolytic conductivity ! "
References.
1 S. Arrhenius, Zeit. phys. Chem., 4. 381, 1889 ; H. C. S. Snethlage, ib., 85. 211, 1912 ; Zeit.
Elektrochem., 18. 539, 1912 ; S. F. Acree, Amer. Chem. Journ., 43. 352, 1912 ; H. C. Robertson
and S. F. Acree, ib., 49. 474, 1913 ; F. P. Worley, Phil. Mag., (6), 27. 459, 1914 ; G. Senter, Journ.
Chem. Soc, 91. 460, 1907 ; 95. 1827, 1909 ; G. Senter and A. W. Porter, ib., 99. 1049, 1911 ;
A. Lapworth, ib., 93. 2187, 1908 ; 107. 857, 1915 ; S. Arrhenius, ib., 105. 1424, 1914 ; H. M. Dawson
and T. W. Crann, ib., 109. 1262, 1916 ; H. S. Taylor, Medd. Vet. Nobel-InsL, 2. 34, 35, 37, 1913 ;
3. 1, 1914 ; H. Goldschmidt and A. Thusen, Zeit. Elektrochem., 18. 39, 1912 ; J. W. McBain and
F. C. Coleman, Journ. Chem. Soc, 115. 1517, 1919 ; J. W. McBain and J. Kam, ib., 115. 1332,
1919; S. Arrhenius and E. Andersson, Medd. Vet. Nobel-Irist., 3. 1, 1918.
§ 15. The Neutralization of Acids and Bases
The importance for chemistry of the fact that hydrogen and hydroxyl ions cannot
remain in the presence of one another uncombined is difficult to over-estimate. Could
these ions remain separate, then an acid would not neutralize a base, and all salt formation
from the process of the neutralization of acids and bases would be excluded. — H. C. Jones
(1913).
The term neutral has been used somewhat vaguely, implying that the substance
is neither acidic nor basic. The test for acidity or basicity depended upon the
behaviour of the solution towards a solution of litmus. If other indicators are used,
the conclusions might be different, because a substance might appear acidic towards
ELECTROLYSIS AND THE IONIC HYPOTHESIS 1007
one indicator, and neutral towards another. The ionic hypothesis, as we have seen,
refers acidity to the presence of hydrogen ions, and alkalinity to the presence of
OH'-ions, and the term " neutrality " refers to the case where the concentration of
both ions are the same, or both ions are absent. We have seen that water is a con-
stant product of the reaction between the solution of an acid and of a base :
HCl+KOH^KCl+HaO ; H2S04+2NaOH^Na2S04+2H20, etc. When solutions
of an acid and base are mixed, the hydrogen and hydroxyl ions of acid and base
respectively combine to form water, because water only ionizes to an inconceivably
small extent, and the two kinds of ions — H* and OH' — cannot remain in the presence
of one another uncombined. Hence when aqueous solutions of acids and bases are
mixed together, OH'- and the H*-ions are removed from the solution, and the
reaction is almost completed :
H--f cr+K--f OH' ^ K-+cr-f H2O
What is here stated with respect to hydrochloric acid and potassium hydroxide
applies, mutatis mutandis, to any strongly ionized acid and base ; and consequently,
the neutralization of strongly ionized acids and bases involves little more than the
formation of water : H*+0H't==H20, because the other ions present before the acids
and bases are mixed remain after the reaction is over. If, however, the water be
evaporated from the solution, the ions recombine to form the salt, and the result
of the reaction is then correctly symbolized : HCl+K0Hr=^KCl+H20. This
reaction probably also occurs if very concentrated solutions or solids are mixed,
whereas the neutralization o£ acids and bases in dilute solutions involves the
formation of water, not salt molecules.
The heat of neutralization of dilute solutions. — This view is further supported
by the fact that with dilute solutions of the strong acids and bases, the thermal
value of the process of neutralization — heat of neutralization — is the same. For
example,
Ba(0H)2
13-8 Cals.
HIO2
13-5 Cals.
Hence, neutralization is an isothermal process ; the heats of neutralization of
dilute solutions of the strong acids and bases do not depend upon the specific nature
of the acid or base ; and the formation of water in these reactions is accompanied
by the evolution of approximately 13* 7 Cals. of heat.
The law only describes the thermal effect attending the neutralization of solutions
sufficiently diluted to ensure complete ionization of acid, base, and salt ; it presup-
poses that no new electrically neutral molecules are formed. As a corollary, it follows
that if two completely ionized salts are mixed, there will be no thermal change pro-
vided the salts are completely ionized before and after the mixing, and no other
electrically neutral molecules are formed. The fact that if two neutral salt
solutions at the same temperature are mixed together, no change of temperature
occurs, was discovered by H. Hess in 1841, and is called Hess' law of thermo-
neutrality. For example :
Hydrochloric acid
LiOH
. 13-7
NaOH
13-7
KOH
13-7
Ca(0H)2
13-8
Sodiiun hydroxide
HCl
. 13-7
HBr
13-8
HI
13-7
HNO3
13-7
Before Mixing.
Atter Mtxtng.
Calcium nitrate, Ca(N03)2
. 451
Calciima sulphate, CaS04.2HaO
. 642
Potassium sulphate, K2SO4
. 601
Potassium nitrate, KNO3
. 409
Thermal value
. 1052
Thermal value
. 1051
The ionic hypothesis indicates clearly the conditions which must be fulfilled before
Hess' law of thermoneutrality is applicable, and it would be difficult to find such a
strikingly successful explanation by any other known hypothesis. r' w
Heat of ionization. — If the acid and base are but partially ionized, the heat
1008
INORGANIC AND THEORETICAL CHEMISTRY
of neutralization is not only determined by the beat of formation of water — 13" 7
Cals. — but it is also determined by the thermal value of the energy required to com-
plete the ionization of acid and base. When a dilute solution of hydrofluoric acid
is neutralized by sodium hydroxide, for example, the sodium fluoride formed during
the reaction is fully ionized, whereas the hydrofluoric acid at the commencement of
the process : HF+NaOH=NaF+H20+16-27 Cals. is not fully ionized. Hence in
addition to the formation of water, there is a continuous ionization of hydrofluoric
acid during the process of neutralization, and the fact that more heat is produced
has been assumed to prove that the ionization of the acid is accompanied by the evolu-
tion of heat. The heat of neutralization of hypochlorous acid, HOCl, by sodium
hydroxide, NaOH, is: HOCl+NaOH=NaOCl+H204-9-8 Cals., a number less
than the normal value 13*7 Cals. The salt, NaOCl, and the base, NaOH, are com-
pletely ionized ; while the acid, HOCl, is but feebly ionized. Hence, it is assumed
that the ionization of HOCl is an endothermal process. Similarly, when ammonia
is neutraUzed : NH40H+HCl=NH4Cl-f H20+12-1 Cals., it is assumed that the
low results are due to the absorption of heat during the ionization of ammonium
hydroxide. S. Arrhenius gives the following Table VI, in his Recherches sur la
conductibilite galvanique des electrolytes (Stockholm, 1883) :
Table VI.— Heats of Neutralization of Acids
AND Bases.
HCl
HNOa
CHsCOOH
HCOOH
KCOOH),
15-85
iHaS
HCy
IHjjCO,
NaOH .
13-7
13-7
13-3
13-4
14-3
3-85
2-9
10-2
KOH .
13-7
12-8
13-3
13-4
14-3
15-7
3-85
3-0
101
NH^OH.
12-45
12-5
12-0
11-9
12-7
14-5
3-1
1-3
5-3
|Ca(0H)2
140
13-9
13-4
13-5
18-5
15-6
3-9
. —
9-8
iBa(OH),
13-8o
13-9
13-4
13-5
16-7
18-4
. —
. —
111
iSr(0H)2
141
13-9
13-3
13-5
17-6
15-4
• —
""-
10-5
law of thermoneutrality without the ionic hypothesis. — It would
not be fair to pass by this explanation without indicating how H. Crompton (1897)
showed that the phenomena could have been deduced without the aid of the ioniza-
tion hypothesis. It follows from the calorimetric observations of J. Thomsen and
of F. Stohmann and his co-workers i that for non-associated organic compounds —
presumably non-ionized — the replacement of hydrogen, in a compound RH, by one
and the same radicle M is attended by a constant heat change which is independent
of the radicle R to which the hydrogen is attached ; and generally, in non-associated
organic compounds if M is constant, the thermal change in the reaction RH-f M
=RM-|-H is constant and independent of R ; and similarly, in the reaction ROH-f M
=RM+OH, the heat change will remain constant so long as M is constant.
In the neutralization of an acid by a base, in dilute solution, where the solutes
may be assumed monomolecular, the changes which occur involve the splitting of
the acid HR=H+R+Oi and of the base MOH=M+OH-f ^2» and the combination
M-f R=MR-f Qg, and of H-f 0H=H20-f ^4- First, if the base be constant and
the acid varied, the terms Q2 and Q4 will be constant. The changes HR=H+R-f ft
and M+R=MR-fQ3 are analogous with the changes which occur in the replace-
ment of H by M in the non-associated organic compound RH. Hence, unless
inorganic compounds behave quite differently from organic compounds, the heat of
neutralization of an equivalent of any non-associated acid HR by an equivalent
of one and the same non-associated base MOH, is always the same, and independent
of the character of the acid. H. Crompton applies a similar argument to show that
the heat of neutralization of an equivalent of any non-associated base MOH by the
equivalent of one and the same acid HR is always the same and independent of
the character of the base. These two conclusions are then generalized : The heat
of neutralization of any acid by any base is independent of the character of the
ELECTROLYSIS AND THE IONIC HYPOTHESIS 1009
acid or base, so long as these are non-associated. H. Crompton further showed that
in dilute solutions, the solute assumes a non-associated state.
The heat of neutralization of an acid by a base involves the heat of replacement
M0HH-H=H20-f-M, and the heat of replacement HR+M=MR+H, it therefore
follows that since OH is a negative radicle like R, changing the radicle does not
affect the heat of replacement of H by M or of M by H. Hence the heat of the
first process should exactly balance that of the second, and an acid should be neutra-
lized by a base without any thermal change whatever. With organic compounds
the heat of formation of an ethereal salt from an acid and alcohol is usually very
small — sometimes positive, sometimes negative, but the results are complicated by
association phenomena and the heats of association are an integral part of the thermal
values of the reactions. Again, with dilute solutions of acids and bases, the acid,
base, and salt are in a very attenuated condition more or less comparable with the
gaseous state ; but the molecules of water produced in the reaction MOH+HR
=MR4-H20 will immediately pass to the liquid state characteristic of the solvent.
The observed heat of the reaction is therefore mainly that caused by the condensa-
tion of the molecule of water from the gaseous to the liquid state and their subsequent
polymerization or association. The molecular heat of condensation of water vapour
is nearly 10*8 Cals. at 0°, and this is the same order of magnitude as the 13'4 Cals.
observed by J. Thomsen for the heat of neutralization in dilute solutions. J. Thom-
son also found that the heat of neutralization decreases with a rise of tempera-
ture in a way comparable with the effect of a rise of temperature on the
heat of vaporization. These comparisons are closer than might have been
anticipated. Hence, argues H. Crompton, the assumption that salts are ionized
in aqueous solution is unnecessary to explain Hess' law of thermoneutrality ; and
" it is not only unnecessary, but it is inadequate, for it does not bring the behaviour
of electrolytes, as far as the heat changes which accompany the formation of salts
in aqueous solution are concerned, into line with the behaviour of non-electrolytes."
Hydrolysis. — It will be remembered that in hydrolysis, a salt reacts with water
to form the free base and free acid, or free acid and a basic salt. Hydrolysis is thus
a reversion o£ the process of neutralization. Reactions like those previously
discussed are probably slightly reversible, but, in addition to those examples, there
are many others where the back reaction is more pronounced. Hydrocyanic acid,
HCy, for instance, ionizes : HCy^H+Cy^ With potassium cyanide, KCy, in
aqueous solution, KCy^K'+Cy'. In the latter case, some of the H*-ions of the
water unite with the Cy'-ions of the salt to form molecules of hydrocyanic acid, HCy.
The equilibrium is disturbed, and more molecules of water ionize : H20^H*-|-0H'.
The new H'-ions combine with more Cy'-ions and the process continues until the
concentration of the OH'-ions becomes large enough to prevent the further ionization
of the water. The solution then contains an excess of OH'-ions, and free hydro-
cyanic acid, as well as potassium cyanide, and K'-and Cy'-ions. The free hydro-
cyanic acid can be recognized by its smell ; and the OH'-ions can be recognized by
the alkalinity of the solution.
The ionic hypothesis in analytical chemistry. — The language of the ionic
hypothesis has penetrated analytical chemistry — more particularly the qualitative
analysis taught in our schools — and as a result, tests for metals and acid radicles
are described as tests for the ions. Many, however, doubt if anything is really gained
by describing the facts of an essentially practical art in the language of so hypothetical
a doctrine. Be that as it may, since both chlorides and hydrochloric acid are supposed
to furnish chlorine ions : HCl^H'-f-Cr, or NaCl^Na*-f CI', it is assumed that
the test for hydrochloric acid or for a chloride is a search for chlorine ions. The
silver nitrate solution used in making the test is, supposed to be ionized : AgNOs
x=^Ag*+N03' ; consequently, when silver nitrate is added to sodium chloride solution,
the mixed solution momentarily contains: Ag'+N03'-f-Na*-|-Cr ; but, since a
small proportion of silver chloride is ionized and the salt is but very slightly soluble
in water, it precipitates at once. Silver chlorate, AgClOs, is soluble in water, and
VOL. I. 3 T
1010 INORGANIC AND THEORETICAL CHEMISTRY
accordingly, when silver nitrate is mixed with a solution of, say, potassium chlorate,
there is no precipitation. The solution contains four different kinds of ions but no
chlorine, Cl'-ions : AgNOg-fKClOa^K'+ClOg'+Ag'+NOg'. Hence, silver nitrate
is a test for chlorine ions, but not for chlorate ions. If potassium cyanide KCy
in aqueous solution be added to a solution of silver nitrate, AgNOs, a precipitate of
silver cyanide, AgCy, is obtained: Ag-|-N03'+K-+Cy'=AgCy+K--l-N03'. If
an excess of potassium cyanide be added the precipitate redissolves, and it can now
be shown that the solution no longer contains Ag'-ions in appreciable quantities,
since (1) sodium chloride gives no precipitation of silver chloride ; (2) on electrolysis
silver is deposited on the anode not on the cathode, as is the case when a solution of
silver nitrate is electrolyzed ; (3) a crystalline compound, KAgCy2, is obtained on
concentrating the solution. The solution of silver cyanide in potassium cyanide
indeed ionizes thus: KAgCy2=F^K"+AgCy2'. Here again it matters very little
whether the facts be described in terms of the ionic hypothesis or in terms of basic
and acidic radicles. The choice can only be decided by personal opinion since the
ions in solution still remain hypothetical units.
References.
* J. Thomsen, ThermochemiscTie Untersuchungen, Leipzig, 4. 263, 1886 ; F. Stohmann,
C. Kleber, and H. Langbein, Journ. prakt. Chem., (2), 40, 341, 1889 ; F. Stohmann, P. Rodatz, and
W. Heizberg, ib., (2), 36. 1, 1887 ; F. Stohmann, Zeit phys. Chem., 2. 29, 1882 ; 6. 334, 1890 ;
10. 410, 1892 ; H. Crompton, Journ, Chem. Soc., 71. 925, 946, 953, 1897.
CHAPTER XVI
ELECTRICAL ENERGY
§ 1. The Factors of Energy
All the differences discoverable in the effects of electricity (obtained from different
sources) may be owing to its being less intense but produced in much larger quantity from
some sources rather than from others.- — -W. H. Wollaston (1801).
The term electromotive force in electricity is equivalent to the term chemical activity
or affinity just as the term quantity of electricity corresponds to the chemical notion of
valency.— G. Salet (1867).
It has been shown that every form of energy has a dual nature, for all the better-
known forms of energy appear as if they were two dimensional in that they are com-
pounded of two factors — one the capacity factor, the other the intensity factor.
The latter determines whether a given change will occur. The capacity and intensity
factors of heat energy are respectively entropy and temperature. The flow of heat
is not determined by the quantity of heat in a given system, but rather by the
difference of temperature. The heat in a furnace can do work not because it is hot,
but because it is hotter than its surroundings. With electrical energy, the quantity
of electricity expressed in suitable units — say coulombs — is the capacity factor,
and the electromotive force expressed in suitable units — say, volts — is the intensity
factor. The product of these two factors expresses the magnitude of electrical
energy.
What are the factors of chemical energy ?— A. Butleroff (1861) pointed out
that it is necessary to distinguish the quantity of affinity from its intensity, that is,
the smaller or greater energy with which it tends to become active. If chemical
energy can be resolved into two factors, the one factor must be analogous to
the capacity, and the other to the intensity factor of thermal or electrical energy.
J. W. Gibbs 1 calls the intensity factor of chemical energy the chemical potential ;
G. H. Helm calls it the chemical intensity; and it is often called the driving force of a
reaction. These terms are employed with the idea of evading the vagueness of the
old term, chemical aflftnity, which is undoubtedly the correct designation for
" chemical intensity." Now, the quantity of a substance which takes part in any
chemical change is proportional to the equivalent weight of the substance, where
the term equivalent weight refers to quantities of matter which have the same
valency. Assuming then that the chemical equivalent is the capacity factor of
chemical energy, we may write —
Chemical energy = Equivalent weight x Chemical affinity ; or,
Chemical energy = Equivalent weight X Chemical intensity.
Some follow G. Salet, vide supra, and hold that valency is the capacity factor, but
that does not seem the right thing to do.
If two bodies at the same temperature be placed in contact, there will be no
apparent conduction of heat from the one to the other ; but when the temperature
of the one body — i.e. the intensity factor — is higher than that of the other, heat
will pass from the hot to the cold body, so that the cold body is warmed and the
hot body is cooled. So with chemical energy. We assume that the molecules
of every substance possess a specific amount of chemical energy, which has a
1011
1012 INORGANIC AND THEORETICAL CHEMISTRY
definite intensity under certain specified conditions. One substance can only
react with another when the intensity of the energy associated with the original
mixture is greater than that of the final system. If the intensity of the energy
associated with the original mixture be the same as that associated with the products
of the reaction, no reaction will take place, for the system will be in stable equilibrium ;
if the intensity factors are not equal, the energy will not usually be in stable equi-
librium. Just as the value of heat energy as a source of power depends on its tempera-
ture, so does the availability of chemical energy depend on the magnitude of its
intensity factor.
W. Ostwald has drawn attention to the fact that if the chemical process be
performed in a voltaic cell, the work derived from that process will be transformed
into an equivalent amount of electrical energy. The quantity C of electricity
generated when w grams of a compound are formed or decomposed in a cell will
be w=€C, where e denotes the electrochemical equivalent. This means that the
capacity factor — quantity of electricity — is proportional to the quantity of matter
decomposed, and that the capacity factor of the electrical energy is proportional to
the capacity factor of the chemical energy. If the current does no work other than
the chemical decomposition of the compound into its elements, and the difference
of potential is E, the work done by the current will be EC. Again, if chemical
aflSmty performs an amount of work W in building up w grams of a substance from
its elements, wW units of work will be required. Accordingly, wW=EG. By
substituting w=€C, and reducing the expression to its simplest terms, €W=E ;
otherwise expressed, the electromotive force of a cell E is equal to the chemical
affinity per gram equivalent of the compound in question. Consequently, as
G. Salet showed, the product of the quantity of electricity into electromotive force
not only represents the electrical energy of a battery, but also measures the work
of affinity which that energy transformation can perform. It therefore follows that
the respective intensity factors of chemical and electrical energies are proportional,
and since the electromotive force is proportional to the intensity factor of electrical
energy, it follows that the electromotive force is proportional to chemical afl&nity.
We see, then, that electromotive force and chemical affinity are manifestations of
one form of energy ; or, in the words of M. Faraday, " the forces called electricity
and chemical affinity are one and the same." The problem is solved for conductors
of electricity — electrolytes. Chemical action takes place when the electrical potential
or the chemical affinity of the reacting substances is greater than that of the reacting
products. We can to-day express the " affinity " between a number of reacting
substances roughly in terms of difference of potential. The measurement of electrical
potential of voltaic combinations under conditions where disturbing effects due to
thermal changes, secondary reactions, etc., are eliminated, will represent the free
energy or affinity of the reaction in question. How this may be done for non-
conductors of electricity has not yet been determined.
The temperature or intensity factor of heat energy required for the decomposition
of many substances — say calcium or potassium chloride — is so great that commercial
methods of decomposing these substances by thermal energy are not profitable.
A great many compounds thus appear to be very stable when heated at high tempera-
tures ; these can often be decomposed by electrical energy at a comparatively low
voltage (intensity factor). This illustrates how the commercial production of
metals like aluminium, calcium, etc., were not particularly successful until electrical
methods were adopted. The prediction of C. L. BerthoUet (1803) has been fulfilled :
The electrical current has furnished chemistry with an agent whose energy may be
carried to a degree, which as yet can scarcely be imagined, and which will furnish the means
of producing in the formation and decomposition of chemical combinations, effects unfore-
seen, and superior to those which it is possible to obtain by the action of heat.
It has been suggested, too, that if a source of energy with a particularly high
intensity factor were available, it would most likely be possible to decompose many
J
ELECTRICAL ENERGY 1013
of the so-called elements into still simpler substances, but this, of course, is merely
a speculation.
References.
1 W. Ostwald, Zeit. phys. Chem., 15. 399, 1895 ; G. Salet, Laboratory, 1. 248, 1867 ; J. Popper,
Die physikalischen Orundsdtze der elektrischen KraftubertraguTig, Leipzig, 1884; J. W. Gibbs,
Trans. Connecticut Acad., 3. 108, 343, 1876-8 ; G. Helm, Grundzuge der mathematischen Chemie,
Leipzig, 1894 ; M. Faraday, Phil. Trans., 124. 77, 1834 ; C. L. Berthollet, Essai de statique
chitnique, Paris, 1803.
§ 2. Electrochemical Series of the Elements
An electrochemical series is obtained by arranging substances in accord with their
electrical properties, and this series is better than any other for giving a general idea of
chemistry. — J. J. Berzelius (1825).
The metals precipitate one another after a certain order. — ^T. Bergmann (1779).
Near the beginning of the fourth century, Zosimus mentioned the fact that when
iron is immersed in a solution of a copper salt, the iron acquires a coating of copper ;
and miners have frequently noticed that their iron tools become coated with copper
when brought in contact with the water percolating through certain mines. We now
know that such water may hold in solution copper sulphate from the oxidation of
ores containing copper sulphide associated with iron sulphide, and we have also
learned that when the copper is deposited, an equivalent amount of iron passes into
solution. The reaction is represented in symbols, Fe+CuS04=Cu+EeS04. Curi-
ously enough, even as late as the sixteenth century Paracelsus attributed the
phenomenon to the transmutation of iron into copper. The transmutation hypo-
thesis certainly appeared a very plausible explanation of the facts. As T. Bergmann
emphasized near the middle of the eighteenth century in his De prcecipitatis metallicis :
The man who first saw a metal corroded by a limpid menstruum, in such a manner that a
body so extremely ponderous and so opaque should gradually and entirely disappear, and
afterwards, upon the addition of a suitable precipitant to a liquid which appeared to be
simple and homogeneous, saw the metal separate and again come into view, that man,
I say, who first saw this, must have been struck with astonishment and admiration.
Persons accustomed to these wonderful phenomena neglect, perhaps too much, the accurate
investigation of them, though these operations are of the highest importance, and form as
it were the whole of the effective part of chemistry.
Metallic magnesium will displace hydrogen from dilute acids :
Mg-[-H2S04=MgS04+H2 ; or in the language of the ionic hypothesis :
Mg-j-2H -l-S04"=Mg'+S04''4-H2. Magnesium will also precipitate zinc from a
solution of a zinc salt : Mg+ZnS04=MgS04+Zn ; or in the terms of the ionic
hypothesis : Mg+Zn"+S04"=Mg"+S04"+Zn. Zinc in turn will precipitate iron
from iron salts ; iron will precipitate copper from copper salts ; copper will precipi-
tate silver from silver salts, etc. By treating the different metals in a similar manner
it has been found possible to arrange them in a series such that, under like conditions,
any metal in the list will displace those which follow it, and be displaced by those
which precede it.
Again, when zinc is treated with dilute acids under suitable conditions in a voltaic
cell, so as to eliminate disturbing effects, the reaction produces an electric current
at a certain voltage. If the zinc be replaced by some metals — aluminium, magnesium,
etc. — the voltage of the cell is increased ; and conversely, if the zinc be replaced by
other metals — cadmium, iron, cobalt, etc. — the voltage of the cell is diminished. It
is thus possible to arrange the elements in a series representing the potential difference
in volts which is developed between the metals and solutions of their salts. The list
1014
INORGANIC AND THEORETICAL CHEMISTRY
of the elements so arranged is called the electrochemical series. A more complete
list is indicated in Table I. The order of the metals in the electrochemical series
not only depends on the nature of the elements themselves, but also on the chemical
composition of the solution in which they are placed, the degree of concentration,
and its temperature. In an acid solution, for example, the following six metals
have this order : zinc, iron, lead, copper, silver, antimony ; and in 12 5 per cent,
potassium cyanide solution the order zinc, copper, silver,
Table I ^Electro- antimony, lead, iron. When silver and potassium cyanide
CHEMICAL Series replace zinc and sulphxiric acid in the Daniell cell, a strong
OF THE Elements, current is developed and copper is precipitated ; copper is
usually supposed to precipitate silver, but silver here precipi-
tates copper. These variations promised to obscure the use
of the electrochemical series for predicting the course of
chemical reactions. Again, tin precipitates lead from its
solution in acetic acid, and lead precipitates tin from its
solution in nitric acid. Calcium iodide at about 740° is
reduced by metallic sodium, but the reverse action occurs
above 800° for sodium is displaced by calcium ; hence sodium
is more electropositive than calcium at low temperatures, and
less 80 at high temperatures. Similar phenomena occur with
potassium and strontium. In many cases the displacement is
so complete that the reaction is employed in quantitative
analysis. The further apart any two elements are in the
series : (1) The greater the electromotive force of the currents
generated when the two elements are used as plates in a voltaic
couple ; (2) The greater the amount of heat liberated when
the displacement occurs, e.g. when the zinc precipitates silver
more heat is evolved than when it precipitates tin. (3)
Similar remarks apply, mutatis mutandis, to the speed of
precipitation. (4) The greater the amount of heat or electrical
energy required for the decomposition of their compounds ;
(5) The greater their chemical affinity for one another.
A similar table would be obtained if the elements were
arranged in the order of their chemical activity. Thus,
(1) The earlier members on the list oxidize or rust on exposure
to the air. (2) Oxides of the metals succeeding manganese
are reduced to metals when heated in a stream of hydrogen,
while the metals which precede manganese, under the same
conditions, may be reduced to lower oxides, but not to the
metallic condition. (3) The oxides of the metals ranging
from mercury to osmium may be decomposed into their
elements by simply heating them to a comparatively low
temperature. (4) The metals preceding hydrogen on the
list can give hydrogen when treated with acids, although
secondary actions may simultaneously lead to the formation
of some product other than hydrogen. The metals succeeding
hydrogen do not usually displace hydrogen from the acids.
(5) With the possible exception of tin and lead (metals close
to hydrogen) the freed elements are rarely, if ever, found
in nature excepting possibly in meteorites. This arises from
the fact that natural waters containing carbonic and other acids in solution
attack these metals ; consequently, even if these elements were produced by
subterranean agents — volcanic or otherwise — ^they must eventually succumb to
attack by natural waters.
It will be noticed that the series refers only to the action of the free elements,
and it has no direct reference to the mutual action of chemical compounds of
Caesium \
Rubidium
Potassium
Sodium
Lithium
Barium
Strontium
Calcium
Magnesium
Aluminium
Chromium
Manganese
Zinc
Cadmium
Iron
".
Cobalt
\'^
Nickel
(^
Tin
^
Lead
3
Hydrogen
Antimony
Bismuth
Arsenic
Copper
Mercury
SHver
Palladium
Platinum
Gold
Iridium
Rhodium
Osmium /
Silicon '
Carbon
Boron
^
Nitrogen
!^
Selenium
3
Phosphorus
Sulphur
Iodine
§
Bromine
^
Chlorine
-
Oxygen
Fluorine >
ELECTRICAL ENERGY 1015
the elements upon one another. The order of the elements in the electro-
chemical series depends to some extent upon the temperature as well as on the
nature and concentration of the electrolyte. For exami)le, zinc and copper
behave in what appears to be an abnormal manner in the presence of potassium
cyanide. Thus copper and iron will precipitate zinc from potassium zinc cyanide,
whereas zinc will precipitate copper from copper sulphate ; and iron from neutral
ferrous sulphate. Again, silver will displace hydrogen from aqueous hydriodic
acid ; copper will precipitate nickel from sodium nickel chloride ; and platinum
will liberate hydrogen from aqueous solutions of potassium cyanide. R. Abegg and
G. Bodliinder i have developed H. von Helmholtz's assumption that the ions hold
their charges with different degrees of tenacity. The ions — K*, Na", NO 3', CI', etc.
— which hold their charges very tenaciously, are called strong ions ; and ions — Hg",
Ag', OH', Cy', etc. — which readily lose their charge, are called weak ions. The
degree of tenacity with which the ions of an element hold their charges has been
called the electro-affinity of the element. The electro-affinities of the elements
are roughly measured as decomposition voltages. Ions with strong electro-
affinity are difficult to prepare in a free state, and conversely. If an element with
a strong electro-affinity comes in contact with the ion of an element with a
weak electro-affinity, the charge on the latter passes over to the former. Thus
zinc has a stronger electro-affinity than copper, and, in consequence, as indicated
above, zinc will precipitate copper from solutions of its salts : Zn+Cu*'=Zn*'-f Cu.
Zinc also has a stronger electro-affinity than hydrogen, and consequently zinc
dissolves in dilute acids with the evolution of hydrogen: Zn-f2H'=Zn"-f H2.
Similarly, chlorine has a stronger electro-affinity than bromine, and bromine a
stronger electro-affinity than iodine. In consequence, chlorine will displace bromine
from aqueous solutions of the bromides : Cl2-f 2K'-[-2Br'r=^2K*+2Cr+Br2 ; and
bromine will displace iodine from the iodides : Br2-[-2K'+2I'v=^2K-f 2Br'-f I2.
The attempt of R. Abegg and G. Bodlander to show that certain properties and
reactions of the inorganic salts are directly dependent upon the electro-affinities of
the respective + and — ions. These properties are especially (1) solubility — where
it is supposed that the greater the electro-affinity of the -f- and — ions, the greater
the solubility — e.g. the electro-affinity of sodium ion in sodium salts is usually
very great and the salts are usually very soluble in water ; whereas with silver salts,
the silver ion has a weak electro-affinity and the salts are sparingly soluble. If one
of the ions has a high and the other a low electro-affinity the salt is usually soluble,
e.g. silver nitrate. (2) Tendency to form complex positive or negative ions — if
the neutral compound of the complex ion has low electro-affinities, the tendency to
form complex salts will be great — e.g. in potassium ferricyanide KsEeCyg, the
complex ion FeCye'" is composed of single ions 3Cy' and the neutral component
FeCy3, where the electro-affinities are weak. A further examination of these hypo-
theses by J. Locke shows that the relations deduced by R. Abegg and G. Bodlander
are not generally applicable.
References.
1 R. Abegg and G. Bodlander, Zeit. anorg. Chem.y 20. 453, 1899 ; J. Locke, Amer. Chem.
Jo2irn., 27. 105, 1902.
§ 3. Solution Pressure — Contact Differences of Potential
Whenever there is free energy present, or whenever there is potential energy seeking to
become free when chemical afiinities come into play in the presence of differential molecular
structures or conditions, this energy takes the form of electrical potential. — J. T. S PR ague
(1892).
If a spherical globule of mercury at the bottom of a watch glass — a, Fig. 1 — be
electrified at either the positive or negative pole of an electric machine, the globule
1016
INORGANIC AND THEORETICAL CHEMISTRY
Not dectrified tJectrified
Fig. 1.— H. Her-
wig's Experi-
ment.
^;> '-c"*^
r
Aj.
■n n-
B
1
1
1 -^l ^1 0- 1
flattens — h. Fig. 1 — showing that the surface tension of the metal has diminished,
and that changes in the surface tension of mercury accompany variations in its
electrical condition. The greater the electric charge impressed on the metal, the
smaller the surface tension. If g denotes the surface tension ; E, the potential
difference between the metal and electrolyte ; and e, the quantity of positive elec-
tricity per unit surface of metal, G. Lippmann,i H. von Helmholtz,
a b and W. Ostwald find dGldE=e.
In P. Lippmann's experiment (1873), two glass (J -tubes, A and
J5, Fig. 2, each with one leg drawn to a capillary bore, were partially
filled with mercury, and partially immersed in a beaker of dilute
sulphuric acid (1:6) as shown diagrammatically in Fig. 2. The
mercury rose to a higher level in the wider limb than in the
capillary, and the finer the capillary the greater the difference of level in the two
limbs — care must be taken to displace any bubbles of air in the capillary which would
prevent the acid coming in contact with the mercury. Connect the mercury in the
wider limbs by means of a source of electrical energy at a pressure not exceeding
0*5 volt. The meniscus of the mercury a through which the current enters the elec-
trolyte— dilute sulphuric acid — will rise, and the meniscus h,
.^k>w through which the current leaves the electrolyte, will be de-
pressed. The arrows show the direction of the current. If
the meniscus in one capillary falls, the surface tension of the
mercury must be increasing, and conversely, if the meniscus
rises, the surface tension is diminishing. Hence, when the
electrical connections were made the electrical potential of
a was increased, and that of h decreased. If the potential
difference between A and B were gradually increased from
0"5 volt, the meniscus at a would continue ascending, and
that at h would continue descending until a potential differ-
2.— P. Lippmann's ence of 0'93 volt is attained — after that, any further increase
Experiment. jn the potential difference between a and h causes the mercury
in h to rise. The explanation is that mercury in contact
with dilute sulphuric acid normally acquires a potential difference of 0"93 volt, which,
for convenience, is called the contact difference of potential, or the electrode
potential of the mercury. Under the influence of an increasing potential difference,
the positive charge is gradually augmented at a ; and the increasing negative
charge at h reduces the total charge on h by neutralizing the normal positive charge
already there. When the difference of potential applied at h reaches
0*93 volt, the charge at h will be zero, and the mercury will have
its maximum surface tension and h reaches its lowest level in the
capillary. Any further increase in the difference of potential between
a and h will cause h to be charged negatively and the mercury at h
will begin to rise in the capillary. The mercury in both capillaries
will continue rising until the difference of potential reaches 2 volts.
The acid then begins to decompose.
If the source of the electrical energy be removed and the two
wires in A and B be joined, the mercury in each capillary will return
jTjQ^ 3^ to its former level. This experiment illustrates the principle of the
Helmholtz' 8 so-called capillary electrometer used in various forms for measuring
Double small differences of potential in the laboratory.
^^y^^- It is supposed that if the metal near its surface of separation
from the dilute acid be positively charged — ^the surface of separa-
tion must act as an insulator or dielectric — and an equivalent negative charge
will be induced in the acid — as represented diagrammatically in Fig. 3. The two
layers of electricity of opposite sign are called Helmholtz's double layer— after
H. von Helmholtz's investigations on this subject in 1879. Had dilute hydro-
chloric acid been used in place of dilute sulphuric acid, the maximum
Fig.
ELECTRICAL ENERGY
1017
depression in the capillary h would have occurred with a difference of potential
of 0-56 volt.
The contact difference of potential between the different metals in contact with
normal solutions of their salts are indicated in Table II, due to N. T. M. Wilsmore.
Table II.' — Contact Potentials of the Elements.
Volts.
Volts.
Volts.
K
(+2-92)
Fe
+0-063
Hg .
-1-027
Na
( + 2-54)
Tl
+0-045
Ag
-1-048
Ba
( + 2-54)
Co
-0-046
Pd
< -1-066
Sr
Ca
Mg
Mg
(+2-49)
( + 2-28)
(+2-26)
+ 1-214?
Ni
Sn
Pb
H
-0-049
< -0-085
-0-129
-0-277
Pt
Au
F
CI
< -1-140
< -1-356
(-2-24)
-1-694
Al
+0-999 ?
Cu
-0-606
Br
-1-270
Mn
+0-798
As
< -0-570
I
-1-797
Zn
+0-493
Bi
< -0-668
O
-1-396?
Cd
+ 0-143
Sb
< -0-743
Fig. 4.
A common method used in measuring contact difference of potential, or the relative
e.m.f.- — -electromotive force — of metals in contact with different solutions, is to make a cell
with two electrodes each dipping in a separate solution, (i) One electrode consists of the metal
to be tested held by platimma-tipped forceps electrically connected with a galvanometer,
and dipping in the required solution, (ii) The other is the so-called normal electrode. In
Ostwald's non-polarizable normal electrode (1) mercury is electrically connected with the
galvanometer by means of a glass-coated platinum wire, (2) the
surface of the mercury is connected with a layer of merciu-ous
chloride about 5 cm. thick, and (3) a solution containing a normal
solution of potassiima chloride. This gives an electromotive
force of —0-560 volt, or a positive current tends to flow through
the solution to the mercury which becomes positively electrified,
while the solution itself becomes negatively electrified. When
the solutions about the two electrodes are in contact, the slight
e.m.f. due to the contact of the two liquids is neglected and the
e.m.f. of the whole combination is the algebraic sum of the e.m.f. 's
action at the two electrodes. The e.m.f. of the given metal in
contact with the given solution is obtained by subtracting 0-560 volt from the observed
e.m.f. of the combination. The e.m.f. of the combination is observed in terms of the
deflection of a calibrated galvanometer, or by the compensation or zero method of
J. C. Poggendorff described in laboratory manuals.
W. Nernst (1889) 2 has carried the idea of contact difference of potential still
further. He assumes that if a metal rod be immersed in a liquid, it tends to dissolve.
The supposed tendency of a metal to dissolve in any liquid is called the solution
pressure of the metal. The supposed action is likened to the tendency of liquid
to vaporize as indicated by the vapour pressure of the liquid at any given temperature.
Still further, just as a liquid continues to evaporate at a free surface until the number
of molecules leaving the surface of the liquid in any given time is equal to the number
of molecules returning to the liquid, so W. Nernst suggests that a metal, when placed
in contact with water or other liquid, tends to send charged ions into the solution
and itself to assume an equivalent charge of opposite sign. A force has been invented
to drive the ions into the solution, and it is called the electrolytic solution pressure.
The force is supposed to vary with the nature of the metal, the solution, and the
temperature. The solution pressure must be greatest with the metals at the caesium
end, and least with the metals at the osmium end of the electrochemical series.
Conversely, the tendency of positive metal ions in solution to reprecipitate on the
negative electrode must be least at the caesium end of the series and greatest at the
osmium end. The ionic hypothesis assumes that this back or deposition pressure
represents the osmotic pressure of the ions. The ionization of the metal, so to speak,
1018 INORGANIC AND THEORETICAL CHEMISTRY
is supposed to continue until the concentration of the metallic ions in the liquid
has attained a certain value when a state of equilibrium ensues. The number of
ions passing into the solution is then equal to the number reprecipitated on the surface
of the metal. Direct proof of the presence of iron ions in purified water, which has
been in contact with the highly purified iron, is wanting. The evidence is indirect,
or rather hypothetical.
When zinc is immersed in dilute hydrochloric acid, the supposed H'-ions which
come in contact with the zinc plate lose their charge, and positively charged zinc
ions pass into solution. If a stick of metallic zinc be dipped in a saturated solution-
of zinc sulphate, the solution and deposition or osmotic pressures are balanced, and
no action occurs ; but if a stick of metallic zinc be placed in a dilute, say normal,
solution of zinc sulphate, the solution pressure is greater than the deposition or
osmotic pressure, and positively charged zinc ions pass from the zinc rod into the
solution. In consequence, the zinc acquires a negative charge, and the solution
a positive charge, in agreement with the fact that zinc usually acquires a negative
charge when immersed in a solution of its own salt. Similar remarks apply to
aluminium, iron, etc. Conversely, if the solution pressure be less than the deposi-
tion or osmotic pressure of the ions, as appears to be the case with a stick of metallic
copper immersed in a solution of copper sulphate, copper ions will be deposited on
the metal, and the solution will acquire a negative charge while the metal acquires
a positive charge. This also appears to be the case with the metals of silver, mercury,
etc.
Let P denote the electrolytic solution pressure of the metal, and f the osmotic
pressure of the metallic ions in the solution. It is presumed that the osmotic pressure
of the metallic ions will oppose the tendency of the metallic ions to pass into solution.
When the opposing forces are balanced, no action will take place^as will occur with
zinc in a saturated solution of zinc sulphate.
In a normal solution of zinc sulphate, if P be greater than f, bivalent zinc ions will
pass from the zinc rod into the solution, the solution will become positively, and the
metal negatively charged. The attraction of opposite charges will cause positive
ions to collect about the surface separating metal and solution, and thus form a
Helmholtz's double layer. When the osmotic pressure j) of the metallic ions in the
solution has been augmented by the separation of ions from the metal itself until
it is equal to the electrolytic solution pressure, the opposing forces will be balanced ;
there will be no further increase in the number of metal ions in the solution, and a
definite electromotive force will be established — metal negative, solution positive.
Thus zinc, aluminium, and iron are generally negative when immersed in solutions
of their own salts. If the electrolytic solution pressure P be less than the osmotic
pressure f such as occurs when copper is placed in a solution of copper sulphate,
positively charged copper ions will separate from the solution and be precipitated on
the metal and the rod of metal will become positively, and the solution negatively
charged. An Helmholtz's double layer will be formed as the negative ions remaining
in the solution collect about the surface separating solution and metal. Equilibrium
will be established when P=^, and an electromotive force will appear at the boundary
double layer in a reverse direction to that established when P was greater than f.
Gold, silver, mercury, and copper are illustrations of metals usually positive when
immersed in solutions of their own salts.
The electrical effect, or the contact difference of potential, produced when the
different metals are immersed in a normal solution of their sulphates has been
measured. B. Neumann's results are indicated in Table III. The number +0*524
opposite zinc means that if metallic zinc be immersed in a normal solution of zinc
sulphate, the solution will acquire a positive charge, and the metal a negative charge ;
and the difference of potential between the solution and the metal will be 0-524 volt.
With metallic copper and a solution of copper sulphate, the solution will be charged
negatively, and the copper positively, such that the difference of potential between
the solution and the metal will be 0515 volt.
ELECTRICAL ENERGY
Table III.
1019
Metals.
Sulphate.
Chloride.
Nitrate.
Acetate.
Magnesium .
+ 1-239
+ 1-231
+ 1-060
+ 1-240
Aluminium
+ 1-040
+ 1-015
+0-775
—
Manganese
+0-816
+0-824
+0-560
■ —
Zinc
+ 0-524
+0-503
+0-473
+0-522
Cadmium
+ 0-162
+0-174
+ 0-122
—
Thallium
+0-114
+0-151
+0-112
—
Iron
+0-093
+0-087
. — .
—
Cobalt
-0019
-0-015
-0-078
-0-004
Nickel
-0-022
-0-020
-0-060
. — ,
Lead
, —
-0-095
-0-115
-0-079
Hydrogen
-0-238
-0-249
—
-0-150
Bismuth
-0-490
-0-315
-0-500
—
Arsenic
— ,
-0-560
. —
_-
Antimony
—
-0-376
—
—
Tin
— .
-0-085
■ — •
■ — ■
Copper
-0-515
. .
-0-615
-0-580
Mercury
-.0-980
• —
-1-028
—
SHver .
-0-974
. — .
— 1-055
-0-991
Palladium
^—
-1-066
. — .
—
Platinum
_
-1-140
. —
— .
Gold .
■ — •
-1-356
• — •
If a normal solution of copper sulphate be separated by a porous partition,
Fig. 5, from a normal solution of zinc sulphate, and if a rod of copper immersed in
the copper sulphate be connected by a wire with a rod of zinc immersed in the zinc
sulphate (Fig. 5), the zinc pole on the right of the diagram acquires a negative
charge ex hypothesi on account of the departure of positively charged ions from its
surface, and the copper pole on the left acquires a positive charge on account of the
departure of negatively charged copper ions from its surface. In consequence, an
electrical current will flow through the connecting wire from the positively to the
negatively charged pole and pass in the converse direction through the liquid. This
ZiaePlate.
Porous pot contains ZnSO^ aq.
Illni — Copper Plafe.
- Outerjar contains CuSO^ a^.
Fig. 6.— Darnell's Cell.
Fig. 5.— DanieU's Cell
(Diagrammatic ) .
action continues until all the zinc is dissolved or all the copper precipitated. The
relative solution pressures of the two metals decide the magnitude of the resultant
electromotive force of the current, and this is the difEerence of the two effects. The
resultant electromotive force for the zinc : copper couple just described is -f 0*524:
— (— 0-515)=0-524+0-515==l'039 volts. The combination just described represents
the so-called DanieU's cell (1836),^ which resembled in principle the unpractical
cell described by E. Becquerel in 1829. In reality, J. F. DanieU's cell contains the
zinc rod with the sulphuric acid or zinc sulphate solution in a porous pot, and the
copper plate with the copper sulphate solution in the surrounding jar, as illustrated
by the drawing of an uncharged cell in Fig. 6. The reactions are symbolized
1020 INORGANIC AND THEORETICAL CHEMISTRY
molecularly in the following manner : At the positive zinc plate, ZnH-H2S04=ZnS04
+H2. Instead of the hydrogen atoms being liberated on the negative copper plate,
they are exchanged for copper in the copper sulphate solution : H2+CUSO4
=H2S04+Cu, and the copper is deposited on the copper plate. The net result of
this round of changes is that the copper plate grows while the zinc plate lessens ;
and zinc sulphate increases, copper sulphate decreases. If a solution of zinc
sulphate is used in Daniell's cell, the copper is exchanged for the zinc : Zn+CuS04
=ZnS04+Cu. There are many other modifications of Daniell's cell, and numerous
other types of cell with different " poles " and different solutions.
The quantity of electricity (coulombs) produced by Daniell's cell depends upon
the amount of zinc consumed (Faraday's law) ; and the rate at which electricity
is developed (amperes) depends upon the rate at which the zinc is consumed in the
cell. The difference of potential or the electrical pressure cannot exceed 1*039 volts
for the given solutions. The product of the number of coulombs into the number
of volts gives the amount of electrical energy expressed in joules. If the term
ampere be employed to represent a current equivalent to one coulomb per second,
the product of the number of amperes into the number of volts gives the amount of
electrical energy produced in one second by the cell, expressed in watts — a joule of
electrical energy per second represents one watt.
Examples.- — (1) The electrolysis of 36*5 grams of hydroehloric acid requires 96,540
coulombs of electricity at 1-31 volts. Hence the electrical energy needed for this work is
96,540x1*31 = 126,567 miits, or, defining a joule as the unit of electrical energy consumed
by a current of one coulomb working against a resistance of one ohm (joules= volts
X coulombs), the electric energy needed to decompose 36*5 grams of hydrochloric acid is
126,567 joules.
(2) A current of 100 volts and 1-5 amps, passes through a system, hence 100 x 1*5 = 150
watts of energy are consumed per second.
The relation between the electromotive force and the osmotic pressure.—
The qualitative sketch of Nernst's hypothesis can be described in a quantitative
form. If a substance with an electrolytic solution pressure P be converted into
ions with an equivalent osmotic pressure P^, the opposing forces are balanced, and
no work is done. On the other hand, if a metal with an electrolytic solution pressure
P be converted into ions having an osmotic pressure ^, the maximum work which
can be performed during the transfer is analogous with the work performed when
ions are transferred from an osmotic pressure P to an osmotic pressure jp. If one
gram- molecule be involved, the osmotic work will be RT log (P/^). One gram-ion
carrying nC equivalents of electricity (n denotes valency) at a potential of E volts,
can perform nEG units of electrical work. Assuming that the osmotic work and
electrical work are equivalent, nEC=RT log {P/p), when R=2 cals., one volt-
coulomb=0*24 cal. Hence, in place of nEC we can write w£'x 96500x0*24:
=2T log (P/p), and remembering that natui-al logarithms are converted into
common logarithms by multiplying by 2*3026,
^=0-000198 logio volts
n ° p
This equation represents the electromotive force or potential difference developed,
at the absolute temperature T, when an ?^-valent metal with an electrolytic solution
pressure P is immersed in a solution in which the corresponding ion has an osmotic
pressure p.
Examples. — (1) Let one zinc rod dip in a normal and another in a decinormal solution
of zinc chloride. Fig. 7, assume that ionization is complete in both solutions. What is
the resulting electromotive force of the coll ? Ansr, ^=0-0287 volt.
(2) What is the electrolytic solution pressure of zinc in normal zinc sulphate solutions
at 17° when the contact difference of potential is 0*524 volts ? Here n~2 ; £^=0-524;
p denotes the osmotic pressure of zinc ions when a gram-molecule of the salt is
completely ionized in a litre of solution, P = 22-4 atm. Hence, 0-524=0'000198 x^
X 290(log P-log 22-4), or P=2-7 x 10'» atm.
ELECTRICAL ENERGY
1021
The electrolytic solution pressures cannot be measured directly, but they can be
computed from the observed differences of potential as in the preceding example.
The computed values for magnesium, zinc, iron, hydrogen, and silver expressed in
atmospheres, are :
Magnesium.
Zinc.
Iron.
Hydrogen.
Silver.
Palladium.
10x10"
2-7xlOi»
1-2x10*
9-9 X 10-*
2-3 X 10-"
1-5x10-"
It is difficult to understand what these numbers mean. The number for zinc appears
extraordinarily large; it is nearly equal to a weight of 180000,000000,000000
tons or 1-8x1017— the earth itself is estimated at 6000,000000,000000,000000, or
6 X 1021 tons. Nothing like these pressures have been directly observed ; they are
said to represent " the striving of the metal to overcome the opposing osmotic
pressure of the metal ions already in solution in order that the atom of the metal may
be ionized." We cannot employ the tabulated electrolytic solution pressures to
calculate contact differences of potential E along with observed values of jp and n
in the expression EC=RT log (P/j)), and compare the results with observed values
of E in order to show the validity of the formula, because that would be reasoning
in a circle.
Concentration cells. — ^Although the difference of potential of a given cell, say a
Darnell's cell, is not affected by variations in the size or shape of the poles, or upon
the quantity of liquid in the cells, the difference of
potential is altered by changing the concentration of
the solutions. In general the difference of potential
between a metal and a solution of one of its salts is
greater with increasing dilution. A tenth normal solu-
tion of zinc sulphate, for instance, will give a difference
of potential of 0*551 volt, whereas with a normal solu-
tion a potential difference of 0'524: volt is obtained as
indicated above.
If two rods of zinc be separately placed in a. N-
solution of. zinc sulphate, the difference of potential in
both " tends " to drive an electric current from the
metal to the solution with a pressure of 0'524 volt.. If Uilute Concentrated
both rods be joined by a wire, no electric current will Fig. ?.• — Concentration Cell,
flow because the two equal forces are oppositely directed.
On the other hand, if the zinc rods be dipped in solutions of a different concentration,
the two contact differences of potential will be different, and an electric current will
flow from the concentrated solution to the dilute solution outside the cell as indicated
in Fig. 7. Here a normal solution of zinc chloride is supposed to be placed in one
vessel, A. and a decinormal solution of zinc chloride in the other vessel, B. Zinc
rods connected by a copper wire and galvanometer are dipped into the solutions, as
illustrated in the diagram, and the two cells are connected by a syphon tube S.
The difference of potential of the zinc in the normal solution is +0*524, and in the
more dilute solution -}-0'551 volt. Hence an electric current tends to pass from the
metal to the dilute solution with a force of +0"551 volt, and from the metal to the
concentrated solution with a force of 0'524. The resultant pressure is therefore
0'551— 0*524:=0'027, and this represents the electromotive force of the combination.
Cells in which the electromotive force is generated by the difference potential of
two plates immersed in solutions of the same salt at different concentrations are
called concentration cells. The chemical action which occurs in the two cells tends
to bring the two solutions to the same concentration.
The action is made clear by the following experiment : A layer of a concentrated solution
of stannous chloride in hydrochloric acid, about 10 cm. deep, is placed at the bottom of a
cylinder, and above this a layer of a dilute solution. A rod of metallic tin is fixed through
a hole in the cork so that it is suspended axially in the liquid in the cylinder. The rod of
tin thus represents both electrodes and connecting wire of a concentration cell. Tin is
1022 INORGANIC AND THEORETICAL CHEMISTRY
dissolved by the more dilute solution, and precipitated from the more concentrated solution.
After the system has stood a couple of days, the rod of tin near the surface of the more
dilute solution will be reduced in thickness. Cadmium and cadmium chloride can be
used ; zinc does not work so well ; nickel gives negative results, probably because the
surface of the metal becomes polarized ; antimony and bismuth also give negative results,
probably because they are not attacked by dilute hydrochloric acid.*
There is another interesting feature about a concentration cell. If an external
electromotive force be applied so as to force an electric current to pass in a reverse
direction to that which the combination normally furnishes when it is employed in
a voltaic cell, the chemical actions will be reversed, and the difference in the concen-
tration of the two solutions will be augmented. Such combinations are called
reversible cells in contradistinction to irreversible cells in which the original
condition cannot be restored by sending a current through the cell in a reverse
direction to the current normally delivered by the cell. The Zn|H2S04aq|Pt cell
is an irreversible cell ; Darnell's cell, Fig. 6, and the concentration cell, Fig. 7, are
reversible cells. If two metal rods are arranged so that one dips in a solution of
concentration Oi and the other rod in another solution of concentration C2 ,' and if
the two solutions are separated by a porous partition and the rods be connected
electrically, the electromotive force of the combination will be :
E=^=r( log log - ) ; or E=-=r log —
where pi and p2 respectively denote the osmotic pressures of the solutions of concen-
trations Ci and C2 respectively and P is the solution pressure of the metal. If
the metal be nickel, and the solutions be nickel nitrate, i?=8'31 electrical units ;
£•0=96,540 coulombs ; n=2 ; and at 18°, T=273-f 18=291 ; if, further, common
logarithms are used, then, remembering that the osmotic pressures of dilute solutions
are proportional to their concentrations,
£'=0-0288 log ^
If one solution contains 0*1 gram-molecule of nickel nitrate and the other, 0*05
gram-molecule, the calculated ekctromotive force of the combination on the assump-
tion that ionization is complete is £^==0*0288 X0-301=0'0087 volt; the observed
value is 0"010 volt.
G. Meyer 5 showed that measurements of the e.m.f . of the combination
Concentrated amalgam | Solution of a salt of the solute metal | Dilute amalgam
enabled the molecular weight of the metal dissolved in the mercury to be computed
when there is no combination between the mercury and the metal. Expressing the
preceding equation in the form :
^=0-002 - log ^
n ° C2
divide n by the number of electrical units which go with an atom of the metal in
question and the number of atoms in a molecule, and hence the molecular weight of
the metal can be computed.
Mercury precipitates silver from a solution of silver nitrate, and silver precipitates
mercury from a solution of mercurous nitrate. G. M. Smith (1904) has also found
the following pairs of elements can be reciprocally precipitated : Zn— Cu ; Ca— Cu ;
Fe— Hg ; Fe— Ag ; Hg— Ag ; Hg— Pt ; Hg— Au ; Ag— Au. Again, potassium
from a solution of potassium chloride can replace sodium from its amalgam, and
conversely sodium can displace potassium from its amalgam ; potassium and
barium, and sodium and barium are reciprocally replaceable in spite of the
fact that potassium is more electropositive than sodium, and sodium more than
ELECTRICAL ENERGY 1023
barium. M. Berthelot called the phenomenon an anomaly, and supposed it to
be a consequence of the greater loss of energy suffered by potassium in comparison
with sodium when the respective amalgams are formed. According to Nernst's
theory, when one metal Mi is immersed in the salt of another metal M2, if M2 is to
be precipitated by Mi,
^1 ^1 »*2 V2 ^ Vl ^ V'
2
where the subscripts refer to the corresponding metals ; Pj, F^ are the electrolytic
solution pressures ; ^1, p2 *^® osmotic pressures of the univalent ions ; and Wj, n^
the valencies of the metals in question. Hence the conditions which are favourable
for the precipitation of M^ are (i) a high osmotic pressure, that is, a high concentration
of the ions of the second metal ; and (ii) a high solution pressure, that is, a low
osmotic coilnter pressure or low concentration of the ions of this metal. It is there-
fore to be anticipated that two metals might be reciprocally replaceable — all depends
on the relative magnitudes of the above terms under different conditions.
There is a difference of potential at the surface of contact between two solutionst
If two solutions of, say, hydrochloric acid of different concentration be in contact,
the more concentrated solution will diffuse into the other ; and, according to the ion
theory, the hydrogen and chlorine ions can travel independently at different rates —
the positively charged hydrogen ion being the faster will be in the van, the negatively
charged chlorine ions will lag behind, and accordingly a difference of potential will
be established. The electrostatic attractions set up by the two sets of oppositely
charged ions will prevent their separating very far, but it will be sufficient to cause a
layer of positively charged ions to accumulate near the surface of the more dilute
solution, and a layer of negatively charged ions near the surface of the more concen-
trated solution. If a salt had beeii chosen such that the negatively charged ions
moved the faster, the charges on the dilute and concentrated solution surfaces would
have been reversed, for the more dilute solution must be charged by the more quickly
moving ions.
Let v and v' respectively denote the relative velocities of the ions — assumed to
be univalent. Suppose that the cation moves the faster, and that a unit charge of
electricity — 96,450 coulombs — is carried from the dilute to the concentrated solu-
tion. The total current, v'-\-v\ is shared between both ions, so that ^7(^*4"^') repre-
sents the cation's share, and v'l{v'-\-v') the anion's share; oTV'l(v'-\-v')g'£B.'m-ioTi% of
the cation will have gone from the concentrated to the dilute solution, and v*l(v'-\-v')
gram-ions of the anion will have gone in the opposite direction. If 'p^ denotes the
osmotic pressure of the cations in the dilute solution it will also represent the
osmotic pressure of the anion in the same solution ; similarly, let ^2 denote the osmotic
pressure of the anions, and of necessity, also, of the cations in the dilute solution.
This means that v'l(v'-\-v') gram-ions of the cation pass from the higher osmotic
pressure f^ to ^^^ lower osmotic pressure 'pi. Hence the energy converted into
electric energy by the cations, and the maximum work done upon the anions moving
in the opposite direction from the higher to the lower pressure, will be respectively
Work done by cations = — . — JtT log —
Work done on anions = — , — ,RT lo2 —
V'-{-V ° p
The available electrical energy EC obtained will be the difference between these
two quantities. This difference simplifies, as before, into
£'=0000198T^^^, log ^'rvolts
1024 INOEGANIC AND THEORETICAL CHEMISTRY
More generally, (i) if the cation has a valency m and the anion a valency n ; (ii) if
the ratio of the concentrations of the two solutions C1IC2 be substituted, as before,
for the ratio of the osmotic pressure ^1/^2 J ^.nd (iii) if the degree of ionization of
the solution be x,
0-000198 XT ^-^'
-^^ ;;: • «,. 1 j log ?r volts
X v'-\-v C-i
an expression which represents the contact difference of potential of two concen-
trations of the same solute at the absolute temperature T. This expression
represents the contact difference of potential between two concentrations of an
electrolyte. Obviously, if the velocities of the two ions were equal, v'=v' ; and if
the solutions have the same concentrations ^1=^92 ; in either case the electro-
motive force would be zero.
Example. — Calculate the electromotive force of the cell Aglj-J^iV-AgNOgl
^^yAT-AgNOglAg at 17°, when the degree of ionization of the solution is 0'935 and the
velocity of the cation is 52 and of the anion 58. There are three contacts to consider, and
Ihe electromotive force E will be the algebraic sum of the contact differences of potential
of the two silver electrodes with their respective solutions, and of the two solutions with
one another.
J, 0-000198T, C^.v'~v' Cg ,.
Ez=— _— log — ^H ; — , log 77^ volt
Here (78 = 100, n=m = l, ^- = 52, «;' = 58, ^ = 290, it follows that JS7=:{0-000198 x290
X 0-935 X 2 X 58 X 2-3026)/(52 + 58) =0-058 volt. The observed value is 0-055 volt.
W. Nernst's method of calculating the e.m.f., E, of a concentration cell from
the transport numbers v and v' and the ratio of the concentrations :
^=0-0002-^, -^log^
v'-\-v n ° Gi
is approximately correct for dilute solutions, but for concentrated solutions, and
solutions in non-aqueous solvents of low-ionizing power, there are discrepancies
attributed to incomplete ionization ; and in that case, the ratio of the ionic concen-
trations calculated from the conductivities are substituted for the ratio of the concen-
trations. The prediction of the e.m.f. of a cell is dependent on the applicability of
the ionic hypothesis to the solutions. Anomalies are harmonized by introducing
hypotheses respecting the association of solute or solvent, or the combination of
solvent and solute. H. von Helmholtz (1878) calculated the e.m.f. of a concentration
cell from the vapour pressures of the two solutions, and the result is the same as that
obtained by W. Nernst's method when the liquid : liquid potential difference is
negligible.
L. Hermann has shown that when a current is passed from a dilute solution of a
salt to one more concentrated, acid is liberated at the boundary layer, and if the
current is sent in the opposite direction an alkali is set free at the same place. G. S.
Walpole 6 showed that with all the neutral salts he examined, acid is always liberated
when the current passes from the more dilute to the more concentrated solution,
and an alkali when the current passes from the concentrated to the more dilute
solution. The quantities of acid or alkali liberated bear no relation to the quantity
of electricity passing through the circuit ; but when the other conditions are the
same, the same quantity of acid or alkali is liberated whatever neutral salt be used.
G. S. Walpole calculated the effect which would be produced on the assumption
that the phenomenon is due to the difference in the velocities of the hydrogen and
hydroxyl ions on opposite sides of the boundary layer, and obtained numbers in
agreement with the olDserved data.
When a substance in solution passes from a lower to a higher state of oxidation
ELECTKICAL ENERGY 1025
the change may be regarded as an increase in the number of electric charges on the
ions in question ; for instance, the oxidation of ferrous to ferric iron is represented,
Fe"->Fe'" ; and the oxidation of stannous to stannic tin, Sn"->Sn'"'. Conversely,
with the reverse changes, the reduction of stannic to stannous tin, Sn'"*"->Sn*',
or of ferric to ferrous iron, Fe"'-»Fe". The tendency of an ion to pass from one
state of oxidation to another can be expressed in terms of an electromotive force.
A cell is filled with unattackable electrodes, say, with platinum electrodes immersed
respectively in solutions of a ferric and a ferrous salt, and a current is passed from an
accumulator in the direction of the arrow: Pt+|FeCl2->FeCl3|Pt-, the ferrous
ions on the one side will be oxidized to ferric ions, and the ferric ions on the other side
will be reduced to ferrous ions. If the accumulator is now removed and the circuit
closed, ferric ions at one electrode will be reduced because they will give up positive
charges to the one electrode, and, as a result, the other electrode will be charged
positively and it will oxidize the ferrous ions in the vicinity to ferric ions. This
action will continue until the ratio of the ferrous and ferric ions about the two
electrodes is the same. If Ci denotes the concentration of the ferric ions and Cq
that of the ferrous ions, K^^Ci/CQ ; Nernst's equation assumes the form :
where Eq is a constant representing the observed potential difference at the electrodes
when the concentrations of the -ous and -ic ions are equal — say one gram-molecule
per litre — so that log (C'i/Oo)=log l=zero ; n is the difference in the valency of the
two ions — e.g. in the case of iron, n=l, and in the case of tin, n=2 ; the other symbols
have their usual meaning ; E is sometimes called the oxidation or reduction potential
of a solution because it can be taken as a measure of its oxidizing or reducing power.
Examples. — (1) The potential difference between solutions of thallic nitrate and thallium
were measured, and the mean value for Eq for thallium solutions was 1-191 volts. What is
the reduction potential T1'"->T1* for solutions containing 0'0505 gram-molecule of thallic
nitrate, T1(N03)3 and 0-00108 gram-molecule thallious nitrate TINO3 ? Here E=Eq
+ 0-029 log 46-7, or J^=- 1-233 volts.
(2) The observed value of Eq for solutions of uranyl and uranous sulphates is 0-615 at
27-5°, what is the electromotive force with solutions respectively containing 48-4 and 51*6
per cent, of uranyl and uranous sulphate ?
Sacrificial metals. — If metallic zinc dissolving in, say, dilute sulphuric acid, be
in contact with a piece of copper or platinum the rate of dissolution of the zinc
is augmented. The combination Zn : Pt is called a couple, and it really forms a
small galvanic cell with zinc, and, say, platinum electrodes connected together by
metallic contact. Much of the hydrogen is evolved from the surface of the platinum.
We have seen that any metal in the electrochemical series can be made one plate of
a cell against a metal lower down in the series. Zinc, for instance, can be made the
positive plate against a negative plate of iron, tin, lead, etc. ; and iron the positive
plate against a negative plate of tin, lead, etc. The further apart the elements in
the series, the greater the electromotive force of the combination. Tin-plate is
iron or steel coated with a thin layer of tin. If a little moisture be precipitated on
the surface in contact with both the iron and the tin, the moisture, with its dissolved
carbonic acid, dissolves the iron, and produces salts of iron ; these ultimately form rust
iq.v.). The iron is covered with a layer of tin to protect it from rust, but if there be
a flaw in the protecting surface of tin so as to expose the underlying iron, rusting
takes place more rapidly than if the iron had not been tinned at all. The tin remains
untarnished. Zinc is also used as a protecting layer over the surface of thin iron
plates — galvanized iron. The voltaic action developed when the protecting layer
is damaged is much less than when tin is used. These facts can be illustrated by
fitting up a cell like Fig. 8 with iron and tin plates, and another cell with iron and
zinc plates. Water saturated with carbon dioxide is used in both cells. A feeble
VOL. 1. 3 u
1026 INOKGANIC AND THEORETICAL CHEMISTRY
electric current will flow from the tin to the iron outside the cell in one case, and
from the iron to the zinc in the other as illustrated graphically in the adjoining
diagram. In the iron : tin cell, iron dissolves and rusting occurs ; while in the
iron : zinc cell, the zinc dissolves and no rusting occurs as long as the circuit is closed.
These results might also have been predicted from
our study of Table I.
An iron : lead cell behaves like an iron : tin cell.
Iron railings are often fixed in a bed of lead ; the iron
corrodes first and the lead remains intact. H. Davy
(1824) 7 once proposed to prevent the corrosion of
Fig. 8. the copper sheathing of ships by fixing pieces of
metallic zinc here and there on the sheathing. The
zinc was corroded and the copper preserved. When the zinc was all consumed,
the copper ceased to poison the barnacles, and the bottom fouled as if the wood
had not been sheeted with copper. In all these cases it has been fancifully said
that one metal is sacrificed to ensure the safety of the other ; and all the cases
quoted are examples of galvanic couples : Fe : Sn ; Zn : Fe ; Fe : Pb ; and Zn : Cu.
References.
1 G. Lippmami, Pogg. Ann., 14-9. 547, 1873 ; Ann. Chim. Phys., (5), 5. 494, 1875 ; (5), 12.
265, 1877 ; N. T. M. Wilsmore, Zeit. phys. Chem., 35. 291, 1900 ; N. T. M. Wilsmore and
W. Ostwald, ib., 36. 91, 1901 ; W. Palmaer, ib., 25. 265, 1898 ; 28. 275, 1899 ; 36. 665, 1901 ;
F. Paschen, Wied. Ann., 39. 43, 1890 ; 40. 47, 1890 ; 41. 42, 1890 ; F. Braun, ib., 41. 448, 1890 ;
H. Herwig, ib., 11. 661, 1880 ; A. Konig, ib., 16. 1, 1882 ; W. Ostwald, Lehrbuch der allgemeinen
Chemie, Leipzig, 2. i, 927, 1893 ; H. von Helmholtz, Phil. Mag., (5), 5. 348, 1878; Wied. Ann.,
7. 337, 1879 ; Physical Memoirs, 1. 1, 1891.
2 W. Nemst, Zeit. phys. Chem., 4. 129, 1889 ; 9. 1, 1892 ; R. S. Milner, Phil. Mag., (5), 49
417, 1900 ; 0. J. Lodge, ib., (5), 49. 614, 1900 ; R. A. Lehfeldt, ib., (5), 48. 430, 1899.
3 J. F. DanieU, Phil. Trans., 125. 117, 1836 ; E. Becquerel, Ann. Chim. Phys., (2), 41. 5, 1829.
« A. Ditte and R. Metzner, Compt. Rend., 115. 936, 1303, 1892 ; 117. 691, 1893 ; A. Ditte,
»6.,116. 1128, 1893.
6 G. Meyer, Zeit. phys. Chem., 7. 447, 1891 ; F. Haber, ib., 41. 399, 1902 ; V. von Turin, ib.,
5. 340, 1890 ; A. Scholler, Zeit. Elektrochem., 5. 259, 1898 ; G. McP. Smith, Journ. Phys. Chem.,
8. 208, 1904.
« G. S. Walpole, Proc. Roy. Soc, 91. A, 134, 1915 ; L. Hermann, Gott. Nadir., 326, 1887.
' H. Davy, Phil. Trans., 114. 151, 242, 1824 ; 115. 328, 1825.
§ 4. The Ionic Hypothesis and Chemical Reactions
The knowledge of nature as it is- — not as we imagine it to be — constitutes true science. — -
Paracelsus.
There are some fervid enthusiasts who claim that " all chemical reactions are
reactions between ions ; molecules as such do not react at all." This statement
is not quite in harmony with known facts. The same might be said of the assump-
tion that " chemical activity is proportional to the number of available ions."
L. Kahlenberg (1902),i J. L. Sammis (1906), and C. B. Gates (1911) have brought
forward a large number of exceptions to these statements that it appears necessary
to modify the hypothesis very materially before it can be accepted as an accurate
description of the facts. Some chemical reactions proceed very rapidly in solutions
which are considered to be virtually non-conductors of electricity, and which,
ex hypothesi, are almost free from ions. For instance, dry hydrogen chloride precipi'
tates metal chlorides from benzene solutions of the oleates of copper, cobalt, and
nickel ; dry hydrogen sulphide precipitates sulphides from benzene solutions of the
same salts and of arsenic chloride. All this in spite of the fact that these solutions
do not conduct electricity appreciably. Again, dry ammonia does not unite with
dry hydrogen chloride, but union does take place if a trace of non-conducting benzene
vapour be present. One metal can displace another from a non-conducting solution
ELECTRICAL ENERGY 1027
in a non-aqueous medium. Thus, metallic lead, zinc, tin, silver, iron, etc., will
precipitate metallic copper from solutions of various salts in carbon disulphide, carbon
tetrachloride, ether, alcohol, etc. Hence, L. Kahlenberg claimed that in spite of
the ionic hypothesis, chemical reactions do take place in non-conducting solutions,
and these reactions are similar in result and speed to those which occur in conducting
aqueous solutions.
H. C. Allen and H. P. Cady and H. 0. Lichten waiter have re-examined the experi-
mental work on this subject, and they claim that the benzene and toluene solutions
of the salts showed some conductivity which increased when dry hydrogen chloride
was introduced. The solutions of the salts also showed polarization which indicated
that they possessed properties similar to those of an ordinary electrolyte. It is
thus possible that these reactions may be due to ionization, although an alternative
explanation is suggested in the assumption that the reactions with hydrogen chloride
and the salts of the unsaturated acids take place in two stages : (1) the formation
of additive products with hydrogen chloride, and (2) the decomposition of the additive
products into metal chloride and organic acid.
G. Senter showed that when an electrolyte enters into chemical reactions, the
non-ionized molecules as well as the ions may simultaneously take part in the change,
and later, S. F. Acree and J. M. Johnson came to the same conclusion. It was also
assumed in the earlier form of the ionic theory that the H'-ions are catalytically
active, but this did not explain the catalytic activity of neutral salts. A. Lapworth
suggests that the difficulty could be overcome by assuming that the non-ionized
molecules of the acid are also catalytically active, and this hypothesis is now generally
accepted as a result of the work of H. Goldschmidt, G. Bredig and W. S. Millar,
H. C. S. Snethlage, S. Arrhenius, H. S. Taylor, H. M. Dawson and T. W. Crann, etc.2
These facts leave the uncomfortable feeling that the ionic hypothesis is rather
an encumbrance on the theory of chemical reactions ; and with slight modifications
of phraseology, A. L. Lavoisier's indictment of the phlogiston hypothesis could be
applied to the ionic theory of chemical action. The ionic hypothesis cannot ignore
these observations if it is to win a permanent place among the conquests of science.
As 0. D. Chwolson (1910) has emphasized that there is as yet no solidly established
theory of solutions which will take into account all the known phenomena, and
on which reliance can be safely placed. Hence, advocates of the ionic hypothesis
revert to the ordinary language of the molecular theory where the ionic hypothesis
fails.
References.
1 L. Kahlenberg, Journ. Amer. Chem. Soc, 25. 380, 1903 ; Chem. News, 83. 312, 1913 ; Journ.
Phys. Chem., 5. 339, 1901 ; 6. 1, 1902 ; J. L. Sammis, ih., 10. 593, 1906 ; C. B. Gates, ih., 15.
97, 1911 ; N. Dhan, Zeit. Elektrochem., 22. 245, 1916; H. E. Armstrong, Journ. Chem. Soc,
55. 78, 1896 ; G. Senter, ib., 91. 460, 1907 ; 95. 1827, 1908 ; G. Senter and A. W. Porter, ih.,
99. 1049, 1911 ; H. P. Cady and H. 0. Lichtenwalter, Journ. Amer. Chem. Soc, 35. 1434, 1913 ;
H. C. Allen, Instantaneous Chemical Reactions in Benzene and Toluene, Kansas, 1905 ; 0. D.
Chwolson, Traite de physique, Paris, 1910 ; S. F. Acree and J. M. Johnson, ib., 38. 258, 1907 ;
S. F. Acree, Trav.s. Faraday Soc, 15. 18, 1919.
2 A. Lapworth, Journ^ Chem. Soc, 93. 2187, 1908 ; 107. 857, 1915 ; S. Arrhenius, ib., 105.
1424, 1914 ; H. M. Dawson and T. W. Crann, ib., 109. 1262, 1916 ; H. Goldschmidt, Zeit. Elek-
trochem., 15. 6, 1909 ; G. Bredig, W. S. Millar, and H. Braune, ib., 18. 535, 1912 ; H. Goldschmidt
and A. Thuesen, ib., 18. 39, 1912 ; H. C. S. Snethlage, ib., 18. 535, 1912 ; Zeit. phys. Cliem.y
85. 211, 1913; 90. 1, 139, 1915; H. S. Taylor, Medd. Vet. Nobel-Jnst., 2. 34, 35, 37, 1913; 3.
1, 1914.
§ 5. Polarization— Back Electromotive Force
The source of chemical energy in the galvanic cell is certainly the chemical action, a
correction being applied for any reversible heat which the cell absorbs from or gives up
to its surroimdings. — W. C. D. Whetham.
Just as an ordinary steam boiler is a device for transforming the chemical energy
of burning coal into mechanical energy, so can the voltaic cell be regarded as an
1028 INORGANIC AND THEORETICAL CHEMISTRY
engine for converting chemical into electrical energy. In one of the simplest cases,
where the cell Zn : H2SO4 : Pt is working, hydrogen is evolved mainly from the
surface of the platinum. The chemical action is vigorous at first, but gradually
diminishes in intensity, and finally nearly stops altogether. The curve. Fig. 9,
shows the electromotive force of such a cell after different intervals of time when
it is working with a resistance of about ten ohms in the external circuit. The rapid
drop from an initial electromotive force of 13 volts to about half a volt in five
minutes, is indicated by the rapid descent of the curve. After five minutes the
electromotive force remained fairly constant at about 0*4 volt— that is, nearly 66 per
cent, below the initial value. The effect is easily illustrated by connecting an
electric bell with such a cell. The bell rings loudly at first, but gradually weakens,
and finally stops. If the platinum plate be then removed, the surface will be found
covered with a layer of bubbles of hydrogen gas, which have remained on the surface
of the plate instead of passing away. If the circuit is broken, the bubbles of gas
gradually dissipate from the platinum plate, and the cell then resumes its former
electromotive force when the circuit is closed. This temporary reduction in the
electromotive force of a cell is said to be due to the polarization of the cell. Polariza-
tion may be defined as the electrochemical fatigue which is caused either by a modifi-
cation of one or both of the plates, or by the exhaustion of the solution about the
surfaces of the electrodes during the working of the cell.
Whatever changes occur in the body of the electrolyte during electrolysis no
energy is consumed in the process, but at the junctions of electrodes and electrolytes
temporary or permanent changes may occur which produce the counter or negative
electromotive force called polarization. If the change is transient and disappears
when the current is stopped, it is temporary 'polarization in contrast with the so-
called permanent polarization which arises from a more or less permanent alteration
of the surface of the electrodes, and is utilized in the so-called secondary batteries
or accumulators. All the so-called polarization phenomena are secondary effects
of the change in the character of an electrochemical system produced by the
progressive exhaustion of one or more of the components. As a result of polariza-
tion, the normal current furnished by a cell may appear to be weakened owing to
the setting up of counter electromotive force ; or to the increased resistance of a
cell produced by changes in the concentration of the electrolyte or the decomposition
of layers of gas on the electrodes. There are three general methods available for
counteracting the ill effects of polarization ; or for depolarizing a cell :
(1) Mechanical. — The electrolyte or liquid about the electrodes is kept agitated
so that the gas layer may be brushed off, or the negative electrode may have a
roughened surface so that the bubbles of gas can more easily escape from points
on the surface.
In A, Smee's cell (1840) ^ the platinum plate is replaced by a silver plate covered with
finely divided platinum. The idea is to make the hydrogen bubbles collect at the points
of the roughened surface from which they escape more freely than from a plain smooth
plate. The voltage drop with Smee's cell, however, is comparatively large.
(2) Chemical. — The negative plate is surrounded by a powerful oxidizing agent
to oxidize any hydrogen which might be formed thereon. The ohemical agent
employed for this purpose is called a depolarizer — e.g. nitric acid, chromic acid,
manganese dioxide, etc.
W. R. Grove's cell (1839) has a porous pot like Daniell's cell, but the porous pot contains
a platinum plate and nitric acid ; zinc and sulphuric acid are placed in the outer cell. The
hydrogen produced by the action of sulphuric acid on the zinc, instead of accumulating on
the platinum plate is oxidized by the nitric acid, and the latter is reduced say symbolically :
H, + 2HN03 = 2H20 + 2N02. Owing to the high cost of platinum J. T* Cooper (1840)
proposed to use carbon in place of platinum in Grove's cell, and this modification is generally
known as R. Bunsen*s nitric acid cell (1841 ).
In R. Bunsen's dichromate cell (1875) the zinc and carbon plates dip in a solution of
ELECTRICAL ENERGY
1029
sulphuric acid mixed with chromic acid or an alkali dichromate. The zinc plates are attacked
by the solution when the cell is not in use, so that an arrangement is made whereby the
zinc plates can be drawn out of the solution when the cell is not being worked. The object
of the chromic acid is to " burn " up the hydrogen and prevent its accumulating on the
negative carbon plate. The chromic acid is at the same time reduced to chromic oxide,
CraOg. At the zinc plate, therefore: Zn + H2S04=ZnS04 + H2, and then 3H2 + 2Cr03
=Cr203 + 3H20. The solution is acidified with sulphuric acid, and consequently the chromic
acid forms chromic sulphate, Cr2(S04)3 ; potassium (or sodium) dichromate is the usual
source of the chromic acid, so that chrome alum, K2S04.Cr2(S04)3,24H20, usually forms,
and this can be partially separated from the zinc sulphate m dark purple crystals. Th,e
end stages of the reaction are therefore svmbolized : 3Zn+K,Cr207 4-7H2S04+Aq
=Aq + 3ZnS04 + K2S04+Cr2(S04)3 + 7H.20.
In G. Leclanche*s cell (1868) a zinc rod and carbon plate are clamped between two
blocks of compressed manganese dioxide and granulated car-
bon. This combination is dipped in a concentrated solution
of ammonium chloride. When the circuit is closed the ammonium
chloride attacks the zinc: Zn+2NH4Cl=ZnCl2 + 2NH3+H2.
The cell does not give a constant current very long, but if left
for a short time the accumulated hydrogen is '' burnt " by the
manganese dioxide: Ho + 2Mn02=Mn203+H20, and the cell
rapidly recovers. Millions of these cells must be in use where
current is required only for a few momenta — electric bells, etc.
— ^and the circuit is usually open. There are a great many
modifications.
::::-
■i«nigiililiiiB
z?.^ JJ l_J
''^ ^^f ::
- X-"^ -
o^ "" m^-ifn'^ffHi
3
Time
30 m in.
(3) Electrochemical. — The negative plate is immersed in „ q v' '\i'^ "n
a solution of a salt of the same metal — e.g. a copper ' yoitaic Cells,
plate in a solution of copper sulphate — as in Daniell's
cell — is free from polarization because metallic copper is deposited upon the plate,
presumably by the hydrogen which would otherwise accumulate on the copper plate.
In Daniell's cell, the variation in the electromotive force of the working cell is chiefly
due to changes in the concentration of the solution surrounding the battery plates.
The electromotive force is therefore nearly constant. This is illustrated by the
curve shown in Fig. 9, where a Daniell's cell is allowed to work for half an hour
against a resistance of 10 ohms in the external circuit. A comparison of this curve
with that of the Zn|H2S04|Pt cell emphasizes the constancy of the current delivered
by the Daniell's cell. In Clarke's cell, mercury deposits and zinc dissolves ; in
Weston's cell, mercury deposits and cadmium dissolves.
Back electromotive force. — Again, if a current exceeding two volts be directed
through an electrolytic cell containing dilute sulphuric acid, _
and fitted with two platinum plates and a galvanometer in
circuit, bubbles of gas are disengaged at the two electrodes,
oxygen at the anode, hydrogen at the cathode. The direction
of the current is indicated by the deflection of the needle of the
galvanometer. Now let the battery be cut out of the circuit,
and the electrodes immediately joined directly with the gal-
vanometer. The deflection of the needle shows that a feeble
current passes in an opposite direction to that which occurred
when the battery was in circuit. Here, then, when an electric
current is passed through a liquid, a counter-e.m.f. is set up,
relieving the stress set up by the original current. An examination of the plates
of the polarized electrolytic cell shows that gaseous films are present. Obviously,
therefore, after a current has passed through such a cell for a short time, the
plates — originally quite similar — are no longer alike. The plates are polarized
with different gases. The gases adhere to the surface and penetrate the interior of
the plates. The plates then behave as if they were made of two different materials.
Contact differences of potential are established. We have in fact a voltaic cell,
O2IH2SO4IH2, which furnishes a current flowing in an opposite direction to the
original current. The cell acts as a kind of " accumulator " of electrical energy until
the gases absorbed by the plates are used up. This does not take long. The polari-
zation of the plates of an electrolytic cell thus makes them behave like two different
Fig, 10.
1030 INORGANIC AND THEORETICAL CHEMISTRY
metals which exert a hack electromotive force opposing the electromotive force of the
battery. By Ohm's law, the current C (amps.), the resistance R (ohms), and the
electromotive force £"( volts) of a cell are related C=E/R ; and if e denotes the back
e.m.f. C=(E-e)IR.
A high-class steam engine will barely convert 13 per cent, of latent energy of
the fuel into useful work ; while a high-grade gas engine might convert up to about
35 per cent, of the available energy of the fuel into useful work. There is therefore
need for a more efficient and more direct means of converting the chemical energy
of the fuel into mechanical energy. Hence, one of the most important of all technical
problems is to get the largest possible amount of available energy from the combus-
tion of coal. In the ordinary zinc-platinum cell, 90 per cent, of the available energy
of the zinc is converted into electrical current. Zinc is far too costly a fuel for use
on a large scale, and hence chemists and physicists have sought a method of obtaining
electricity directly from the combustion of carbon or coal. In the so-called co7n'
hustion cells,^ carbon or some other cheap fuel is transformed into electrical energy
by consuming the oxygen of the air at one electrode, and some kind of fuel at the
other electrode. A. C. and A. E. Becquerel (1855) tried fusing potassium nitrate in
an iron crucible as one electrode and a carbon electrode dipped in the fused nitre
as the other electrode. The process was far too costly, and so far, all attempts to
dissolve carbon so as to convert the energy of oxidation into electric current have
had no real success. In all the proposed cells yet made, including the carbon mon-
oxide cell of W. Borchers and the coke cell of W. W. Jacques, the source of the electric
current has been traced not to the primary oxidation of the carbon, but either to
a secondary reaction or to a thermoelectric action similar to the development of an
electric current when the junction of two dissimilar metals is in a closed circuit.
References.
1 G. Leclanche, Mondes, 16. 532, 1868 ; Dingler's Journ., 186. 270, 1867 ; 188. 96, 1868 ;
Compt. Rend., 83. 54, 1876 ; 87. 529, 1878 ; R. W. Bunsen, Liebig's Ann., 38. 311, 1841 ; Pogg.
Ann., 54. 47, 1842 ; 55. 265, 1842 ; 155. 232, 1875 ; A. Smee, Phil. Mag., (3), 16. 315, 1840 ;
W. R. Grove, ih., (3), 14. 129, 1839 ; (3), 15. 287, 1839 ; (3), 21. 417, 1842 ; J. T. Cooper, ih.,
(3), 16. 35, 1840; L. Clark, Phil. Trans., 164. 1, 1874; Jmirn. Soc. Teleg. Eng., 7. 83, 1878;
E. Weston, Electrician, 30. 741, 1893.
2 E. Baur, Scient, Amer. Suppl, 75. 346, 1913 ; E. de Fodor, EleUricitdt direkt aus Kohle,
Wien, 1897 ; A. C. and A. E. Becquerel, Traite d'electricite et de magnetisme, Paris, 1. 183, 1855;
W. W. Jacques, Electrician, 36. 768, 1896 ; W. Borchers, Zeit. Elektrochem., 4. 42, 1897.
§ 6. Decomposition Voltages
During the electrolysis of a mixture of electrolytes those substances are set free which
absorb in becoming free tht> least intrinsic energy, or the lowest voltage. —J. T. Spraoue.
Suppose a current of half a volt be sent through the electrolytic cell, containing
normal sulphuric acid and fitted with platinum plates in circuit with a galvanometer.
The current passes through the cell for an instant as indicated by the " throw "
of the galvanometer needle, and then the quick drop to nearly zero. The hydrogen
and oxygen developed on the plate sets up a back electromotive force of nearly
J volt which very nearly stops the current. A minute steady current — residual
current — does flow through the system, but this is only just sufficient to maintain
the polarization, since if no current at all passes through, the plates would gradually
depolarize owing to the dissipation of the gases from the plates. If the current be
now raised to 1 volt, a similar state of things prevails. The amount of oxygen and
hydrogen adhering to the plates increases ; and the increased polarization raises
the back electromotive force to very nearly one volt. The residual current passing
through the cell is slightly larger than before. This is required to maintain the
ELECTRICAL ENERGY
1031
polarization. If the current is now raised to 1*7 volts, the electrodes become satu-
rated with hydrogen and oxygen gases. Polarization reaches a maximum value,
and the back electromotive force also attains its maximum value. Hence any
further increase in the applied electromotive force is available for electrolysis, 1*7
volts is the minimum needed for steady electrolysis. If 2 volts are passed through
the system, there is a back electromotive force of about 1*7
volts, and the " excess " or " residual " current, 0*3 volt, is the
effective electromotive force available for the production of
current, and the steady evolution of gases from the electrodes.
The facts here described can be exhibited very concisely by
plotting the applied electromotive forces as ordinates and
quantities of electricity passing through the system as abscissae.
Fig. 11 shows the results vidth normal solutions of sulphuric
acid, hydrochloric acids, and silver nitrate. The " residual "
current flowing through the cell with normal sulphuric acid
rises very slowly with increasing voltages until the driving force
reaches 1*67 volts. There is then a sudden change in the direc-
tion of the curve. Increasing electromotive forces now augment
the quantity of electricity passing through the system, and also
the amount of electrolysis. Normal hydrochloric acid gives a similar break at
rSl volts ; and silver nitrate, one at 0*70 volt.
The minimum electromotive force required to cause steady electrolysis in any
solution is called the decomposition voltage or discharge potential. The decompo-
sition voltages for a few acids, bases, and salts are shown in Table IV.
J
J
x^''
(^ '
'-i
;-4i
i
' 1
Current
Fig. 11.— Effect of
an increasing
E.M.F. on some
Electrolytes.
Table IV.^ — Discharge Potentials or Some Electrolytes.
Salts.
Acids.
Bases.
^-solutions.
Decom-
position
voltages.
JV-solutions.
Decom-
position
voltages.
^-solutiona.
Decom-
position
voltages.
Zinc sulphate
Nickel sulphate
Lead nitrate
Silver nitrate
2-35
2-09
1-52
0-70
Sulphuric acid
Hydrochloric acid
Nitric acid
Phosphoric acid
1-69
1-31
1-69
1-70
Sodium hydroxide
Potassium hydroxide
Ammonium hydroxide
1-69
1-67
1-74
While the values for the metallic salts vary from metal to metal, the acids and
bases have a decomposition voltage approaching 1*7 volts, and the products of
the electrolysis are oxygen and hydrogen. Those acids which have a lower decom-
position voltage usually give off other products on electrolysis, and attain the final
value — 1'7 volts — on further dilution. Thus hydrogen and chlorine are evolved
when the strength of the hydrochloric acid exceeds 2iV-HCl, and the decomposition
voltage of the 2iV acid is 1'26 volts. The voltage steadily rises with increasing
dilution until, with ^2-^-1101, the decomposition voltage is 1*69, and hydrogen and
oxygen are the products of electrolysis. Not only do the numbers vary with concen-
tration, within certain limits, as exemplified in the case of hydrochloric acid, but
also with the nature of the electrodes. The decomposition voltage of normal sulphuric
acid, for example, with polished platinum electrodes is 1'67 volts, whereas with
platinum electrodes covered with platinum black, the decomposition voltage is
1-07 volts.
The contact potential between metallic zinc and a normal solution of a zinc
salt, —0*493 volt, shows that when a zinc ion is deposited on a zinc electrode it
conveys a positive charge to the electrode and so lessens the negative charge there
present. The system is only in equilibrium when the zinc electrode is negatively
1032
INORGANIC AND THEORETICAL CHEMISTRY
charged to a potential of — 04:93 volt. If, therefore, zinc is to be deposited in an
electrolytic cell, this difference of potential must be counterbalanced by the current.
Hence contact differences of potential may also be regarded as decomposition
voltages.
The discharge potentials of a few anions and cations are indicated in Table lY,
which may be compared with Tables II and III. The numbers refer to normal
solutions. The prefix refers to the electrical state of the electrode in the presence
of a normal solution of its ions, say, 325 grams of zinc per litre. Some of the numbers
have not been measured directly. For instance, the number of zinc sulphate has
been obtained by extrapolation, since, according to the conductivity measurements,
only 23 per cent, of zinc sulphate is ionized in normal solutions.
Table V. — Dischaboe Potentals of Some Anions and Cations.
Cations.
Charge on metal
points.
Anions.
Charge in
volts.
Zn- .
-0-493
I' . . . .
+ 0-797
Fe-
-0-063
Br' .
+ 1-270
Ni"
+ 0 049
0" (in acid)
+ 1-396
Sn--
+0-085
cr .
+ 1-694
Pb-
+0-129
OH' (in acid)
+ 1-96
H-
+0-277
OH' (in bases)
+ 1-16
Cu-
Hg- .
+0-606
+ 1027
NO3'
SO4"
+ 1-75
+ 1-9
Ag-- ....
+ 1-048
HSO4'
+2
Just as different electrical pressures (e.m.f.) are needed to produce in different
solutions equivalent amounts of chemical change, so different chemical reactions in
a voltaic cell generate different amounts of electrical energy, and produce currents
with different electromotive forces. During electrolysis a difference of electrical
pressure must be continuously supplied because the current is consumed, so to
speak, by the separation of chemically equivalent quantities of matter (Faraday's
law). In a voltaic cell electrical energy is produced, so to speak, from the chemical
energy of the dissolving zinc. The question whether or not a given supply of
electrical energy can start electrolysis is determined by the intensity pressure,
or voltage of the current. The total supply of available electrical energy does not
matter. Although a given quantity or electricity, say 96,540 coulombs, will separate
chemically equivalent quantities of different electrolytes, these 96,540 coulombs
must be supplied at definite pressures before electrolysis can take place. In other
words, just as different compounds decompose at different temperatures, and this
quite independent of the total quantity of available heat, so electrical energy at
different voltages is needed for the decomposition of different electrolytes.
Current density. — If the cathode be small in comparison with the anode, the
solution about the former will be very much more quickly exhausted than if a larger
cathode had been used. The decomposition voltage of the substance will rise in a
proportional manner. Hence, the larger the cathode the lower the *' average '■
electromotive force needed for the decomposition of the pure metal. It is convenient
to call the quantity of electricity flowing through the unit surface area, the current
density at the electrode, in other words, " the number of amperes per unit surface,"
" Unit surface " is usually taken in the laboratory to be one square decimetre.
The symbol NDiqq=^ means that a current of 0'5 amp. flows for every 100 sq. cm.
of electrode surface.
Example. —What was the current density at each electrode of an electrolytic cell when
4 sq. cm. of each electrode was immersed in the electrolyte, and a current of 4-25 amperes
was passed through the system for one hour V One square decimetre = 100 sq. cm. Hence,
ELECTRICAL ENERGY
1033
r0625 amps, passed per sq. crn. ; or 106*25 amps, per sq. decimetre,
at the anode was therefore 106-25 amps., or iV/>ioo = 106-25.
The current density
Current density is one of the most important factors in electrolysis, since it deter-
mines the character and nature of the products obtained at the different electrodes.
Thus, by using a large current density and a concentrated solution of sulphuric
acid, hydrogen, oxygen, ozone, and free sulphur can be obtained, whereas under
ordinary laboratory conditions the last substance does not appear.^
References,
1 J. W. Turrentine, Journ. Phys. Chem., 14. 152, 1910.
§ 7. aas Cells
The term gas cell is applied to cells devised in 1839 by W. R. Grove i in which
gases are dissolved in inert electrodes, which are then treated as if they were electrodes
consisting of the gases alone — one gas may be hydrogen and the other oxygen or
chlorine, or even hydrogen at a different pressure. In the first case, water is
produced ; in the second, hydrogen chloride ; in the third, hydrogen passes from the
electrode, where it is at the greater pressure to that where it is at
the less. Such a cell with platinum electrodes is shown in Fig.
12. The e.m.f. of the cell is very irregular ; a day or two after
the preparation it approximates 1'08 volts in nearly all electrolytes,
although higher voltages have been observed by F. G. Smale,
T. N. M. Wilsmore, and E. Bose. Some days later the voltage
falls below the value just indicated. According to theory, the
potential difference between the two poles should be constant,
1-231( ±0001) volts (17°) in all electrolytes. In his review of the
potential of the oxygen electrode, E. P. Schoch has pointed out
that the value 1*08 volts cannot be due to the maximum e.m.f. of
the gases, i.e. the potential with which the action of the poles is
reversible. If the cell be reversed, by discharging oxygen at the
poles by electrolysis, the gases are not evolved at a potential
slightly greater than 1-231 volts; instead, a potential difference
exceeding 1-5 volts must be applied before any current greater
than those due to diffusion, convection, etc., will pass, and the
evolution of gases occur.
The hydrogen electrode is quite reversible in its action, and
its potential is independent of the metal used in its production,
and can be obtained in all kinds of electrolytes, all the observed irregularities
must be due to the oxygen electrode. The work of F. Forster, E. Miiller, R. Lorenz,
L. Wohler, and R. Ruer has shown that the irregularities with the oxygen electrode
are due to the oxidation of the electrode.
The discharge potential of an ion is determined by the opposing potential of
the electrode ; thus, to discharge oxygen or hydrogen at an electrode devoid of any
electrochemical activity requires but a small e.m.f., but if the respective electrodes
be charged with these gases the e.m.f. required is over one volt. If the products
of the electrolysis react with the electrodes, then the discharge potential is
determined by the nature of the film formed at the surface of the electrode. Thus,
the discharge potential of chlorine at an inert electrode, say, graphite, is quite
different from its value at a silver electrode because silver chloride is formed ; simi-
larly, oxygen gas is liberated at a lower potential at platinum than at a lead electrode,
because in the latter case lead dioxide is formed ; R. Luther and F. J. Brisbee found
the discharge-voltage of chlorine from hydrochloric acid at a polished platinum
Fig. 12.— W. R.
Grove's Oxy-
gen-Hydrogen
Gas Cell.
1034 INORGANIC AND THEORETICAL CHEMISTRY
electrode is higher than is required for the discharge of chlorine from platinum
electrodes, possibly because of the formation of a film of some product on the
surface of the electrode ; and E. P. Schoch found evidence of the formation of a
hydride on the surface of iron or nickel electrodes. In cases of this kind, the main
process is irreversible because the extra potential entails a loss of free energy.
Hence, the term discharge potential may have two meanings, one refers to the
reversible, the other to the irreversible process ; it usually has the former meaning.
G. Premier showed that with the gas cell PtlHIHgOIOalPt, when the pressures of the
gases at the electrodes are each one atmosphere, and the hydrogen electrode the partial
pressure of the hydrogen in saturated water vapour are respectively reduced to Pi and p„,
the e.m.f. E is
^ RT ^ I RT ^ I ^ RT ^ 1
E= -- log - +_ log^ ; or ^= — log ^-
W. Nemst and H. von Wartenberg calculate at 290° K., Pi =0-0191 x 1-80 x 10"" atm.
and P2=JPi ; hence, ^ = 1-2322 volts (17°) ; in agreement with this value, J. N. Bronsted
and G. N. Lewis respectively obtain 1-224 and 1-234 volts. The difference between these
theoretical values and the observed lower value 1-15 is attributed to the observations being
made on what virtually amounts to a platinmn oxide, not platinum electrode. T. N. M.
Wilsmore calculated d^/dlT =--0-00121 ; F. J. Smale found -0-00142.
When a gradually rising e.m.f. is applied to platinized electrodes in dilute sulphuric
acid or sodium hydroxide, only small currents of the order of diffusion currents
pass, the potential of the anode rises rapidly until it passes 1*50 volts, when bubbles
of oxygen appear ; the anode potential still rises but less rapidly than before. If
it were a reversible electrode, which had turned its reversible point at VbO volts,
the current voltage curve would not have risen but continued nearly parallel to the
current axis. The polarization potential also increases steadily with time so long
as the current is continuous, but there is no indication of a definite maximum for a
particular current density. This polarization is due to a specific surface attraction
between the platinum and the gas, as H. G. Moller showed to be the case with
hydrogen, because (i) the range of potential extending to nearly three volts is too
great, and (ii), as F. Forster showed, the platinum is able to function at a low or
high potential which would require the assumption of an arbitrary change in the
absorption power of the gas. The effect is also incompatible with the difference of
potential being caused by a resistance film. In this manner, E. P. Schoch argues
that the potential rise requires that the active substance formed on the anode have
a physical form which can change its concentration continuously. Again, if a single
solid were formed, the potential would remain practically constant ; and it is assumed
that the material which is formed should be dissolved by the remainder of the
electrode so that it exhibits an increase of potential corresponding with the increased
concentration.
When the polarized oxygen electrode is left at rest, the self-discharge results
in a continuous decrease in the e.m.f. until a potential difference of 1'08 volts is
attained, the potential retains this value for some time — steady state — and after-
wards falls to still lower values. R. Lorenz found that, unlike the platinized platinum
electrode investigated by F. Forster, a polished platinum electrode exhibits a large
number of abrupt changes during its discharge. The halting stages, so to speak,
occur at I'S, 1'05, 0'94, 0-74, 0'64, 0*57, 0-43, 0*27, 0-12, 0*05, and O'OOS volts. This
is supposed to correspond with the formation of definite oxides, with different
potentials. Generally, however, the step-by-step discharge is not shown because
the drop of potential during discharge is continuous. F. Forster showed that the
self-discharge of an iridium anode is quite analogous to that of the platinum anode,
but the drop is rather more rapid and it falls to 0'865 volt. The behaviour of lead
is rendered familiar through the accumulator ; and F. Streintz studied the different
potentials exhibited by the hydrated lead oxides. The nickel oxide anode in alkaline
solutions has been studied by F. Forster in connection with the nickel accumu-
lator ; J. Zedner, and F. Forster and V. Herold the iron electrode ; E. Miiller studied
ELECTRICAL ENERGY
1035
the copper anode in alkali lye ; and R. Lorenz has shown that oxygen gas electrodes,
with the metals lead, silver, nickel, copper, iron, and zinc, exhibit potentials analogous
to those shown by their oxides. The two metals usually regarded as non-oxidizable
by gaseous oxygen are therefore supposed to be oxidized, and to owe their potential
to the presence of oxides. In general, therefore, (1) During the discharge of " oxygen
yielding " anions all metal electrodes are oxidized. (2) The potential of the electrode
is that of the oxide irrespective of any (adsorbed) oxygen gas also present. (3) The
oxides specifically determine the potentials with which oxygen is evolved. (4) The
amount of an oxide that must be actually present to give all characteristic effects
may be less than is optically perceptible. (5) Oxygen gas does not appear to be
directly electromotively active.
The oxides PtO, Pt02, PtOs, with several hydrated forms, were prepared by
L. Wohler, and R. Lorenz determined the potentials of the different oxides used as
anodes as well as thei potential at which the steady state occurs during the discharge
of the platinum anode.
PtO„.4H20
PtOg-SHgO
Pt02.2H20
PtOa-HgO
PtO,
Pt0.2H20
PtO.HgO .
PtO
Oxide
otential.
0-93
Potential during
steady state.
0-94
0-86
. —
0-74
0-74
0-63
0-64
0-53
0-57
0-45
0-43
0-34
—
0-25
0-27
The value for PtOg has not been measured because it decomposes so rapidly, but it
is supposed to lie above 0*43, and it is therefore thought to be the oxide to which
all higher potentials are due. For potentials above one volt, R. Lorenz believes the
different potentials are due to a number of distinct oxides or hydrates, while F. Forster
attributes them to the formation of a solution of a higher oxide in the material of
the electrode.
F. Forster and E. Miiller believe that the evolution of oxygen from the oxygen
electrode is a secondary effect, due to the formation and decomposition of a higher
oxide ; if this means that oxygen is not evolved except through such action, it is
not in agreement with G. Schulze's observation that the formation of oxygen also
occurs while other oxides are present which are not capable of such decomposition —
e.g. alumina, magnesia, etc.
References.
1 W. R. Grove, Phil Mag., (3), 14. 129, 1839; (3), 21. 417, 1842; Phil. Trans., 133. 91,
1843 ; The Correlation of Physical Forces, London, 253, 1874 ; F. Forster, Zeit. phys. Chem.,
69. 236, 1909 ; 38. 1, 1901 ; G. Schulze, ib., 69. 236, 1909 ; R. Luther and F. J. Brisbee,
ib., 45. 216, 1903; K. Bennewitz, ib., 72. 202, 1910; J. B. Westhaver, ib., 51. 65, 1905;
T. N. M. Wilsmore, ib., 35. 298, 1900 ; 36. 91, 1901 ; F. J. Smale, ib., 14. 577, 1894 ; G. Preuner,
ib., 42. 57, 1903 ; W. Nemst and H. von Wartenberg, ib., 56. 534, 1906 ; J. N. Bronsted, ib.,
65. 84, 744, 1909; G.N. Lewis, i6., 55. 465, 1906; Journ. Amer. Chem. Soc, 28. 158, 1906;
V. Czepinsky, ZeiL anorg. Chem., 30. 1, 1902 ; H. G. MoUer, Zeit. phys. Chem., 65. 226, 1908 ;
E. Bose, ib., 34. 701, 1900 ; 38. 1, 1901 ; E. Bose and H. Kochan, ib., 38. 28, 1901 ; E. Muller, Zeit.
KUktrochem., 13. 133, 1907 ; E. Bose, ib., 5. 169, 1898 ; R. Lorenz, ib., 15. 293, 349, 1909 ;
F. Forster, ib., 13. 414, 1907 ; 14. 17, 1908 ; J. Zedner, ib., 13. 752, 1907 ; F. Forster and V. Herold,
ib., 16. 46], 1910; 0. Faust, ib., 13. 161, 1907; R. Ruer, ib., 11. 661, 1905; L. Wohler, ib.,
15. 769, 1909; Zeit. anorg. Chem., 51. 81, 1906; Ber., 36. 3475, 1903; E. P. Schoch, Journ.
Phys. Chem., 14. 665, 1910 ; Amer. Chem. Journ., 41. 226, 1909 ; F. Streintz, Wied. Ann., 49.
564, 1893.
1036 INORGANIC AND THEORETICAL CHEMISTRY
§ 8. The Relation between Electrical and Thermal Energy
No chemical development will be satisfactory and permanent xinless erected on a
thoroughly physical basis. — O, J. Lodge (1885).
The total amount of electrical energy required for the liberation of chemically
equivalent quantities of different electrolytes can be approximately determined by
multiplying 96,540 coulombs (or one farad) of electricity by the voltage needed
for electrolysis. Hence, the decomposition voltage is proportional to the energy
needed for the decomposition of a gram equivalent of a given electrolyte, and the
product of the quantity of electricity into its electromotive force not only representing
the energy of a battery, but it also measures the chemical energy which was trans-
muted into electrical energy by the battery. As previously indicated, a joule, the
unit of electrical energy, is numerically equivalent to the product of one volt into
one coulomb. The amount of heat evolved when a given compound is decomposed
can be measured and the minimum amount of electrical energy required to decompose
a given compound must be at least equivalent to the amount of heat developed when
the separate substances re-unite to form the original compound. Measurements
show that a joule is equivalent to 0*24 calorie of thermal energy ; and a calorie is
equivalent to 4*2 joules. Hence, just as thermochemistry writes Na-f Cl=NaCl
+97,900 cals., so electrochemistry writes Na+Cl=NaCl-i-411,000 joules. As a
first approximation, it may be assumed that the heat of formation of any given
compound is a measure of the thermal equivalent of the electrical energy required
to break up the compound by electrolysis.
From the first law of thermodynamics, the law of the equivalence of the different
forms of energy, if the work W ergs done by an electric current be wholly expended
in decomposing a substance, W=JQ, where Q cals. denotes the thermal equivalent
of the electrical energy, and J is a numerical conversion factor required, to make the
units of work and heat comparable. Here «/=42 X 10^ because one calorie is nearly
equivalent to 42x10^ ergs or 4*2 joules. By definition, the work done in a circuit
per unit of electricity conveyed is W=EC ; and from the definition of electrochemical
equivalent m=eC. Again, the heat q liberated during the formation of one gram
of a compound from the radicles into which it has been decomposed by a current in
q=Qlm provided none of the energy remains in any form other than heat. Substi-
tuting these values of Q, m, and W in W=JQ, there remains E=Jeq, which is a
symbolic form of the statement that the electromotive force required to decompose
a substance into given constituents is equal to the product of the heat of formation
of a gram equivalent to the substance and the number of joules per calorie ; or
E=4:-2eq joules.
Example.- — The heat of formation of sodium chloride is 97,900 calories : what is the
equivalent electrical energy needed for the electrolysis of a gram equivalent of the fused
salt, and what is the decomposition voltage required ? Here, 97,900 calories are equivalent
to 97,900x4-2=411,000 joules. But 96,540 coulombs will liberate chemically equivalent
quantities of sodium and chlorine, and 411,000 joules are needed for this purpose. Conse-
quently, since electrical energy = volts x coulombs ; 411,000= volts x 96,540 ; or volts = 4-3.
This means that in order to liberate 23 grams of sodium and 35-5 grams of chlorine from
58-5 grams of fused sodium chloride, 411,000 joules of electrical energy must be supplied
at a minimum voltage electromotive force of 4*3 volts. The minimum voltages so calculated
are usually a little higher than are needed in practice. It will be observed that this arith-
metic is summarized in the formula : volts =0*043 X Cals. Where Calories are employed
to represent the heat developed in the reaction of an equivalent weight of a given compound
expressed in grams, the equivalent of water is half the molecular weight in grams, the
equivalent of aluminium chloride is one-third the molecular weight expressed in grams.
The computation of the electromotive force of various battery cells from the
heat of combination of the " elements " of the cell has been of great value in tech-
nology. It is commonly assumed that the electrical energy which a battery can
supply may be calculated directly from the thermochemical data. According to
the old observation of Lord Kelvin {ante W. Thomson), i the electrical energy which
ELECTRICAL ENERGY 1037
can be obtained from a galvanic element is equivalent to the thermal value of the
chemical processes producing the current when the current is not doing any special
work in the circuit — Kelvin's rule. In illustration, the thermal value of the reac-
tion in a Darnell's cell is Zn+CuSO4=Cu+ZnSO4+50-ll Cals. That is to say,
every gram-atom of zinc dissolved in the reaction is attended by the evolution of
SO'll units of heat. Every gram-atom of bivalent zinc carries in the cell 2 X 96,540
=19,300 coulombs of electricity. The electromotive force developed during the
action is 1 '096 volts. Hence the thermal equivalent of the electrical energy developed
by the dissolution of one gram-atom of zinc in Daniell's cell is 0*24 X 1*096 x2 X 96,540
=50,000 cals., provided the electrical energy produced is equal to the chemical
energy used up. The difference between the thermal value of the chemical processes
and the thermal value of the electrical energy derived from the cell is 50" 11 and
50-00=0H Cal. The two quantities differ by about one-fourth per cent. ; the
difference is within the range of experimental error, and Kelvin's rule, cited above,
is valid. J. P. Joule immersed a cell in a calorimeter, and its outer circuit in another,
and he found that the heat energy of the cell can be made to appear in the outer
circuit. Further experiments showed that some cells directly heat and others
directly cool themselves.
As a matter of fact, observation shows that cells sometimes furnish more, and
sometimes less energy than corresponds with the thermal data, and the difference
between the thermal values of the chemical process and the electrical energy exceeds
the limits of experimental error ; thus, with the Pb : Pb(N03)2 : AgNOs : Ag cell
the difference is —16 per cent. ; with Pb : lead acetate : copper acetate : Cu the differ-
ence is nearly +15 per cent. The assumption that the electrical energy derived from
a cell is equivalent to the heat of the reaction of the components of the cell, is one of
the half truths illustrated further by Berthelot's law of Inaximum work. The
electrical energy is equivalent to the free energy, and that alone is a measure of
the maximum work obtainable from a chemical reaction.
If the cell in which the electrical energy is being produced rises in temperature,
less electrical energy than is represented by Kelvin's rule will be obtained, because
part of the chemical energy may also be converted into heat in working against the
resistance of the cell ; and a correction factor is required to allow for the energy
dissipated in this way. Conversely, if a cell becomes cooler while it is working, more
than the calculated quantity of electrical energy might be expected from the cell.
Only when the working cell suffers no change of temperature is the electrical energy
produced equal to the chemical energy expended. Similar remarks, mutatis mutandis,
apply to the reverse action during electrolysis. In electrolyzing fused cryolite,
for example, J. Hopkinson obtained but 60 per cent, of the amount of aluminium
corresponding with the electrical energy expended. In addition, some energy may
be expended during electrolysis in overcoming polarization, in secondary chemical
reactions, etc. Suppose that the temperature changes ever so little, say dT, while
the cell is working, there will be a corresponding change, dE, in the electromotive
force of the cell ; let q denote the amount of heat absorbed or evolved when a gram-
equivalent of the electrolyte is decomposed. Let the cell be placed in a bath main-
tained at a constant temperature, the units of heat must be added to or given up by
the cell if its temperature is kept constantly at T°. If S and T respectively denote
the capacity and intensity factors of thermal energy, while C and E denote the
corresponding factors of electrical energy, then q=ST, and €=CE, and when the
two forms of energy are in equilibrium, CE=ST, and when T and E change as
indicated above, C{E-\-dE)=S(T-\-dT), or CdE=S.dT, and by substituting q=ST,
we get
CdE^Q^^, or ?=Crg (1)
This expression represents the change in the thermal value of the electrical energy
of a cell at T° when the temperature changes dT during the working of the cell.
1038 INOKGANIC AND THEORETICAL CHEMISTRY
and the electromotive force increases by dE per degree rise of temperature. The
factor dEjdT is called the temperature coefficient of a cell, and it is usually evalu-
ated by measuring the electromotive force of the cell at two different temperatures.
Example.' — The e.m.f. of a cell at 0° was 01483 volt ; and at 43-3°, 0-1846 volt. Hence,
the increase in the e.m.f. per degree is (0-1846-0-1483)-r-43-3 = +0000838 volt^dE/dT.
If one farad of electricity be passed through a reversible cell, and it is found
necessary to apply q calories of heat in order to maintain the temperature constant,
the electrical energy which would be furnished by the cell working in the opposite
direction will be equivalent to the thermal value of the reaction, Q-{-q, or expressed
in suitable units, electrical energy =Q+g'. Substituting the value of q obtained in
(1) above, and electrical energy e=CE,'we get CE=Q-^CTdE/dT, or
^ __ ^p e=Q+Cl§., and E^+jf^
This is called the Gibbs-Helmholtz eauation— after J. W. Gibbs (1878) and
H. von Helmholtz (1882). The equation shows that in order to calculate the electro-
motive force ^ of a galvanic element, from Q, the thermal value of the chemical
processes which occur during the working of the cell, it is necessary to know the
temperature coefficient showing the variation of e.m.f. with temperature. By
measuring the electromotive force E and the temperature coefficient dEjdT of a
zinc-iodine combination, A. P. Laurie (1885) 2 was able to estimate the heat of
combination of these two elements.
1. If the temperature coefficient of a cell he negligibly small, dEjdT may be taken as
zero, and Kelvin's equation E=QjG remains. Hence, Kelvin's rule indicating the
relation between the equality of the thermal and electrical energy of a cell is a limited
equation which is valid only when the electromotive force of the cell does not change
with variations of temperature. This is nearly the case with the Daniell's cell,
and in consequence, the electromotive force calculated from the thermal value of
the reactions in the cell is nearly equal to the observed value.
Examples. — -(1) Assume the temperature coefficient of Daniell's cell is zero, and the
heatof the reaction 50,110cals. ; then 50,110 calories equal 50,110-7-0-24 = 209,900=^ joules ;
and C=2 X 96,540. Hence, ^ = 1*087 volts ; the observed voltage is 1-096.
(2) According to F. Haber and S. Tolloczko (1904) the thermal value of reaction in the
AgCl:CuCl cell is Cu-|-AgCl=Ag+CuCl + 3500 cals., and at 200° the temperature
coefficient of the cell is zero. Show that the calculated e.m.f. is accordingly 0-151 volt.
(The observed value is 0*149 volt.)
2. If the temperature coefficient of the cell dEjdT he negative, the electromotive
force will diminish with rising temperature, and the electrical energy derived from
the cell will be less than that computed from the thermal energy. The cell will
therefore become hot while it is working. This does not refer to heat due to the
internal resistance of the cell. L. Clark's cell has a temperature coefficient of
— 0-0012345 volt per degree at about 18°.
Examples.- — (1) Show that the e.m.f. of the oxyhydrogen cell is l-075-|-0-0014^ when
Ha -f-0 = H20 -1-67,520 cals., and the temperature coefficient of the cell is —0-0014 volt per
degree.
(2) The reaction in a Clark's cell evolves 340,000 joules. Show that the temperature
coefficient of the cell is —0-00114 when the e.m.f. at 18° is 1-429.
(3) According to G. N. Lewis and C. A. Kraus (1910), the difference of potential of a
cell with sodium and 0-206 per cent, sodium amalgam as electrodes in a solution of sodium
iodide in ethylamine is 0*8456 volt at 25°, and the temperature coefficient —0-0000408
volt per degree. Hence calculate the heat evolved during the solution of one equivalent
of sodium in an excess of the 0-206 per cent, amalgam. Substitute in Helmholtz's equation,
C is the Faraday equivalent = 96,540 coulombs; i^' =0-8456 ; dE/dT=— 0-0000^08 ;
T = 25+213. The result furnishes Q = 82,850 joules ; and since one joule is equivalent to
4'186 cals., the thermal value of the process is nearly 19,800 cals.
ELECTKICAL ENERGY 1039
3. If the temperature coefficient of the cell dE/dT he positive, the electromotive
force of the cell will increase with rise of temperature and the electrical energy of
the cell will be less than that computed from the thermal value of the chemical
processes, and heat will be abstracted from the surrounding objects in order to
maintain the temperature of the cell constant. Such a cell will become colder in
action.
Examples.' — (1) The slight discrepancy between the observed and calculated values of
the electromotive force of Daniell's cell indicated in a preceding example is due to the
small but measurable positive temperature coefficient of the cell, for dE/dT — -}- 0-000034:
vol. per degree at 15°. If this factor be introduced into the previous calculation for the
cell working at about 15°, T = 280° K., we must add 0-000034x288 to the 1-087 volts
obtained in the previous computation. This furnishes 1-096 volts; the observed value is
1-092 volts.
(2) According to F. Haber and S. Tolloczko (1904), the thermal value of the reaction
Pb-|-2AgCl=PbC]2+Ag + 24 Cals. The temperature coefficient is +0-152 millivolt per
degree at about 250°. What is the calculated e.m.f. of the lead chloride ; silver chloride
cell ?
The self-cooling of a working cell has been compared with the self-cooling of
a freezing mixture, or of a jet of compressed gas.
References.
1 W. Thomson, Phil Mag., (4), 2. 429, 1851; J. P. Joule, ib., (3), 19. 260, 275, 1841;
(3), 20. 204, 1842; J, W. Gibbs, Proc. Connecticut Acad., 3. 501, 1878; H. von Helmholtz,
Sitzher. Mad. Berlin, 22, 825, 1882,
2 A. P. Laurie, Phil. Mag., (5), 21. 409, 1886; Journ. Chem. Soc, 49. 700, 1886; F. Haber
and S. Tolloczko, Zeit. anory. Chem., 41. 407, 1904.
§ 9. Fractional Electrolysis— G. Magnus' Rule
During the electrolysis of a number of mixed electrolytes, there is a selective power at
the electrodes which is based solely on the ratio of the voltage required to free the several
ions.- — J. T. Sprague.
When a solution containing salts of different metals is subjected to electrolysis,
there is a certain voltage at which one and only one of the metals will be deposited
on the cathode — G. Magnus' rule (1856). If a mixed solution of nickel and copper
sulphates, for example, be subjected to electrolysis, copper alone is precipitated when
the applied electromotive force has reached 1*29 volts ; the nickel is not precipitated,
since its decomposition voltage is 1*95 volts. On the other hand, if a mixture of
nickel and iron sulphates be similarly treated, a mixture of iron and nickel will be
simultaneously deposited. The decomposition voltage of these salts are too close
to allow an effective separation of the two elements by electrolysis. Hydrogen is
also evolved during the electrolysis of these salts. This arises from the fact that
the decomposition voltage of sulphuric acid — 1"67 volts — renders it also susceptible
to the influence of the same current as liberated nickel and iron.
Many useful methods of analysis are based upon these principles. In metallurgy,
too, electrolytic processes for refining metals — nickel, copper, lead, tin, silver, gold,
etc. — have been developed. For example, in copper refining, as we shall soon see,
anodes made of crude copper are dipped in a solution of copper sulphate acidified
with sulphuric acid ; the cathodes are sheets of pure copper. Zinc, iron, and
copper from the anode pass into solution during electrolysis. The decomposition
voltage is kept below that needed for the deposition of zinc and iron. In conse-
quence, refined copper is deposited upon the cathode. Other impurities affecting
the crude copper are but slightly soluble in the electrolyte, and are deposited about
the anode as a thin mud — anode mud.
The effect of concentration on the decomposition voltage. — The decomposition
1038 INORGANIC AND THEORETICAL CHEMISTRY
and the electromotive force increases by dE per degree rise of temperature. The
factor (lEIdT is called the temperature coefficient of a cell, and it is usually evalu-
ated by measuring the electromotive force of the cell at two different temperatures.
Example.— The e.m.f. of a cell at 0° was 01483 volt ; and at 43-3°, 0*1846 volt. Hence,
the increase in the e.m.f. per degree is (0-1846-0-1483)-r-43'3 = +0000838 volt=dE/d2\
If one farad of electricity be passed through a reversible cell, and it is found
necessary to apply q calories of heat in order to maintain the temperature constant,
the electrical energy which would be furnished by the cell working in the opposite
direction will be equivalent to the thermal value of the reaction, Q-{-q,0T expressed
in suitable units, electrical energy =Q+5'. Substituting the value of q obtained in
(1) above, and electrical energy e=CE,'we get CE=Q-\-CTdEldT, or
^ . ^p e=Q+Cl§; an. ^=§+rg
This is called the Gibbs-Helmholtz eauation— after J. W. Gibbs (1878) and
H. von Helmholtz (1882). The equation shows that in order to calculate the electro-
motive force ^ of a galvanic element, from Q, the thermal value of the chemical
processes which occur during the working of the cell, it is necessary to know the
temperature coefficient showing the variation of e.m.f. with temperature. By
measuring the electromotive force E and the temperature coefficient dE/dT of a
zinc-iodine combination, A. P. Laurie (1885) 2 was able to estimate the heat of
combination of these two elements.
1. If the temperature coefficient of a cell he negligibly small, dE/dT may be taken as
zero, and Kelvin's equation E=QIC remains. Hence, Kelvin's rule indicating the
relation between the equality of the thermal and electrical energy of a cell is a limited
equation which is valid only when the electromotive force of the cell does not change
with variations of temperature. This is nearly the case with the Daniell's cell,
and in consequence, the electromotive force calculated from the thermal value of
the reactions in the cell is nearly equal to the observed value.
Examples. — '(1) Assume the temperature coefficient of Daniell's cell is zero, and the
heat of the reaction 50,110 cals. ; then 50,110 calories equal 50,110-4-0-24 = 209,900=^ joules ;
and (7 = 2 X 96,540. Hence, ^ = 1-087 volts ; the observed voltage is 1-096.
(2) According to F. Haber and S. Tolloczko (1904) the thermal value of reaction in the
AgCl:CuCl cell is Cu+AgCl=Ag+CuCl + 3500 cals., and at 200° the temperature
coefficient of the cell is zero. Show that the calculated e.m.f. is accordingly 0-151 volt.
(The observed value is 0*149 volt.)
2. If the temperature coefficient of the cell dE/dT he negative, the electromotive
force will diminish with rising temperature, and the electrical energy derived from
the cell will be less than that computed from the thermal energy. The cell will
therefore become hot while it is working. This does not refer to heat due to the
internal resistance of the cell. L. Clark's cell has a temperature coefficient of
—00012345 volt per degree at about 18°.
Examples.- — (1) Show that the e.m.f. of the oxyhydrogen cell is 1-075 + 0-0014^ when
Ha + 0 = H20 + 67,520 cals., and the temperature coefficient of the cell is —0-0014 volt per
degree.
(2) The reaction in a Clark's cell evolves 340,000 joules. Show that the temperature
coefficient of the cell is —0-00114 when the e.m.f. at 18° is 1-429.
(3) According to G. N. Lewis and C. A. Kraus (1910), the difference of potential of a
cell with sodium and 0-206 per cent, sodium amalgam as electrodes in a solution of sodium
iodide in ethylamine is 0-8456 volt at 25°, and the temperature coefficient —0-0000408
volt per degree. Hence calculate the heat evolved during the solution of one equivalent
of sodium in an excess of the 0-206 per cent, amalgam. Substitute in Helmholtz's equation,
C is the Faraday equivalent = 96,540 coulombs; i; =0-8456 ; dE/dT = —0-00004.08 ;
jr = 25+273. The result furnishes Q = 82,850 joules ; and since one joule is equivalent to
4-186 cals., the thermal value of the process is nearly 19,800 cals.
ELECTRICAL ENERGY 1039
3. If the temperature coefficient of the cell dE/dT he positive, the electromotive
force of the cell will increase with rise of temperature and the electrical energy of
the cell will be less than that computed from the thermal value of the chemical
processes, and heat will be abstracted from the surrounding objects in order to
maintain the temperature of the cell constant. Such a cell will become colder in
action.
Examples.' — (1) The slight discrepancy between the observed and calculated values of
the electromotive force of Daniell's cell indicated in a preceding example is due to the
small but measurable positive temperature coefficient of the cell, for dE/dT — -\- 0-000034:
vol. per degree at 15°. If this factor be introduced into the previous calculation for the
cell working at about 15°, T==280° K., we must add 0-000034x288 to the 1-087 volts
obtained in the previous computation. This furnishes 1-096 volts; the observed value is
1-092 volts.
(2) According to F. Haber and S. ToUoczko (1904), the thermal value of the reaction
Pb + 2AgCl=PbCl2+Ag + 24 Cals. The temperature coefficient is +0-152 millivolt per
degree at about 250°. What is the calculated e.m.f. of the lead chloride : silver chloride
cell ?
The self-cooling of a working cell has been compared with the self-cooling of
a freezing mixture, or of a jet of compressed gas.
References.
1 W. Thomson, Phil. Mag., (4), 2. 429, 1851 ; J. P. Joule, ib., (3), 19. 260, 275, 1841 ;
(3), 20. 204, 1842; J. W. Gibbs, Proc. Connecticut Acad., 3. 501, 1878; H. von Hehnholtz,
Sitzher. Alcad. Berlin, 22, 825, 1882,
2 A. P. Laurie, Phil. Mag., (5), 21. 409, 1886; Journ. Chem. Soc, 49. 700, 1886; F. Haber
and S. ToUoczko, Zeit. anory. Chem., 41. 407, 1904.
§ 9. Fractional Electrolysis — G. Magnus' Rule
During the electrolysis of a number of mixed electrolytes, there is a selective power at
the electrodes which is based solely on the ratio of the voltage required to free the several
ions.- — J. T. Sprague.
When a solution containing salts of different metals is subjected to electrolysis,
there is a certain voltage at which one and only one of the metals will be deposited
on the cathode — G. Magnus' rule (1856). If a mixed solution of nickel and copper
sulphates, for example, be subjected to electrolysis, copper alone is precipitated when
the applied electromotive force has reached r29 volts ; the nickel is not precipitated,
since its decomposition voltage is r95 volts. On the other hand, if a mixture of
nickel and iron sulphates be similarly treated, a mixture of iron and nickel will be
simultaneously deposited. The decomposition voltage of these salts are too close
to allow an effective separation of the two elements by electrolysis. Hydrogen is
also evolved during the electrolysis of these salts. This arises from the fact that
the decomposition voltage of sulphuric acid — 1'67 volts — renders it also susceptible
to the influence of the same current as liberated nickel and iron.
Many useful methods of analysis are based upon these principles. In metallurgy,
too, electrolytic processes for refining metals — nickel, copper, lead, tin, silver, gold,
etc. — have been developed. For example, in copper refining, as we shall soon see,
anodes made of crude copper are dipped in a solution of copper sulphate acidified
with sulphuric acid ; the cathodes are sheets of pure copper. Zinc, iron, and
copper from the anode pass into solution during electrolysis. The decomposition
voltage is kept below that needed for the deposition of zinc and iron. In conse-
quence, refined copper is deposited upon the cathode. Other impurities affecting
the crude copper are but slightly soluble in the electrolyte, and are deposited about
the anode as a thin mud — anode mud.
The effect o! concentration on the decomposition voltage. — The decomposition
1040 INORGANIC AND THEORETICAL CHEMISTRY
voltage of an electrolyte is greater the more dilute the solution. The concentration
of any given salt about the electrode naturally decreases during the process of
electrolysis. Hence also the decomposition voltage for that particular salt in the
mixed electrolyte also increases. When the concentration of the copper sulphate
in a mixture of copper and nickel sulphates has become so small that the decomposi-
tion voltage of the dilute solution approaches that of nickel, any further electrolysis
will bring down a mixture of both metals. There is, therefore, a limit to the process
of electrolytic separation, just as there is a limit to the separation of substances in
ordinary analysis. The limit in the former case is determined by the decomposition
voltages of the respective metals ; and in the latter case, the limit is determined by
the solubility of the precipitates in the given menstruum. The limiting concentration
can be approximately estimated from the rule : A decrease of one-tenth in the
concentration of the electrolyte raises the decomposition voltage of any given
ion 0'058/w volt, where n is the valency of the particular ion.
As the cation is deposited about the cathode, the loss in concentration is made
up by diffusion from the surrounding electrolyte. To hasten diffusion, and prevent
undue attenuation of the electrolyte in the vicinity of the cathode, stirring by
rotating one of the electrodes is sometimes used.
INDEX
" Abnormal " in chemistry, 192
Absolute boiling-point, 166
temperature, 160
zero, 160
Absorption coefficient, 627
Abu-r-Raihan, 42
Acad^mie des Sciences, 5
Academy of Nature's secrets, 2
Accademia del Cimento, 4
dei Segreti, 2
Acetamide and hydrogen, 304
Acetic acid and hydrogen, 303
Acetone and hydrogen, 304
Acicular crystals, 597
Acid, anhydrides, 390
history, 382
primitive, 384
primordial, 384
salts, 387
Acidimetry, 391
Acidity, principle of, 384
Acids, 385
and bases, neutralization, 1006
strength measurement, 1004
salts, reactions, 1002
basicity, 1002
binary, 387
theory, 404
constitution theories, 402
Graham's theory, 402
hydro-, 386
ion theory, 1000
Laurent and Gerhardt's theory, 404
Liebig's theory, 403
oxy-, 386
■ oxygen theory, 385
source of acidity, 384
strength of, 1003
strong, 981
ternary, 387
luiitary theory, 404
weak, 981
Acidum pingue, 384
Active oxygen, 925
valency, 207
Activity of colloids, 777
, optical, 608
Adhesion, 821
Adiabatic compression gases, 863
elasticity, 820
expansion gases, 863
Adsorption, 311
iEolotropic crystals, 610
■ solids, 820
Aero, 122
VOL. I.
iEther, 33
Affini-valencies, 225
Affinities, neutral, 213
Affinity, 205. 785
and electromotive force, 1012
chemical, 291, 1011
constant, 296
Davy's electrical theory, 398
elective, 223
hygroscopic, 81
measurement, 294
of degree, 205, 223, 224
of kind, 205
pressure, 235
reciprocal, 298
selective chemical, 786
tables, 297
units, 224
Agricola, G., 51
Air, 61, 122, 123
(element), 32
fire, 344
inflammable, 125
phlogisticated, 125
preservation liquid, 873
• ■ pressure of, 149
respirable, 69
solubility of, 534
vital, 69
vitiated, 344
weight of, 143
Albertus Magnus, 46
Alchemy, 49
in China, 23
Alcogel, 771
Alcohols, 389
Alcosol, 771
Alkahest, 50
Alkali halides, 579
history, 382
salts, catalysis by, 487
Alkalimetry, 391
Alkalinity, principle of, 384
Al-Khazini, 42
AUotropism and heat of reaction, 700
Aluminium, Eka, 261
solubility of hydrogen, 306
X-radiogram, 642
Alums, X-radiograms, 642
Amalgam, lead, 3
Amicrons, 770
Amidopropionic acid and hydrogen, 304
Ammonia, effect on catalysis, 487
Ammonium chloride and hydrogen, 302
iodide, X-radiogram, 642
salts, 919
sulphide, effect on catalysis, 487
1041
3 X
1042
INDEX
Amonton's law, 160
Ampere, 963
Amphoteric oxides, 394
Amyl acetate and hydrogen, 304
alcohol and hydrogen,- 303
Analysis, 91
ionic hypothesis, 1009
Anaxagoras, 32
Anaximenes, 32
Angle of optical extinction, 608
Angles, axial, 615
of crysttils, interfacial, 593
Anhydrides, 395, 396
acid, 396
basic, 397
Anhydrite, X-radiogram, 642
Aniline €uid hydrogen, 304
Anion, 92
Anisotropic crystals, 610
liquids, 645
Anode, 92
Anthropomorphical chemistry, 2
Anticatalysts, 938
Antimony, solubility of hydrogen, 306
Antozone, 899
Apatite, X-radiogram, 642
Apparent equilibrium, 715
Applied chemistry, 11
Aragonite, X-radiogram, 642
Arc, high-tension, 882
low-tension, 882
Archimedes, 36
Architecture of crystals, 616
Aristotle, 30, 36
Arithmetic, chemical, 202
Arnold ViUanovanus, 47
Artiads, 208
Aryans, 20
Associated liquids, 856
Association of liquids, 858, 860
Atmolysis, 342
Atmosphere, 147, 148
extent of, 150
pressure of, 149
Atom, 103, 187
volume, 188
Atomic CO- volume, 240
heat. See Heat, atomic.
— — heats, effect of state of aggregation, 803
motion, 783
source of, 785
theory, 103
Boscovich's piuictual. 111
history of, 105
Lucretius', 106
volume, 259
volumes, 228
weights, 104, 180, 181, 198, 199
and Dulong and Petit's rule, 804
and isomorphism, 668
molecular heat, 807
volumes, 763
vmit of. 200
Atomicity, 224
Atoms, 740
Dalton's, 177
distance apart in molecules, 783
energy of, 785
individuality in molecules, 782
kinetic theorv, 782
Atoms, motion in molecules, 783
primitive, 225
vibration frequency, 828
weighing, 179
weights of, 179
Attraction, intermolecular, 625, 755, 822,
841
molecular, 865
Aura, 122
electrica, 877
tonante, 137
Autoclave, 437
Autoxidation, 925
Available energy, 717
Averroes, I. R., 42
Avicenna, E. S., 41
Avogadro's constant, 753
for colloids, 778
hypothesis, 172
and kinetic theory, 748
solutions, 545
190
Berzelius on, 187
Cannizzaro on, 191
deviations from, 192
Dumas on, 189
Gaudin on, 190
history of, 1 86
Gerhardt and Laui'ent on,
W. Prout on, 190
WoUaston on, 187
Axes, crystal, 614
of symmetry, 614
optic, 607
topic, 656,
Axial angles, 615
Azeotropic mixture, 550
Azote, 69
B
Babo's ozonizer, 885
Back electromotive force, 1029
Bacon, Roger, 46 •
Bar, 149
Barium, action on water, 135
chloride and hydrogen, 303
nitrate, X-radiogram, 642
peroxide, action of heat, 356
salts, catalysis by, 487
Base, acidifiable, 385
history, 382, 383
Bases, 393
and acids, neutralization, 1006
strength measurement, 1004
salts, reactions, 1002
ion theory, 1001
strength of, 1003
strong, 981
weak, 981
Basic anhydrides, 397
Basicity, 224
acids, Ostwald and Walden's rule, 1002
of acids, 389
Becker, J. J., 64
Belonites, 628
Benitoite, X-radiogram, 642
Bernoulli's equation, 744
Berthelot's law limiting density, 196
Bertollides, 519
INDEX
1043
Beryl, X-radiogram, 642
Beryllium, solubility of hydrogen, 306
Berzelius' electrochemical theory, 399
Biaxial crystals, 607
Biblical chemistry, 28
Biref ringent liquids, 645
Bismuth, solubility of hydrogen, 306
Bivariant systems, 447
Blagden's law, 516
Blood and hydrogen, 304
Boehme, J., 48
Boiling, 436
constant, 562, 564
curve, 167
point, 436, 438
absolute, 165
determination, 563
• — ■ Beckmann's process, 563
effect, volatility of solvent,
565
Landsberger's process, 564
and molecular weight, 561
osmotic pressure,
vapour pressure,
568
561
points colloids, 774
Boltzmann's constant, 809
distribution theorem, 792
Bonus, P., 48
Boron, eka, 261
Boscovich's theory of matter, 112
Bose's swarm theory, liquid crystals, 649
Bound energy, 716
Boyle, R., 52, 53
Boyle's law, 151
and kinetic theory gases, 742
solutions, 543
deviations, 152
• — • — • effect of molecular weight on, 194
British thermal unit, 699
Brodie's ozonizer, 886
Bronze age, 19
Brownian movement, 775
Bulk modulus, 820
Bumping, 453, 847
Bunsen's dichromate cell, 1028
• nitric acid cell, 1028
Burning, 59
Butyl(iso) alcohol and hydrogen, 303
Byzantium. See. Constantinople.
C
Cadmium, 521
mercury, 520
solubility of hydrogen, 306
Cailletet and Mathias' law, 169
Calcination, 55, 68
Calcite, X«radiogram, 641
Calcium, action on water, 135
■ chloride and hydrogen, 303
light, 326
sulphite, 520
Calcopyrite, X-radiogram, 642
Calor coelestis, 309
Calorie, 693, 698, 699
■ big, 699
gram, 699
kilogram, 699
pound, 699
Calx, 55
Capacity factor of energy, 712
Capillary electrometer, 1016
Caput mortuum, 55
Carbon atom, tetraliedron, 214
disulphide and hydrogen, 304
effect ou catalysis, 487
Carborundum, X-radiogram, 642
Camot's equation, 720
principle, 713
Cassiterite, X-radiogram, 641
Catalysis, 325, 357, 936
by contact, 486
mechanism of, 488
negative, 358
Catalysts, 937
negative, 938
poisoning of, 937
Catalytic reactions, 358
Cathode, 93
Cation, 93
Cause, 13, 57
Cellular structure metals, Quincke's theory,
603
Centibar, 150
Centre of symmetry, 614
Cerium, solubility of hydrogen, 307
Chaldea, 20
Chancourtois' telluric screw, 253
Characteristic equation, 161
Charcoal absorption, oxygen, 371
adsorption of hydrogen, 310
Charles' Law, 158
and Kinetic theory, 747
solutions, 545
deviations, 162
effect molecular weight on, 194
Charnock, T., 48
Chemical action, polar theory, 397
affinity, 1011
change, 83
combinations, 658
constant, 434, 737
composition and refractive index, 677
— surface tension, 853
■ energy, 1011
equilibria, 730
-, effect of temperature, 732
equivalent, 964
intensity, 1011
mixtures, 658
potential, 1011
reaction, work, 730
Chemistry, 3 -dimensional, 213
anthropomorphical, 2
— — applied, 1 1
Arabian, 40
Aryan, 20
Biblical, 28
Byzantium, 38, 39, 44
Constantinople, 44
Chaldean, 20
Chinese, 22
Egypt, 24
Grecian, 29
Hindu, 22
history of, 1
Indian, 21
language, 114
mythological, 2
1046
INDEX
Disperse phase, 769
Dispersion and refractive index, 677
atomic, 673
degree of, 769
mediiim, 769
molecular, 673
specific, 673
Dispersive power, 673
molecular, 673
specific, 673
Dispersoid system, 772
Dispersoids, 770, 772
ionic, 773
molecular, 773
Dissipation of energy, 704, 711
Dissociation, 492, 707
in solution, 570
pressure, 348
Distance energy, 712
Distillation, 553
in vacuo, 437
with reduced pressure, 437
Distortion of crystals, 598
Distribution, colloidal particles, 776
■ of molecular velocities, 792
Boltzmann's theorem, 792
Maxwell's theorem, 792
Dobereiner's triads, 253
Dolomite, X-radiogram, 641
Double refraction, 607
Drummond's light, 326
Drying gases, 288
Duhem and Margule's vapour pressure law,
555
Dulong and Petit's constant, 809
• law, 798
and atomic weights, 804
quantum theory of
energy, 811
meaning of, 808
Dumas' process vapour density, 184
Duralumin, 279
Dyad, 224
Dyads, 206
Dyne, 692
£
Earth (element), 31
inflammable, 64
mercurial, 64
Earths, history, 383
Eau oxygenic, 936
Ebers' papyrus, 26
Ebullition. See Boiling.
Effect, 13
Efflorescence, 81, 502
Effusion gases, 342
Egypt, 24
Einstein's theory, atomic heat, 811
Eka-aluminium, 261
boron, 261
silicon, 261
Elastic constants and isomorphism, 657
limit, 819
Elasticity, 819
adiabatic, 820
cubic, 820 i
isothermal, 820
Elasticity, longitudinal, 820
modulus 820
volume, 820
Electric acid, 137
discharge, glow, 882
invisible, 881
non-luminous, 881
silent, 882
Electrical and thermal energy, relation,
1036
conduction, velocity of, 967
discharge, 881
brush, 882
dark, 882
energy, 712
flame, 882
pressure, 963
resistance, 963
theory chemical action, 398
imits, 963
Electricity, 89
quantity of, 963
Electroaffinity, 1000, 1015
Electrochemical equivalent, 964
series, 1013, 1014
Electrochemistry, 711
Electrode, 92
potential, 1016
Electrolysis, 92, 962
Clausius' ionization hypothesis, 971
effect of solvent, 968
Faraday's laws, 963
fractional, 1039
Grotthus' chain hypothesis, 969
Helmholtz's strain hypothesis, 971
ion hypothesis, 969
of water. Bell cells, 278
diaphragm cells, 278
filterpress cells, 277
tank cell, 278
Electrolyte, 92
Electrolytes, Hall effect, 982
Electrolytic gas, 137, 483
solution pressure, 1017
Electrometer, capillary, 1016
Electromotive force, 963
and chemical affinity, 1012
osmotic pressure, 1020
back, 1029
Element, 74
Elementi primi, 60
secundi, 60
tertii, 60
Elements, Anaxagoras, 32
Anaximedes, 32
Aristotle, 33
bridge, 257
classification, 249, 263
distribution of, 272
Empedocles, 33
electrochemical series, 1013
extinct, 257
four, theory of, 33
five, theory of, 33
group, 257
Heracleitus, 32
Heterologous, 267
missing, 261
multivalent, 267
naming, 114
INDEX
1047
Elements, occurrence and periodic law, 272
Pherecydes, 31
Thales, 31
transition, 267
twin, 266
typical, 257
Elixir of life, 49
vitae, 49
Empedocles, 33
Empirical facts, 8
Emptiness, optical, 768
Tyndall's test, 768
Emulsoids, 770
Emulsions, 769
Endosmosis, 539
Endothermal compounds, 707
Endrometer, volta, 144
Enantiomorphism, 596
Energetic hypothesis of matter, 691
Energetics, first law of, 693, 694
, second law, 713
Energy, 688, 689
atomic, 785
available, 717
bound, 716
capacity factor, 712
chemical, 1011
conservation matter and, 695
cost of reaction, 716
degradation and entropy, 726
of, 711, 712
dissipation of, 704, 711
• distance, 712
electricity, 712
factors of, 712, 1011
forms of, 688
free, 716
and entropy, 726
intensity factor, 712
interval, 695, 717
of gases, 792
kinetic, 696, 712
energy of gases, 744,
persistence, 692
latent, of reaction, 728
• transformation, 689
law of conservation, 692
mass factor, 712
non-productive, 721
potential, 696, 727
• quantity factor, 712
quantum theory, 811
relation of electrical and thermal,
1036
stability function, 727
strength factor, 712
surface, 712, 846, 847
total, 717
transformations of, 689
units of, 693 -•
volume, 712
Enstatite, 521
Entropic series, 654
Entropy, 721
analogus, 723
and degradation of energy, 726
diffusion, 725
free energy, 726
law of maximum, 725
measurement, 722
Eotvos' rule, 855
Epidote, X -radiogram, 642
Episomorphs, 662
Equation building, 361
characteristic, 161
gas, 161, 754
Clausius', 761
Dieterici's, 758
Van der Waals',
756
of state, 161
of solids, 834
state solids, Guldberg's, 836
Van der Waals', 836
Equations, chemical, 202
Equilibrium, apparent, 715
chemical, 730
effect of temperature, 732
conditions of, 445, 714
effect of temperature on chemical,
732
false, 775
metastable, 716
pressure, 348
stable, 714
Equivalent, 187
chemical, 964
electrochemical, 964
weights, 79, 99
Erbium, solubility of hydrogen, 307
Erg, 692
Error, probable, 131
Etch figures, 611
Ether, solubility in water, 523
Etherine theory, 217
Ethers, 389
Ethyl acetate and hydrogen, 304
■ alcohol and hydrogen, 303
Ethylene, effect on catalysis, 487
ozonide, 899
Europium, solubility of hydrogen, 307
Eutectics, 517
Eutectoid, 518
Eutexia, 517
Evaporation, cooling during, 426
kinetic theory, 425
speed of, 424
Evidence, circumstantial, 90
cumulative, 90
negative, 83
Evolution, chemistry, 119
nomenclature, 119
Exosmosis, 539
Exothermal compounds, 707
Expansion and isomorphism, 658
coefficient and heat fusion, 837
gases, thermal effects, 862
thermal, of colloids, 774
Experience, 5
Experiment, 5, 12
Experiments, blank, 67
control, 57
dummy, 57
Explosion wave, velocity of, 486
Explosions, 485, 705
External work, 695
Extinction, angle of optical, 608
oblique, 608
parallel, 608
straight, 608
Extraordinary ray,' 607
1048
INDEX
Factor, capacity of energy, 712
intensity of energy, 712
mass of energy, 712
quantity of energy, 712
strength of energy, 712
Factors of energy, 712, 1011
Facts, 5
empirical, 8
False equilibrium, 715
Farad, 963
Faraday's laws, electrolysis, 963
Fedoroff's crystallochemical analysis, 616
Ferments, inorganic, 937
Ferrous ammonimn sulphate, electrolysis,
962
gas, 123
Fick's law of diffusion, 536
Figures, corrosion, 611
etch, 611
interference, 610
Fire, 55, 59
air, 344
astral, 64
(element), 32
elemental, 64
matter, 384
sacred, 59
Flame, 56, 61
electrical, 882
musical, 127
philosopher's, 126
Flamel, N., 48
Flatus, 61, 122
Fluorspar, catalysis by, 487
X-radiogram, 640
Force, 689
Formula weight, 179
Formulae, chemical compounds, 223
constitutional, 206
empirical, for properties of solids, 834
graphic, 206
minerals, 668
mixed crystals, 668, 670
of compounds, 179
structural, 206
Fractional electrolysis, 1039
Free energy, 716
and entropy, 726
path of molecules, 748
Freezing constant, 566
curves, 519
point and molecular weight, 565
and osmotic pressure, 568
vapour pressure, 565
colloids, 774
determination, 567
Beckmann's process, 567
pressure, 457
temperature, 457
Fusion curve, 445
heat of, 426
G
Gadolinium, solubility of hydrogen, 307
Galen, C, 38
Galilei, Gallileo, 47
Gallium, solubility of hydrogen, 307
Garnet, X-radiogram, 642
Gas, 122
analogy hypothesis, osmotic pressure,
557
analysis, 144
and vapour, 435
cells, 1033
constant, 161
cuprous, 123
detonating, 137
electrolytic, 137
equation, 161, 754
ferrous, 122
fuliginosum, 122
laws and osmotic pressure, 543
pingue, 122
sicum, 122
sylvestre, 122
Gases, Are molecules alike ? 342
density, 175
diffusion, 338
drying, 288
effusion, 342
equilibrium, 152
kinetic theory, 742
liquefaction, 868
molecular heat, 795
of, effect of pressure, 796
effect of tempera-
ture, 796
permanent, 869
refractive index, 681
separation by diffusion, 341
— — solubility, and volume of solvent, 527
effect of pressure, 529
salt solutions, 535
of mixed, 533
specific gravity, 175
heat, constant pressure, 786, 787
constant volume, 786, 787
■ thermal effects, compression, 862
expansion, 862
two specific heats, 786
Gay Lussac's law, 171
Geber, 40
Latin, 40
Pseudo, 40
Gel, 771
Germanium, 261
Gibbs' phase rule, 444, 446
Gibbs and Helmholtz's equation, 1038
Glace-du-fond, 464
Glaser, C, 52
Glass, catalysis by, 487
permeability to gases, 305
permeability to oxygen, 371
solubility of hydrogen, 309
Glauber, J. R., 52
Globulites, 628
Glucose and hydrogen, 304
Glycerol and hydrogen, 304
Glycocol and hydrogen, 304
Glycozone, 946
Gold, catalysis by, 487
diplosis of, 49
palladium alloys. See Palladium.
solubility of hydrogen, 305, 306 '
X-radiogram, 641
Goldschmidt and Wright's law, 612
INDEX
1049
Graham's diffusion law and kinetic theory,
744
law of diffusion, 340
Gram-calorie, 699
molecule, 392
Graphite, X -radiogram, 642
Gravitation, 292
Gravity, 786
acceleration of, 693
Greece, 29
Grotthus' chain hypothesis, electrolysis, 969
Groups of elements, 255
Grove's cell, 1028
Growing face of crystals, 629
Growth of crystals, 623
Griineisen's formula, 834
Guldberg and Waage's law, 300
Guldberg's equation of state for solids, 836
Gypsum, X -radiogram, 642
H
Habit of crystals, 597, 598
prismatic, 597
tabular, 597
Haematite, X-radiogram, 642
Halitus, 122
Hall effect with electrolytes, 982
Hambergite, X-radiogram, 642
Hardness and isomorphism, 657
Harmonicon, chemical, 127
Hauerite, X-radiogram, 641
Haiiy's law, 594
• rational indices, 616
Heat, atomic, 798, 811, 812, 813
• and atomic weights, 804
Debye's theory, 815
effect of pressure, 799
temperature, 801
Einstein's theory, 811
fusion and coefficient expansion, 837
vibration frequency, 833
mechanical equivalent, 693
molecular, 805
of gases, 795
effect of pressure, 796
— temperature,
796
of combustion, 710
• fusion, 426
— — ■ • and freezing point, 440
ionization, 1007
■ reaction, 698
and allotropism, 700
• isomerism, 900
in solution, 700
temperatiure coefficient, 702
solution and osmotic pressure, 547
- vaporization, 426
and surface tension, 851
external, 427
internal, 427
specific and surface tension, 852
Debye's theory, 815
- gases, constant pressure, 786, 787
volimie, 786, 787
of molecules, 832
solids, 798
Heat, work value of, 719
Heating curve, 450
curves, 518
Heats, molecular, and atomic weights, 807
Helmholtz and Gibbs' equation, 1036
Helmholtz's equation, 720
double layer, 1016
strain hypothesis, electrolysis, 971
Helmont, J. B. van, 61
Hemihedral symmetry, 613
Hemimorphite, X-radiogram, 642
Henry's law, kinetic theory of, 531
solution of gases, 527
Heracleitus, 32
Hermes Trismegistus, 24
Hero, 37
Hesiod, 19, 31
Hess' law of heat of reaction, 708
thermoneutrality, 1007
Hexagonal system, 617
Hippocrates, 32
History of chemistry, 1
kinetic theory, 767
Hittorf's transport numbers, 985
Hoar-frost curve, 444
Hofmann's process, vapour density, 185
Holmium, solubility of hydrogen, 307
Holohedral symmetry, 613
Homoeomeriae, 33
Homogeneous substances, 86, 96
Homomorphism, 663
Hooke's law, 819
Hydracids, 386
Hydrate ozone, 908
Hydrated salts, 498
salt, 397
vapour pressure, 501
Hydrates, 397, 498
Hydrates, distinction hydroxides, 499
Hydrides, 326
and periodic law, 328
Hydrochloric acid and hydrogen, 303
Hydrogel, 771
Hydrogen, 264
action on oxides, 328
salt solutions, 328
activated, 321, 322
auto -combustion process, 282
atomic, 336
magnetism, 322
refraction, 316
volume, 313
weight, 335, 380
boiling point, 315
by-product, 286
calx, 128
combustibility, 325
compressibility, 314
critical pressure, 315
• temperatiu*e, 316
volume, 316
theorem, Nemst's, 735
vaporization and boiling point, 440
degree ionization, 320
density, 313
detection, 334
determination, 334
dielectric constant, 322
discharge tension, 319
discovery, 125
sulphide, effect on catalysis, 487
electrode, 320
1050
INDEX
Hydrogen, entropy, 316
free energy ionization, 32 1
from decomposition water, 278
from metal hydrides, 283
metals and acids, 282
■ alkalies, 283
heat combustion, 489
ionization, 321
index of refraction, 316
ionization of gas, 319
ionizing potential, 319
latent heat fusion, 316
vaporization, 316
magneto-optic rotation, 316
magnetic susceptibility, 322
melting point, 316
molecular heat, 315
rotation, 316
molecule, collision frequency, 313
diameter, 313
dissociation, 335
free path, 313
number per c.c, 313
volume of, 313
molecules, velocity of, 313
nascent, 331
occurrence, 270
overvoltage, 333
ozonized, 321
permeability of metals, 304
peroxide, 277
action, alcohols, monohydric, 946
polyhydric, 946
alkali bromides, 940
chlorides, 940
alkaloids, 946
aluminium, 942
ammonia, 94
animal extracts, 938
antimony, 941
sulphide, 941
arsenic, 941
benzene, 946
bismuth, 941
nitrate, 941
sulphide, 943
blood, 938, 946
bromic acid, 940
bromine, 939
cadmium hydroxide, 943
carbon, 942
dioxide, 946
carbonyl chloride, 946
— — catalase, 938
• cerium oxide, 943
salts, 942
chlorates, 939
chloric acid, 940
chlorine, 939
chromic oxides, 944
cobalt hydroxide, 943
copper, 941
cupric hydroxide, 943
• salts, 943
diastase, 938
didymium oxide, 943
enzymes, 938
ferrous salts, 943
fibrin, 946
gallic acid, 946
Hydrogen peroxide, action, glycerol, 946
glycol, 946
gold, 941
oxide, 942
guaiacum, 946
haemoglobin, 938
hydriodic acid, 939
hydrogen bromide, 939
chloride, 939
selenide, 941
sulphide, 941
hydroxylamine sulphate,
941
hypochlorous acid, 939
hyposulphites, 941
iodates, 940
iodic acid, 940
indigo, 946
iodine, 939
iron, 943
sulphide, 943
lanthanum oxide, 943
lead, 941
dioxide, 943
monoxide, 943
sulphide, 943
magnesiiun, 941
hydroxide, 943
manganese compounds, 944
mannite, 946
mercuric oxide, 943
mercury, 941, 942
sulphide, 943
metals, 941
milk, 938
molybdenum, 943
salts, 942
sulphide, 943
nickel, 941
hydroxide, 943
nicotine, 946
nitric oxide, 941
nitrous acid, 941
oxalic acid, 946
perchlorates, 939
perchloric acid, 940
periodates, 940
periodic acid, 940
— — phenyl carbonate, 946
phosphorus, 941
platinum, 941
potassium cyanide, 942, 946
ferricyanide, 943
fluoride, 940
■ iodide, 940
pyrogallol, 946
quinine, 946
samarium oxide, 943
selenium, 941
serum, 946
silver, 941, 942
carbonate, 943
chloride, 940
• dioxide, 942
nitrate, 942
oxide, 942
peroxynitrate, 942
sulphide, 943
sodium periodate, 940
stannous salts, 943
INDEX
1051
Hydrogen peroxide, action, starch, 938
sugars, 946
sulphides, 941
sulphuric acid, 941
sulphurous acid, 941
tannin, 946
tellurium, 941
dioxide, 941
tetrathionates, 941
thallium oxide, 943
tin, 941
sulphide, 943
titanium salts, 942
tungsten, 943
salts, 942
uraniiun salts, 942
vanadic acid, 942
vanadium salts, 942
vegetable extracts, 938
water, 939
white of egg, 946
yttrium oxide, 943
zinc hydroxide, 943
■ oxide, 943
zirconia, 943
boiling point, 929
catalytic decomposition,
938
iridium, 938
Hydrogen peroxide, solubility, ethyl, ]
valerianate, 932
in petroleum ether, 932
isobutyl alcohol, 932
butyrate,
i isoamyl propionate, 932
■ nitrobenzene, 930
phenol, 932
propyl butyrate, 932
formate, 932
quinoline, 932
sodium carbonate, 932
in water, 932
specific gravity, 929
heat, 929
surface tension, 929
tests, 961
thermochemistry of, 931
uses, 946
solubility in acetamide, 304
acetic acid, 303
a,cetone, 304
amidopropionic acid, 304
— ammonium chloride, 303
boron.
manganese dioxide, 938
minerals, 938
— palladium, 938
. platinum, 934, 938
wood charcoal, 938
— chemical properties, 936
— colour, 929
— composition, 952
— concentration of solutions, 927
— constitution, 952
— decomposition, action pressure,
938
catalytic, 934, 936
in light, 933
— dielectric constant, 931
— dihydrate, 939
— electrical conductivity, 931
— fractional distillation, 927
— free energy, 930
— heat of formation, 930
neutralization, 939
solution, 930
vaporization, 934
- higher, 945
- history, 877
- index of refraction, 931
- melting point, 929
- monohydrate, 939
- occurrence, 891, 892
- partition coefficient with organic
solvents, 932
- physical properties, 929
- preparation, 922
quantitative determination, 949
solubility acetophenone, 932
amyl acetate, 932
aniline, 932
benzene, 932
chloroform, 932
ether, 932
ethyl acetate, 932
amyl acetate, 304
alcohol, 303
aniline, 304
barium chloride, 303
blood, 304
calcium chloride, 303
carbon disulphide, 304
chloracetic acid, 303
ethyl acetate, 304
alcohol, 303
glucose, 304
glycerol, 304
glycocoU, 304
hydrochloric acid, 303
isobutyl acetate, 304
alcohol, 303
lithium chloride, 303
magnesium sulphate, 303
in metals, 305
methyl alcohol, 303
nitrobenzene, 304
petroleum, 304
potassium carbonate, 303
chloride, 303
hydroxide, 305
nitrate, 303
' propionic acid, 303, 304
. serum, 304
sodium carbonate, 303
chloride, 303
hydroxide, 303
nitrate, 303
sulphate, 303
sugar, 304
sulphuric acid,. 303
toluene, 304
urea, 304
water, 301, 302
xylene, 304
zinc sulphate, 303
preparation, 125, 275
properties, 126
purification, 275, 287
reducing power, 332
sical process, 279
silicol process, 284
1052
INDEX
Hydrogen, specific gravity, 313
heat, 315
ratio two, 315
spectrum, 317
absorption, 319
Balmer's series, 318
Lynman's series, 318
Paschen's series, 318
Stark effect, 318
storage, 288
surface tension, 314
thermal conductivity, 314
expansion, 314
triple point, 316
valency, 335
vapour pressure, 315
Verdet's constant, 316
viscosity, 313
weight of atoms, 313
Htre, 313
Zeeman effect, 318
Hydrogenite, 286
Hydrogeniimi, 309
Hydrol, 461
Hydrolith, 283
Hydrolysis, 391, 496, 1009
Hydrone, 279
Hydroperoxide, 956
Hydrophile, 771
Hydrophobe, 771
Hydrosol, 771
Hydrotropism, 493
Hydroxides, 396
• distinction hydrates, 499
Hydrozone, 946
Hygroscopicity, 81
Hy) Jtropic mixture, 656
Hypereutectic, 618
Hyperol, 932
Hypo-, 118
Hypoeutectic, 518
Hypothesis, 13
Hypotheses, 57, 58, 59
rival, 16
verification, 15, 30
Hysteresis, 162
latro-chemistry, 50
Ice, anchor, 464
bending moment, 466
bottom, 464
crystalloluminescence, 465
curve, 445
elasticity, 466
flow of, 466
frazil, 464
friction, 467
ground, 464
hardness, 466
plasticity, 466
See Water, 435
sheet, 464
• slush, 464
X-radiogram, 465
Yoimg's modulus, 466
Ideal crystals, 598
Idiomorphs, 696
Igneous corpuscles, 56
Ignis caelestis, 64
subtilis, 64
tenuis, 64
Ignition temperatures, 485
Imagination in chemistry, 9
Impure substances, 80, 82
Indefinite compounds, 658
Index of crystals, 615
refraction, 670, 671
and specific gravity, 672
India, 21
Indiarubber, permeability to gases, 309
• oxygen, 371
Indicator, 389
Indices of crystals, 615
rational, Haiiy's law, 615
Indium, solubility of hydrogen, 307
Induction, 17
period, 296
Inert gases, 263
Inoculation solutions, 451
Inorganic ferments, 937
Insoluble substances, 508
Intensity, chemical, 1011
factor of energy, 712
Intercrystalline cement, 605
Interference figures, 610
Intermediate oxides, 394
Intermolecular attraction, 525
Internal energy, 695, 717
■ ■ of gases, 792
friction, 749
pressure, 841
work, 695
Intrinsic pressure, 841
and latent heat, 843
solubility, 852
• surface tension, 842
liquids, 841
Invariant systems, 446, 447
Inversion temperature, 866
Iodine, 264
Ion, 93, 965
hypothesis, electrolysis, 969
theory acids, 1000
bases, 1001
precipitation, 996
solubility, 995
unit charge, 965
Ionic dispersoids, 773
hypothesis, analysis, 1009
Ionization, 971
and osmotic pressure, 990
collision hypothesis, 973 •
constant, 992
dielectric hypothesis, 974
heat of, 1007
mechanism of, 973
modes of, 991
percentage, 981, 992
solvent attraction hypothesis, 974
Ions, concentration, 981
effect, hydration on speed, 989
migration of, 983
number in solution, 978
strong, 1015
weak, 1015
Iridium, catalysis by, 487
INDEX
1053
Iron, action on water, 1354
age, 19
and steam, 297
catalysis by, 487
■ solubility of hydrogen, 305, 306
-vanadium, 520
X-radiogram, 642
Irreversible cells, 1022
colloid, 771
processes, 717
Isaac of Holland, 48
Isobutyl acetate and hydrogen, 304
Isodimorphism, 664
Isogonism, 663
Isomerism and heat of reaction, 700
refractive index, 685
Isomorphism, 661
• and atomic weights, 668
cleavage, 657
corrosion figures, 658
• elastic constants, 657
• hardness, 657
• — — magnetic properties, 658
optical properties, 658
specific gravity, 657
— — thermal conductivity, 658
expansion, 658
Mitscherlich's law, 651, 652
Isomorphous mixtures, 658
Isothermal compression gases, 863
elasticity, 820
expansion gases, 863
Isotonic solutions, 539
Isotropic crystals, 610
solids, 820
Joule, 693
Kelvin effect, 864, 866
Thomson effect, 864, 866
Joule's law, 864
rule, 805
K
Kanaka, 22
Kelley, E., 48
Kelvin's equation, 1038
— rule, 1037
Kepler Johann, 47
Kilogram-calorie, 699
Kinetic energy, 696
• gases, 744
theory and Avogadro's hypothesis, 748
Charles' law, 747
■ Dalton's law, partial pres-
sures, 744
diffusion, 744
solution, 524, 528
atoms, 782
history, 767
gases, 742
• and Boyle's law, 743
hquids, 840
moleciJes, 765
of Henry's law, 531
solids, 819
Kirchhoff's equation, 702
Knowledge, empirical, 8
, scientific, 8
Kohlrausch's conductivity equation, 987
law, 987
law, 979
Kunckel, J., 62
Labile states, 454
Laevorotatory, 608
Lamp, perpetual, 50
Langmuir's theory, liquids, 642
solids, 642
Lanthanum, solubility of hydrogen, 307
Laplace's constant, 841
Latent energy of reaction, 728
heat and intrinsic pressure, 843
Lattice, clinorhombic prism, 626
cubic, body -centred, 626
face-centred, 625
double, 626
■ simple, 625
hexagonal prism, 626
monoclinic parallelepiped, 626
rectangular prism, 626
body -centred, 626
rhombic prism, 626
' body -centred, 626
rhombohedron, 626
space, 624
squaie-prism, 626
120°, 626
face-centred, 626
triclinic, 626
Laue's spots, 634
X-radiograms, 634
Lavoisier and Laplace, law of, 698
Law, 10, 13, 31
continuity, 14
of chemical composition, 95
compound proportion, 100
• — constant composition, 76, 78
• Dalton, 93
— • definite proportion?, 77
■ — equivalent ratios, 79
• indestructibility of matter, 101
mass action, 933 —
— multiple proportions, 93, 96
• persistence of weight, 101
• — proportionality, 79
reciprocal proportions, 97
— three states, 1
Proust's, 76
Richter's, 79, 97
Laws, 157
Lead amalgam, 3
nitrate, X-radiogram, 642
solubility of hydrogen, 306
X-radiogram, 641
Leclanche's cell, 1029
Leduc's molecular volimae method, molecu-
lar or atomic weights, 763
Lemery, N., 52
Leonardo da Vinci, 47
Leyden payprus, 26
Libavius, A., 51
Liesegang's rings, 637
1054
INDEX
Light, calcium, 326
Dnunmond's, 326
lime, 326
matter of, 66
zircon, 326
Lignit«, absorption oxygen, 371
Lime-light, 326
Liquefaction gases, 868
by cooling, 870
^- Joule-Thomson effect, 872
— rapid evaporation, 871
cascade method, 871
Liquid air. See Air, liquid.
crystals, 645
Bose's swarm theory, 649
Liquids, anisotropic, 645
associated, 856
association of, 858, 860
birefringent, 645
intrinsic pressure, 841
kinetic theory, 840
Langmuir's theory, 642
normal, 856
polymerized, 860
solubility in liquids, 522
Litharge, absorption oxygen, 371
Lithium chloride and hydrogen, 303
solubility of hydrogen, 307
X -radiogram, 642
Longitudinal elasticity, 820
Longulites, 628
Lucretius, 19, 37
atomic theory, 107
LuUy, Raymond, 407
Lumen philosophicum, 126
Luminescence, crystallo-, 601
tribo-, 601
Lyophile, 771
Lyophobe, 771
M
Macromolecules, 657
Magnesium, action on water, 135
solubility of hydrogen, 306
sulphate and hydrogen, 303
■ X-radiogram, 642
Magnetic properties and isomorphism, 658
rotatory power and refractive index,
682
Magnetite, X-radiogram, 640
Magnitudes of molecules, 766
Magnus' nile, 1039
Maier, M., 48
Manganese, 520
— — dioxide, action, heat, 359
• salts, catalysis by, 487
• solubility of hydrogen, 306
sulphite, 520
Margarites, 628
Margules and Duhem's vapour pressure
law, 555
Mass action, law of, 933
active, 299
chemical, 299
factor of energy, 712
Materia coelestia, 60, 64
ignis, 64
subtilia. 61
Matter, 688
conservation energy and, 695
energetic hypothesis, 691
law of indestructibility, 101
molecular structure, 740
perdurability of, 100
weight of, 64
Maximum entropy, law of, 725
work, principle of, 703
Maxwell's distribution theorem, 792
Mayer's equation, 787
Measurement of entropy, 722
Mechanical equivalent of heat, 693
Medicine, universal, 49
Medico-chemistry, 50
Medium dispersion, 769
Megabar, 149
Melting point and solubility, 585
• surface tension, 852
Membrane, semipermeable, 539
MendeleefE's periodic law, 255
Mercuric bromide, 520
iodide, 520
oxide, action heat, 3, 347
Mercury, 521
absorption oxygen, 371
cadmium, 521
catalysis by, 487
(element), 34
hydride, 321
solubility of hydrogen, 307
Metal, 248
Metalloids, 248, 250
Metals, base, 248
cellular structure, Quincke's theorv^
603
influence of planets on, 3, 21
noble, 248
perfect, 248
permeability of hydrogen, 304
sacrificial, 1025
semi, 248
transmutation, 49
Metastable equilibrium, 715
states, 454
Methyl alcohol and hydrogen, 303
Meyer's process, vapour density, 185
Microbalance, 184
Microns, 769
Microscope polarizing, 608
Micro-weighing, 184
Migration of ions, 983
Miller's system, crystal notation, 614
Millibar, 150
Millimol, 392
Millival, 392
Mimetic twinning, 595
Minerals, formulae, 668
Mitscherlich's law, isomorphism, 652
Mixed crystals, 658
formulae, 668, 670
Kuster's rule, 660
— law of, 658
Retger's colour test, 659
Retger's law, 659
Mixing limit, 665
Mixture, eutectic, 517
law and refractive index, 678
Mixtures, 85
law of, 88
INDEX
1055
Models, molecular, 783
Modulus, bulk, 820
of elasticity, 820
sheer, 820
Young's, 820
Moisture efiect on catalysis, 487
Mol, 392
Molar weight, 176
Molecular attraction, 755, 822, 841
865
complexity and crystal form, 622
co-volume, 239, 755
dispersoids, 773
heat. See Heat, molecular.
magnitudes, 766
models, 783
motion, source of, 785
structure matter, 740
volume, affinity and, 233
chemical activity and, 237
compressibility and, 234
density and, 234
volimies, 176, 195, 228
and atomic weights, 763
molecular weights, 763
• Traube's theory, 233
weight, 202
and boiling point, 561
critical constants, 762
freezing point, 565
solubility, 568
vapour pressiu-e, 548
of colloids, 773
weights, abnormal, 569
and molecular volumes,
volumes, 201
763
ratio of two specific heats and.
788
Molecules, 174, 740
are all alike ? 342
■ average diameter, 752, 755
collision frequency, 761
elementaires, 173
free path, 748
■ integrantes, 173
kinetic theory, 765
number per c.c, 753
specific heat, 832
• velocity of, 744
vibration frequency, 828
weight of, 174
Molybdenum, solubility of hydrogen, 306
Monad, 224
Monads, 35, 111, 206
Monoclinic system, 621
Morphotropic series, 654
Morphotropy, 655
Motion, perpetual, 50
Motochemistry, 227
Motus caloris, 60
ignis, 60
Moufette atmospherique, 68
Mythological chemistry, 2
N
Nascent state, 331
Negative catalysts, 938
evidence, 83
Neodymium, solubility of hydrogen, 307
Neo-platonists, 39
Nephelene, X-radiogram, 642
Nernst's heat theorem, 735
vapour pressure formula, 434
Neumann's rule, 805
Neutral salts, 388
Neutrality, Richter's law of, 391
Neutralization, 389, 391
ion theory, 1007
of acids and bases, 1007
Newland's law of octaves, 252, 254
Newton, Isaac, 47
Nickel, 264, 520
ammoniimi sulphate, electrolysis, 962
catalysis by, 487
solubility of hydrogen, 306
X-radiogram, 642
Nitre, volatile, 56
Nitrogen, 69
■ manufacture from liquid air, 874
Claude's process,
875
Linde's process,
874
Nomenclature, chemistry, 114
chemist's, evolution, 119
Werner's, 209
Non-metals, 248
Nonproductive energy, 721
Non-valence, 206
Normal liquids, 856
salts, 387
Norton, T., 48
Notation crystals, Miller's system, 614
Nucleus theory, 218
Number of molecules per c.c, 763
Number, polar, 211
Numerical prefixes, 117
Obach's formula, 835
Oblique extinction, 608
Observation, 5
Occlusion, 306
Octaves, law of, 252, 254
Ohm, 963
One Thing, 48
Opalescence of gases, 166
critical, 166
Optic axes, 607
Optical activity, 608
constants and isomorphism, 658
emptiness, 768
TyndaU's test, 768
extinction, angle of, 608
Ordinary ray, 607
Osmium, catalysis by, 487
solubility of hydrogen, 307
tetroxide, solubility of hydrogen, 308
Osmosis, 639
negative, 541
positive, 541
reversed, 541
Osmotic pressure, 538
abnormal, 990
and boiling point, 568
chemical theoi-y , 570
1056
INDEX
Osmotic pressure colloids, 774
and concentration, 643
electromotive force, 1020
freezing point, 568
gas an^ogy hypothesis, 657
gas laws, 543
general formula, 652
heat of solution, 547
ionization, 990
. solubility, 669
solution pressure hypothesis, 668
surface tension hypothesis, 660
temperatiire, 545
theories of, 557
vapour pressure, 550
hypothesis, 668
pressures, abnormal, 570, 673
Ostwald and Walden's basicity rule, 1002
Ostwald's dilution law, 992
Overgrowths, 661
Overvoltage, 333
Oxidation, 64, 69, 117, 210
Oxide, 69
Oxides, 117, 374, 393
amphoteric, 394
heat of formation, 374
higher, 268
intermediate, 394
■ preparation, 374
Oxonium hydroxide, 920
salts, 919
Oxozone, 899
Oxy acids, 386
Oxygen, 69
absorption by solids, 370
allotropic forms, 366
atomic, 366
weight, 380
boiling point, 365
combustion, calcium, 374
charcoal, 374
in, 373
iron, 374
magnesitun, 374
phosphorus, 374
sodium, 374
sulphur, 374
pressure, 365
temperature, 366
critical volume, 365
crystals of, 366
detection, 380
determination, 380
diameter molecule, 363
dielectric constant, 369
diffusion coefficient, 371
discharge potential, 368
discovery, 344
dispersion, 366
electrode, 368
entropy, 365
free path, 363
index refraction, 366 %
ionizing potential, 368
latent heat fusion, 366
vaporization, 365
liquid, absorption, fluorine, 371
nitrogen, 371
magnetic moment, 369
susceptibility, 369
Oxygen, active, 926
quadrivalency, 919
manufacture from liquid air, 874
Claude's pro-
cess, 875
Linde's pro-
cess, 874
melting point, 366
number molecules in gas, 363
occurrence, 351
oxidization, potential, effect of hydro-
gen peroxide, 930
physiological effects, 378
preparation, 352
pv-curves, 364
rate of solution in water, 369
relative density, 363
solubility, 369
acetone, 370
acids, 369
ammonium chlorides, 370
barium chloride, 370
blood, 370
calcium chloride, 370
caesium chloride, 370
ethyl alcohol, 370
— ■ — • lithium chloride, 370
magnesium chloride, 370
methyl alcohol, 370
petroleum, 370
potassiiim bromide, 370
chloride, 370
cyanide, 370
hydroxide, 369
iodide, 370
nitrate, 370
sulphate, 370, 379
rubidium chloride, 370
sea-water, 370
sodium bromide, 370
chloride, 370
hydroxide, 370
sulphate, 370
sugar, 370
sulphuric acid, 369
water, 369
specific cohesion, 364
heat, 365
volume, 363
spectrimi, absorption, 368
spark, 367
Stark effect, 368
storage, 356
surface tension, 364
thermal conductivity, 365
expansion, 365
■ uses, 379
vapour pressure, 365
velocity of molecules, 363
of sound, 364
Verdat's constant, 367
viscosity, 364
weight of atom, 363
___ of litre, 363
Oxyhydrogen flame, 326
Ozobenzene, 899, 911
Ozobutylene, 899
Ozoethylene, 899
Ozonates, 908
Ozone, 277
INDEX
1057
Ozone, absorption spectrum, 895
action, alcohol, 911
alkali hydroxides, 908
alkaline earth hydroxides, 908
aluminium, 908
— ammonia, 907
aniline, 911
antimony, 907
arsine, 907
arsenic, 907
trichloride, 907
— arsenious oxide, 907
— benzene, 911
— bismuth nitrate, 910
— brass, 908
— bromme, 904
— carbon, 907
monoxide, 907
— chlorine, 904
— chromic salts, 911
— cobalt sulphate, 911
sulphide, 909
copper, 909
cork, 911
cupric salts, 910
dynamite, 911
ethyl peroxide, 911
ethylene, 911
ferric salts, 911
ferrochromium, 908
ferrocyanides, 911
ferrous salts, 910
fluorine, 904
gold, 908
chloride, 911
sulphide, 910
hydrazine sulphate, 907
hydrogen, 901
■ chloride, 904
fluoride, 904
halides, 904
peroxide, 903
sulphide, 905
iodine, 904
iron, 908
lead, 909
salts, 910
sulphide, 909
manganese dioxide, 910
sulphide, 909
manganic sulphate, 910
manganous salts, 910
merciirous salts, 910
mercury, 909
methane, 911
nickel, 909
nitrate, 911
sulphide, 909
nitric oxide, 906
nitrogen, 906
• chloride, 911
iodide, 911
tetroxide, 906
trioxide, 906
nitroglycerol, 911
palladium salts, 911
sulphide, 909
permanganates, 910
phenols, 911
phosphine, 907
Ozone, action, phosphorus, 907
pentachloride, 907
iodide, 907
pentabromide, 907
pentoxide, 907
tribromide, 907
— — trichloride, 907
platinum, 908
potassium carbonyl ferro-
cyanide, 911
iodide solutions, 904
• acid, 905
alkaline, 905
neutral, 904
rubber, 911
selenium, 906
silicochloroform, 908
silver, 909
sulphide, 909
sodium sulphide, 905
thiosialphate, 905
stannous chloride, 910
stibine, 907
sulphur, 905
dioxide, 905
trioxide, 906
sulphuric acid, 906
sulphurous acid, 905
tellurium, 906
thallous salts, 910
tin, 909
vegetable colours, 911
water, 903
zinc, 909
as oxidizing agent, 905-910
— reducing agent, 904
boiling point, 894
chemical properties, 901
colour, 894
composition, 914
constitution, 917
formula of, 918
free energy, 895
heat formation, 895
history, 877
hydrate, 908
luminescence, 901
occurrence, 891
physical properties, 893
preparation, 878
quantitative determination, 949
solubility, acetic acid, 897
anhydride, 897
carbon tetrachloride, 898
chloroform, 898
essential oils, 897
ethereal oils, 897
ethyl acetate, 897
fats, 897
in alkaline solutions, 897
in salt solutions, 897
in sulphuric acid, 897
in water, 896
action, acetaldehyde, 897
oxalic acid,
paraldehyde, 897
quinine salts, 897
stabilizing, 897
specific gravity, 894
heat, 895
VOL. I.
3 Y
1058
INDEX
Ozone, specific magnetization, 896
tests, 951
uses, 911
water, 898
Ozonic acid, 906, 908
Ozonides, 897, 899
Ozonizer, Babe's, 886
Brodie's, 886
Siemens', 886
Ozonons acid, 908
Ozonwasserstoff, 321
Ozozobutylene, 899
Ozozonides, 899
Palladiiim, absorption oxygen, 370
catalysis by, 487
gold alloys, solubility of hydrogen,
307
platinum, solubility of hydrogen, 307
silver aUoys, solubility of hydrogen,
307
solubility of hydrogen, 305, 306
Papin's autoclave, 437
digester, 437
Papyrus, Ebers', 26
Ley den, 26
Rhind, 26
Paracelsus, 50
Parallel extinction, 608
Parameters of crystals, 615
topic, 656
Particulae igniae, 56
nitro-aerae, 56
Passive resistance, 152
Per-, 118
Peracids, 956
and periodic law, 960
Perdurability of matter, 100
Perhydral, 946
Perhydrol, 932
Period of induction, 295
Periodic law and occurrence of elements,
272
graphic representation of, 260
Mendeleeff's, 255
misfits, 263
— occurrence of elements, 273
table elements, 256
Periods of elements, 255
long, 257
— short, 257
Perissads, 208
Permanent gases, 869
Peroxal, 946
Peroxide, 966
Peroxides, 394, 956, 968
and periodic law, 960
Perpetual lamp, 60
motion, 60, 693
law of excluded, 694
Persalts, 960
Persia, 20
Persulphurio acid, 276
Petroleum and hydrogen, 304
Pettenkofer's series, 263
Phase colloidal, 771
disperse, 769
Phase rule, 444
and solutions, 514
derivation of, 447
Gibbs', 444, 446
modifications, 449
object of, 448
Phases, 446
Phenacite, X -radiogram, 642
Pherecydes, 31
Philathes, Erenaeus, 48
Eupenius, 48
Philosophical chemistry, 3
Phlogiston, 64, 70, 72, 126
Phoenicia, 28
Physical change, 83
Pictet's formiJa, 834
Piezo-electricity, 648
Plait-point, 168
Planck's constant, 811
Plane of sj^nunetry, 614
polarization of hght, 607
Planets, influence on metals, 3, 21
Plasticity, 819
Platinum absorption oxygen, 370
catalysis by, 487
colloidal, 937
palladium alloys. See palladiima.
solubility of hydrogen, 305, 306
Plato, 35
Pliny, 38
Pneumatic chemistry, 122
trough, mercury, 124
Poisson's ratio, 820
Polar number, 211
theory chemical action, 397
valency, 211
Polarity, 211
Goldschmidt and Wright's law, 611
Polarization, 1028
of hght, 607
plane, 607
rotatory, 608
Polarizing microscope, 608
Polymerization in solution, 570, 673
Polymerized liquids, 860
Polymorphism, 596
Polyoxides, 958
Porcelain, catalysis by, 487
Chinese, 23
permeability to gases, 305
Positive chemistry, 4
Potassium, action on water, 136
amalgam, action on water, 135
bromide, X-radiogram, 638
carbonate and hydrogen, 303
chlorate, 591
action heat, 349, 360
chloride, 521, 591
and hydrogen, 303
X-radiogram, 636
hydroxide, 521
and hydrogen, 303
iodide, X-radiogram, 638
nitrate and hydrogen, 303
zonate, 908
perchlorate, 361, 591
solubility of hydrogen, 308
Potential, chemical, 1011
contact difference of, 1016
— differences, 1016
INDEX
1059
Potential difference, 963
discharge, 1031
electrode, 1016
energy, 696
of energy, 727
thermodynamic, 727
Pouillet effect, 495
Poimd-calorie, 699
Praseodymium, solubility of hydrogen,
307
Precipitation, ionic theory, 996
rhythmic, 537
Prefixes, numerical, 117
Prehistoric chemistry, 19
Pressure affinity, 235
and refractive index, 675
cohesive, 841
critical, 165
deposition, 1017
dissociation, 348
effect on solids, 825
• — vol. gases, 150
equilibrium, 348
freezing, 457
internal, 841
intrinsic, 841
— — of liquids, 841
normal, 149, 161
solution, 538, 539, 1017
■ electrolytic, 1017
standard, 149, 161
siu-face, 846
tmits of, 149 •
Pressures, partial, Dalton's law, 155
Principle of reversibility, 93
— — • sulphurous, 64
Prima materia, 31
• hypothesis, 48
Prismatic habit, 597
Probability, 90
Properties, specific, 84
Propionic acid and hydrogen, 303, 304
Proportion, law of compound, 100
Proportionality, law of, 79
Proportions, law of definite, 77
multiple, 93, 96
reciprocal, 97
Protyle, 257
Proust's law, 76
Pseudo peroxides, 958
Pseudomorphs, 595
Pseudotemary system, 524
Pumice, catalysis by, 487
Pure substances, 80, 82
Pyrites, X-radiogram, 641
Pyro-electricity, 648
I^oxone, 946
Pythagoras, 34
Q
712
Quantity factor of energy,
Quantivalence, 224
Quantum, 811
theory of energy, 811
Dulong and Petit's
rule, 811
Quartz, permeability to gases, 305
■ X-radiogram, 642
R
Radicals. See Radicles.
Radicle theories, 216, 217, 221
Radicles, 197
Rankine's vapour pre.ssure formula, 433
Rare earths, 265
asteroid theory, 265
Rate. See Velocity.
Ratios, law of equivalent, 79
Raumgitter, 624
Ray, extraordinary, 607
ordinary, 607
Reacting weights, 99
Reaction, energy cost, 716
heat of, 698
Reactions and pressure, 300
balanced, 299
— — catalytic, 358
— — chemical, 291
complete, 299
concurrent, 360
consecutive, 359
• incomplete, 299
irreversible, 299
opposing, 299
reversible, 299
side, 360
speed, 294
trigger, 358
with compressed solids, 826
solids, 824, 826
Spring's experiments, 824
Reason, 13
Reduction, 64, 210
by hydrogen, 332
Refraction, atomic, 673
— double, 607
index of, 670, 671
molecular, 673
specific, 673
Refractive constants, 675
• energy, 673
— specific, 673
index and chemical composition, 677
critical temperature, 675
dielectric constant, 683
dispersion, 677
effect of pressure, 675
temperature, 675
of gases, 68 1
and isomerism, 685
magnetic rotatory power,
681
— mixture law, 678
valency, 681
Refractivity, 673
Roseau, 624
Residual current, 1030
Residues, theory of, 219
Resistance, chemical, 293
electrical, 963
passive, 152
specific electrical, 978
Retger's colour test mixed crystals, 660
law mixed crystals, 660
Reticular density, 628
Reversibility, principle of, 93, 706
Reversible cells, 1021
— — colloid, 771
1060
INDEX
Reversible processes, 717
Rey, J., on calcination, 56
R-gas constfiuit, 161
Rhfises, A. M., 41
Rhind's papyrus, 26
Rhodium, catalysis by, 487
solubility of hydrogen, 306
Rhodochrosite, X-radiogram, 641
Rhombic system, 619
Rhythmic crystallization, 599
precipitation, 537
Richards' formula, 835
Richter's law, 79, 97
of neutrality, 391
Rigidity solids, 820
Rings, Liesegang's, 537
Ripley, G., 48
Robertson's formula, 835
Rome, 37
Rose's crucible, 329
Rosicrucians, Society of, 4
Rotatory polarization, 608
power, molecTilar, 609
specific, 609
Royal Society, 5
Ruthenium, solubility of hydrogen, 307
Rutile, X-radiogram, 641
S
Sacrificial metals, 1025
Sala, A., 51
Salt, 389
(element), 34
history, 382, 384
hydratod, 397
neutral, 384
solutions and gas solubility, 534
Salts, 393
acid, 387
and acids, reactions, 1002
reactions, 1002
basic, 394
constitution theories, 403
hydrated, 498
neutral, 388
normal, 387
Samariuun, solubility of hydrogen, 307
Saturation, 384
capacity, 224
Scandium, solubihty of hydrogen, 307
Scheelite, X-radiogram, 642
Science, object of, 10
Scientific chemistry, 4
knowledge, 8
Scolecite, X-radiogram, 642
Seeding solutions, 451
Sendibogius, M., 48
Seneca, A., 38
Sensation, 6
Senses, 6
Series of elements, 255
even, 255
odd, 255
Serum and hydrogen, 304
Sesqui oxides, 118
Settling of particles in water, 774
Shear modulus, 820
Sical process hydrogen, 284
Side reactions, 360
Siderite, X-radiogram, 641
Siemens' ozonizer, 886
Silica, X-radiogram, 642
Silicol, 284
Silicon, eka-, 261
X-radiogram, 642
Silver absorption oxygen, 371
catalysis by, 487
nitrate, 521
electrolysis, 962
palladium alloys. Sec Palladium.
permeability to oxygen, 371
solubihty of hydrogen, 306, 306
sulphide, 520
voltameter, 964
X-radiogram, 641
Sines, law of, 670
Size of molecules, 752, 755
Smee's cell, 1028
Snow, 464
Society of Rosicrucians, 4
Royal, 6
Sodivun, action on water, 135
amalgam, action on water, 135
carbonate and hydrogen, 303
chlorate, X-radiogram, 642
chloride and hydrogen, 303
X-radiogram, 636
hydroxide and hydrogen, 303
nitrate, 521
and hydrogen, 303
X-radlogram, 641
solubihty of hydrogen, 308
sulphate and hydrogen, 303
solubility, 514
X-radiogram, 642
Sol, 771
Solid solution, 659
Solids seolotropic, 820
crystallization of, 602
effect pressure, 825
empirical formiilse for properties, 834
equation of state, 834
isotropic, 820
kinetic theory, 819
Langmuir's theory, 642
reactions with, 824, 826
specific heat of, 798
Solubility, 506
and intrinsic pressure, 862
apparent, 996
chemical composition and, 686
effect grain-size, 608
of pressiu-e, 611
temperature, 510
gases in salt solutions, 636
ion theory, 995
law, 995
and melting point, 586
mixed gases, 533 .
mixtures with common ion, 995
no common ion, 999
molecular, 996
and molecular weight, 568
of gases, effect of pressure, 629
temperature, 532
and osmotic pressure, 669
product, 996
real, 996
INDEX
1061
Solute, 506
Solution and compressibility of solvent, 529
definition, 772
and dielectric constant of solvent, 529
cause of, 574
concentration, 607
definition, 507
kinetic theory and, 524, 528
number ions, 978
pressure, 538, 539, 1015, 1017
electrolytic, 1017
hypothesis, osmotic pressure, 558
rate of, 537
solid, 659
solvate theory, i?94
standard, 391
temperature, critical, 523
Solutions, 95
and Avogadro's hypothesis, 545
phase rule, 514
compressibility, 581
effect on solvent, 509
electrolytic conductivity, 977
freezing, 516
heat of, 582
isotonic, 539
molecular volume, 578
physical properties, 578
specific gravity, 578
surface tension, 853
thermal expansion, 581
viscosity, 681
Solvate theory of solution, 994
Solvent, 606
effect on electrolysis, 968
universal, 50
Sorption, 311
Space -lattice, 624
Spagyric art, 91
Specific gravity, 87
cohesion, 848
gravities, colloids, 774
gravity and index of refraction, 672
isomorphism, 657
gases, 175
heat. See Heat, specific ; heat, atomic ;
heat, molecular,
gases, ratio of two, 788
heats of gases, ratio of two, and degree
of freedom, 790
effect of pres-
sure, 788
effect of tem-
perature, 788
• molecular
weights, 788
volumes, 228
colloids, 774
Speed. See Velocity.
Spectrometer, X-ray, 635
Spectriun, X-ray, 636
Spirit, 122
Spiritus, 122
Spring's experiments on reactions with
solids, 824
Stability function of energy, 727
Stable equilibrium, 714
Stahl, G. E., 65
Standard solution, 391
State colloidal, 771
State critical, 164, 165
States, corresponding, 759, 760
van der Waals' theory, 759
of aggregation, 164
Status nascens, 331
Steam curve, 444
decomposition by red-hot iron, 935
electrolysis, 493
Steel, absorption oxygen, 371
Store, 237
Stereochemistry, 214
Stimulants in chemical actions, 359
Stone age, 19
Straight extinction, 608
Strain, 819
theory, valency, 215
Strength factor of energy, 712
Stress, 819
Strong acids, 981
bases, 981
ions, 1015
Strontium, action on water, 135
nitrate, X-radiogram, 642
Structure, chemical compounds, 223
Struvite, X-radiogram, 642
Stupa, 23
Sublimation curve, 444
Suboxides, 118
Substitution theory, 218
Sugar and hydrogen, 304
Sulphur adustible, 64
ardens, 64, 67
combustible, 64
dioxide, chlorine, 518
effect on catalysis, 487
(element), 34
fixed, 64
of Mars, volatile, 125
wine, 64
wood, 64
phlogistic, 64
sideric, 64
volatile, 64
Sulphuric acid and hydrogen, 303
Sulphurs, 64
Sulphury 1 chloride, 518
Super-, 118
Superoxides, 958
Supersaturation, 450, 451
and phase rule, 454
kinetic theory, 455
Superstition in chemistry, 2
Surface energy,. 712, 846, 847
liquids, 855
pressure, 846
tension, 846, 847
and chemical composition, 853
compressibility, 860
concentration, 854
heat of vaporization, 851
intrinsic pressure, 842
melting point, 852
specific heat, 852
colloids, 774
effect of temperature, 849
hypothesis, osmotic pressure, 560
solutions, 853
Surfusion, 451
Suspensoids, 770
Sutherland's formula, 835
1062
INDEX
Sylvius de la Boe, F., 62
Symmetry, axes of, CI 4
— — centre of, 614
crystals, 613
hemihedral, 613
holchedral, 613
plane of, 614
tetartohedral, 613
Synthesis, 91
System, cubic, 616
hexagonal, 617
monoclinic, 621
rhombic, 619
tetragonal, 619
trigonal, 618
triclinic, 621
Systems, crystal, 616
Tabular habit, 697
Tachen, O., 52
Tantalum, solubility of hydrogen, 307
Taouists, 23
Telluric screw, 253
Tellurium, 264
Temperature, absolute, 160
action on vol. gases, 158, 160
coefficient of reactions, 702
critical, 165
solution, 523
effect on chemical equilibria, 732
solubility of gases, 532
eu tactic, 517
freezing, 457
inversion, 866
normal, 161
— and osmotic pressure, 545
refractive index, 675
standard, 161
transition, 513
Temperatures, transition, 512, 513
Tensile strength, 821, 822
liquids, 421
Terbium, solubiUty of hydrogen, 307
Ternary system, pseudo-, 524
Terra damnata, 55
fluida, 64
lapida, 64
mercurialis, 64
pinguis, 64
vitrescibilis, 64
Tetartohedral, symmetry, 613
Tetrabase paper, 950
Tetrad, 224
Tetrads, 206
Tetragonal system, 619
Tetrahedron theory, carbon atom, 214
Tetramethyl paper, 950
Tetramorphism, 596
Tetra-paper. 950
Thales, 31
Thallium, solubility of hydrogen, 306, 308
Theophrastus, 36
Theories, 72
Theory, 13
Thermal analysis, 518
conductivity and isomorphism, 658
and electrical energy relation, 1036
Thermochemical constant, 710
Thermochemistry, 697, 698, 711
Thermodynamic potential, 727
Thermodynamics, 711
first law, 693, 694
second law, 713
Thermoneutrality, Hess' law, 1007, 1008
Thomas Aquinas, 46
Thompson. See Kelom.
Thorite, X-radiogram, 642
Thorium, solubUity of hydrogen, 307
Thulium, solubility of hydrogen, 307
Tin, solubility of hydrogen, 306
— — X-radiogram, 642
Titanium, solubility of hydrogen, 307
Toluene and hydrogen, 304
Tomlinson's formula, 835
Topaz, X-radiogram, 642
Topic axes, 656
parameters, 656
Total energy, 717
Toiu^maline, X-radiogram, 642
Transition point, 513
action of pressure, 429
Transmutation of metals, 49
Transport numbers, 985, 986
Hittorf 's, 985
Triad, 224
Triads, 206
Bobereiner's, 253
Tria prima, 34
Tribo -luminescence, 600
Triboluminiscope, 600
Trichitic crystals, 597
Triclinic system, 621
Trigger reactions, 358
Trigonal system, 618
Trihydrol, 461
Trimorphism, 596
Triple point, 446
Trough, pneumatic, 123
Trouton's rule, 440
Tungsten, solubility of hydrogen, 306
Twin, 595
Twinning, mimetic, 595
of crystals, 595
Tycho Brahe, 47
Tyndall's test, optical emptiness, 768
Type theoiy, 217, 218, 220
of condensed, 220
mixed, 221
U
Ullmanite, X-radiogram, 641
Ultrafiltration, 772
Ultramicrons, 770
Ultramicroscope, 769
Ultramicroscopic particles, 768
Ultramicroscopy, 768
Undercooling, 450
Uniaxial crystals, 607
Units, electrical, 963
of energy, 693
Univariant systems, 446, 447
Universal medicine, 49
solvent, 50
Unstable states, 454
INDEX
1063
Uranium, solubility of hydrogen, 307
Urea and hydrogen, 304
Val, 392
Valence, 205, 224
Valencies, affini-, 225
crypto-, 208
dormant, 208
electrical double, 213
latent, 208, 213
passive, 208
— — residual, 213
secondary, 213
sleeping, 208
unsaturated, 213
Valency, 204, 224, 784
Abegg's theory, 212
absolute, 209
active, 207, 209
Baeyer's strain theory, 215
Barlow and Pope's theory, 241
bodies, 225
contra-, 212
doctrine, 222
effect of light, 210
pressure, 210
radiant energy, 210
temperature, 210
force, 225
free, 209
history, 216
maximum, 20?
negative, 211
normal, 212
polar, 211
positive, 211
and refractive index, 681
theories of, 225
volume, 241
zero-, 206
ValentLue Basil, 52
Vanadium iron, 520
solubility of hydrogen, 306
Van der Waals' vapour pressure formula,
433
Vaporization curve, 444
heat of, 426
Vapour and gas, 435
density, abnormal, 192
determination, 181
Dimaas' process, 184
Hofmann's process, 186
Meyer's process, 185
pressure, 431
and boiling point, 561, 565
colloids, 774
constant, 551
hypothesis, osmotic pressure, 558
and molecular weight of solute,
548
Nemst's formula, 434
and osmotic pressure, 550
of small drops, 453
Rankine's formula, 433
Raoult's law, 550
van der Waals' formula, 433
Variables, dependent, 446
Variables, independent, 445
of system, 445
Variance of system, 445
Varro, M. T., 38
Vaughan, T., 48
Vedas, 22
Velocity of chemiced reactions, 294
colloidal particles, 776
electrical conduction, 967
of molecular motion, 792
Boltzmann's theorem,
792
Maxwell's theorem, 792
molecCiles, 744
Vibration frequency, 828
and heat fusion, 833
atoms, 828
molecules, 828
Vibratory volume, 755
Vicarious constituents, 651
Virtual work, principle of, 714
Viscosities, colloids, 774
Viscosity coefficient, 749
fluids, 749
Vitiated air, 344
Vitriols, 383
Vitruvius, 37
Volatile sulphxir of Mars, 125
Volt, 963
Volta's law, 158
Voltage, 963
decomposition, 965, 1031
— • and concentration, 1039
Voltameter, copper, 964
silver, 964
Volume, atom, 188
critical, 165
crystal, 656
elasticity, 820
energy, 712
gases, 150
effect of temperature, 158, 160
effect pressure, 150
joint effect, temp, and press., 161
moist gases, measuring, 438
molecular, 416
of atom, oscillatory, 233
vibratory, 233
theory, 188
valency, 241
vibratory, 755
Volumes and molecular weights, 201
atomic, 228
molecular, 176, 195, 228
law of combining, 171
specific, 228
W
Waals' equation of state for sohds, 836
gas equation, 756
theory corresponding states, 769
Water, absorption spectrum, 474
action, alinninium, 494
barium, 135
boron, 494
— calcium, 135
• carbides, 494
chromous oxide, 494
1064
INDEX
Water, action, esters, 494
halogens, 493, 494
hydrides, 494
iodine, 494
iron, ] 34
' magnesium, 135
manganese oxide, 494
metals, 493
metal dioxides, 494
molybdenoas chloride, 494
nitrides, 494
non-metal oxides, 494
organometallic compounds, 494
phosphides, 494
phosphorus, 494
potassium, 135
amalgam, 135
cobaltocyanide, 494
selenides, 494
silicides, 494
sodiimi, 135
amalgam, 135
strontiiun, 135
sulphides, 494
sulphur, 494
uranium oxide, 494
zinc, 134
adsorption by solids, 495
allotropic states, 457
bath, 49
boiling point, 436
colour, 473
composition of. Cavendish, 138
(gravimetric), 129
of, Lavoisier, 140
Watt, 141
compressibility, 418
conductivity, 410
critical temperature, 437
crystallization, 463
crystals, 464
cycle in nature, 405
decomposition, 136, 490
by metals, 134
density, critical, 438
dielectric capacity, 478
diffusion, 469
dispersion, 472
dissociation, 492
distillation, 409
drinking, 408
electrical conductivity, 476
electrolysis, 136, 277, 356
(element), 31
energy formation, 489
entropy, 470
evaporation velocity, 424
formation of, 127
free energy, 490
freezing, 463
fresh, 406
fusion heat of, 428
gas, 281
gravimetric composition, 129
Dimaas, 130
Morley, 132
ground, 406
hard, 407
heat conductivity, 471
formation, 489
Water, heat ionization, 477
influence in chemical action, 377
ionizing constant, 476
potential, 476
Kerr's electro-optic effect, 480
liquid, constitution, 461
molecular state, 460
Sutherland's theory constitution,
461
magnetic susceptibility, 479
magnetization, 479
magneto-optic rotation. 479
maximum density, 413
mineral, 406
molecular formula, 460
■ volume, 416
molecule, diameter, 460
mean free path, 460
number per c.c, 460
velocity, 460
molecules, collision frequency, 460
gasogenic, 410
ice, 411
liquidogenic, 411
water, 410
optical properties, 472
ozone, 898
photo-electric effect, 480
potable, 408
pressure coefficient, 429
critical, 438
purification, 409
rain, 406, 407
refractive index, 472
saline, 407
sea, 407
specific cohesion, 469
— — gravity, 415
heat, 469
spring, 406
(steam) and iron, 297
sulphur, 406
surface, 406
tension, 467
tensile strength, 422
thermal expansion, 412
to earth, transformation, 81
transition point, 429
effect pressure, 429
undergroim.d, 406
vaporization, heat of, 426
vapour pressure, 423, 431, 435
formulae, 433
. See Steam.
velocity formation, 483
soiuid, 469
Verdet's constant, 479
viscosity, 465
volume, 438
effect pressure, 410
temperature, 410, 414
synthesis, 143
Cavendish, 143
Hofmann, 145
voliunetric compositions, 139
Cavendish, 139
Waters, chalybeate, 406
Waterston's hypothesis, 747
Weak acids, 981
981
INDEX
1065
Weak ions, 1016
Weight, formula, 179
increase during calcination, 55
law of persistence, 101
molar, 176
of matter, 66
Weights, atomic, 104, 180, 181
combining, 99
equivalent, 99
reacting, 99
Wilhelmy's law, 294
Woestyn's rule, 806
Work, 689
in changing volume of gases, 690
of chemical reaction, 730
external, 695
internal, 695
maximum, 703
value of heat, 719
virtual, principle of, 714
Wulfenite, X-radiogram, 642
Wiillner's law, 648
Xenotime, X-radiogram, 642
X-radiograms, crystals, 634
X-ray analysis, crystal structure, 633
X-ray spectrometer, 635
— — spectrum, 636
Xylene and hydrogen, 304
Yang, 23
Yield point, 819
Yin, 23
Yoimg's modulus, 820
Ytterbium, solubility of hydrogen, 307
Yttrium, solubility of hydrogen, 306
Zero, absolute, 160
Zinc, action on water, 134
blende, X-radiogram, 640
catalysis by, 487
solubility of hydrogen, 306
sulphate and hydrogen, 303
X-radiogram, 642
Zircon light, 326
X-radiogram, 641
Zirconium, solubility of hydrogen, 306
Zoroaster, 20
Zosimos, 39
END OF VOL. I.
PRINTED IIT GREAT BRITAIN BY WILLIAM CLOWES AND SONS, LIMITED, BECCLES.
rare B&fe IS DUE ON THE LAST DATE
STAJHPED BELOW
;jA\M 1 i riv>/'
JAN 1 1 1938
FEB 2 0 1939
,jUU^-^342 .
jutr 2 6 m^
-
i
lm-4,'88